Files
test/source/blender/blenkernel/intern/node_tree_update.cc

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1933 lines
65 KiB
C++
Raw Normal View History

/* SPDX-FileCopyrightText: 2023 Blender Authors
*
* SPDX-License-Identifier: GPL-2.0-or-later */
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include <fmt/format.h>
#include "BLI_listbase.h"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "BLI_map.hh"
#include "BLI_multi_value_map.hh"
#include "BLI_noise.hh"
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
#include "BLI_rand.hh"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "BLI_set.hh"
#include "BLI_stack.hh"
#include "BLI_string.h"
#include "BLI_string_utf8_symbols.h"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "BLI_vector_set.hh"
#include "DNA_anim_types.h"
#include "DNA_modifier_types.h"
#include "DNA_node_types.h"
#include "BKE_anim_data.hh"
#include "BKE_image.hh"
#include "BKE_lib_id.hh"
#include "BKE_main.hh"
#include "BKE_node.hh"
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
#include "BKE_node_enum.hh"
#include "BKE_node_legacy_types.hh"
#include "BKE_node_runtime.hh"
Refactor: Geometry Nodes: improve lifetime analysis for anonymous attributes This refactors the lifetime analysis of anonymous attributes in geometry nodes. The refactor has a couple of goals: * Use a better and simpler abstraction that can be used when building the lazy-function graph. We currently have a bunch of duplicate code to handle "field source" and "caller propagation" attributes. This is now unified so that one only has to worry about one kind of "reference sets". * Make the abstraction compatible with handling bundles and closures in case we want to support them in the future. Both types can contain geometries and fields so they need to be taken into account when determining lifetimes. * Make more parts independent of the concept of "anonymous attributes". In theory, there could be more kinds of referenced data whose lifetimes need to be managed. I don't have any concrete plans for adding any though. At its core, deterministic anonymous attributes still work the same they have been since they became deterministic [0]. Even the generated lazy-function graph is still pretty much or even exactly the same as before. The patch renames `AnonymousAttributeSet` to the more general `GeometryNodesReferenceSet` which is more. This also makes more places independent of the concept of anonymous attributes. Functionally, this still the same though. It's only used in the internals of geometry nodes nowadays. Most code just gets an `AttributeFilter` that is based on it. [0]: https://archive.blender.org/developer/D16858 Pull Request: https://projects.blender.org/blender/blender/pulls/128667
2024-10-07 12:59:39 +02:00
#include "BKE_node_tree_reference_lifetimes.hh"
#include "BKE_node_tree_update.hh"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "MOD_nodes.hh"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "NOD_geo_closure.hh"
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
#include "NOD_geometry_nodes_dependencies.hh"
Geometry Nodes: support attaching gizmos to input values This adds support for attaching gizmos for input values. The goal is to make it easier for users to set input values intuitively in the 3D viewport. We went through multiple different possible designs until we settled on the one implemented here. We picked it for it's flexibility and ease of use when using geometry node assets. The core principle in the design is that **gizmos are attached to existing input values instead of being the input value themselves**. This actually fits the existing concept of gizmos in Blender well, but may be a bit unintutitive in a node setup at first. The attachment is done using links in the node editor. The most basic usage of the node is to link a Value node to the new Linear Gizmo node. This attaches the gizmo to the input value and allows you to change it from the 3D view. The attachment is indicated by the gizmo icon in the sockets which are controlled by a gizmo as well as the back-link (notice the double link) when the gizmo is active. The core principle makes it straight forward to control the same node setup from the 3D view with gizmos, or by manually changing input values, or by driving the input values procedurally. If the input value is controlled indirectly by other inputs, it's often possible to **automatically propagate** the gizmo to the actual input. Backpropagation does not work for all nodes, although more nodes can be supported over time. This patch adds the first three gizmo nodes which cover common use cases: * **Linear Gizmo**: Creates a gizmo that controls a float or integer value using a linear movement of e.g. an arrow in the 3D viewport. * **Dial Gizmo**: Creates a circular gizmo in the 3D viewport that can be rotated to change the attached angle input. * **Transform Gizmo**: Creates a simple gizmo for location, rotation and scale. In the future, more built-in gizmos and potentially the ability for custom gizmos could be added. All gizmo nodes have a **Transform** geometry output. Using it is optional but it is recommended when the gizmo is used to control inputs that affect a geometry. When it is used, Blender will automatically transform the gizmos together with the geometry that they control. To achieve this, the output should be merged with the generated geometry using the *Join Geometry* node. The data contained in *Transform* output is not visible geometry, but just internal information that helps Blender to give a better user experience when using gizmos. The gizmo nodes have a multi-input socket. This allows **controlling multiple values** with the same gizmo. Only a small set of **gizmo shapes** is supported initially. It might be extended in the future but one goal is to give the gizmos used by different node group assets a familiar look and feel. A similar constraint exists for **colors**. Currently, one can choose from a fixed set of colors which can be modified in the theme settings. The set of **visible gizmos** is determined by a multiple factors because it's not really feasible to show all possible gizmos at all times. To see any of the geometry nodes gizmos, the "Active Modifier" option has to be enabled in the "Viewport Gizmos" popover. Then all gizmos are drawn for which at least one of the following is true: * The gizmo controls an input of the active modifier of the active object. * The gizmo controls a value in a selected node in an open node editor. * The gizmo controls a pinned value in an open node editor. Pinning works by clicking the gizmo icon next to the value. Pull Request: https://projects.blender.org/blender/blender/pulls/112677
2024-07-10 16:18:47 +02:00
#include "NOD_geometry_nodes_gizmos.hh"
#include "NOD_geometry_nodes_lazy_function.hh"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
#include "NOD_node_declaration.hh"
#include "NOD_socket.hh"
#include "NOD_texture.h"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
#include "DEG_depsgraph_build.hh"
#include "BLT_translation.hh"
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
using namespace blender::nodes;
/**
* These flags are used by the `changed_flag` field in #bNodeTree, #bNode and #bNodeSocket.
* This enum is not part of the public API. It should be used through the `BKE_ntree_update_tag_*`
* API.
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
*/
enum eNodeTreeChangedFlag {
NTREE_CHANGED_NOTHING = 0,
NTREE_CHANGED_ANY = (1 << 1),
NTREE_CHANGED_NODE_PROPERTY = (1 << 2),
NTREE_CHANGED_NODE_OUTPUT = (1 << 3),
NTREE_CHANGED_LINK = (1 << 4),
NTREE_CHANGED_REMOVED_NODE = (1 << 5),
NTREE_CHANGED_REMOVED_SOCKET = (1 << 6),
NTREE_CHANGED_SOCKET_PROPERTY = (1 << 7),
NTREE_CHANGED_INTERNAL_LINK = (1 << 8),
NTREE_CHANGED_PARENT = (1 << 9),
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
NTREE_CHANGED_ALL = -1,
};
static void add_tree_tag(bNodeTree *ntree, const eNodeTreeChangedFlag flag)
{
ntree->runtime->changed_flag |= flag;
ntree->runtime->topology_cache_mutex.tag_dirty();
Geometry Nodes: add simulation support This adds support for building simulations with geometry nodes. A new `Simulation Input` and `Simulation Output` node allow maintaining a simulation state across multiple frames. Together these two nodes form a `simulation zone` which contains all the nodes that update the simulation state from one frame to the next. A new simulation zone can be added via the menu (`Simulation > Simulation Zone`) or with the node add search. The simulation state contains a geometry by default. However, it is possible to add multiple geometry sockets as well as other socket types. Currently, field inputs are evaluated and stored for the preceding geometry socket in the order that the sockets are shown. Simulation state items can be added by linking one of the empty sockets to something else. In the sidebar, there is a new panel that allows adding, removing and reordering these sockets. The simulation nodes behave as follows: * On the first frame, the inputs of the `Simulation Input` node are evaluated to initialize the simulation state. In later frames these sockets are not evaluated anymore. The `Delta Time` at the first frame is zero, but the simulation zone is still evaluated. * On every next frame, the `Simulation Input` node outputs the simulation state of the previous frame. Nodes in the simulation zone can edit that data in arbitrary ways, also taking into account the `Delta Time`. The new simulation state has to be passed to the `Simulation Output` node where it is cached and forwarded. * On a frame that is already cached or baked, the nodes in the simulation zone are not evaluated, because the `Simulation Output` node can return the previously cached data directly. It is not allowed to connect sockets from inside the simulation zone to the outside without going through the `Simulation Output` node. This is a necessary restriction to make caching and sub-frame interpolation work. Links can go into the simulation zone without problems though. Anonymous attributes are not propagated by the simulation nodes unless they are explicitly stored in the simulation state. This is unfortunate, but currently there is no practical and reliable alternative. The core problem is detecting which anonymous attributes will be required for the simulation and afterwards. While we can detect this for the current evaluation, we can't look into the future in time to see what data will be necessary. We intend to make it easier to explicitly pass data through a simulation in the future, even if the simulation is in a nested node group. There is a new `Simulation Nodes` panel in the physics tab in the properties editor. It allows baking all simulation zones on the selected objects. The baking options are intentially kept at a minimum for this MVP. More features for simulation baking as well as baking in general can be expected to be added separately. All baked data is stored on disk in a folder next to the .blend file. #106937 describes how baking is implemented in more detail. Volumes can not be baked yet and materials are lost during baking for now. Packing the baked data into the .blend file is not yet supported. The timeline indicates which frames are currently cached, baked or cached but invalidated by user-changes. Simulation input and output nodes are internally linked together by their `bNode.identifier` which stays the same even if the node name changes. They are generally added and removed together. However, there are still cases where "dangling" simulation nodes can be created currently. Those generally don't cause harm, but would be nice to avoid this in more cases in the future. Co-authored-by: Hans Goudey <h.goudey@me.com> Co-authored-by: Lukas Tönne <lukas@blender.org> Pull Request: https://projects.blender.org/blender/blender/pulls/104924
2023-05-03 13:18:51 +02:00
ntree->runtime->tree_zones_cache_mutex.tag_dirty();
Nodes: automatically gray out input values that don't affect the output This patch automatically grays out input values which can't affect the output currently. It works with inputs of group nodes, geometry nodes modifiers and node tools. To achieve this, it analyses the node tree and partially evaluates it to figure out which group inputs are currently not linked to an output or are disabled by e.g. some switch node. Original proposal: https://devtalk.blender.org/t/dynamic-socket-visibility/31874 Related info in blog post: https://code.blender.org/2023/11/geometry-nodes-workshop-november-2023/#dynamic-socket-visibility Follow up task for designing a UI that allows hiding sockets: #132706 Limitations: * The inferencing does not update correctly when a socket starts being animated/driven. I haven't found a good way to invalidate the cache in a good way reliably yet. It's only a very short term problem though. It fixes itself after the next modification of the node tree and is only noticeable when animating some specific sockets such as the switch node condition. * Whether a socket is grayed out is not exposed in the Python API yet. That will be done separately. * Only a partial evaluation is done to determine if an input affects an output. There should be no cases where a socket is found to be unused when it can actually affect the output. However, there can be cases where a socket is inferenced to be used even if it is not due to some complex condition. Depending on the exact circumstances, this can either be improved or the condition in the node tree should be simplified. Pull Request: https://projects.blender.org/blender/blender/pulls/132219
2025-01-21 12:53:24 +01:00
ntree->runtime->inferenced_input_socket_usage_mutex.tag_dirty();
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
static void add_node_tag(bNodeTree *ntree, bNode *node, const eNodeTreeChangedFlag flag)
{
add_tree_tag(ntree, flag);
node->runtime->changed_flag |= flag;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
static void add_socket_tag(bNodeTree *ntree, bNodeSocket *socket, const eNodeTreeChangedFlag flag)
{
add_tree_tag(ntree, flag);
socket->runtime->changed_flag |= flag;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
namespace blender::bke {
/**
* Common datatype priorities, works for compositor, shader and texture nodes alike
* defines priority of datatype connection based on output type (to):
* `< 0`: never connect these types.
* `>= 0`: priority of connection (higher values chosen first).
*/
static int get_internal_link_type_priority(const bNodeSocketType *from, const bNodeSocketType *to)
{
switch (to->type) {
case SOCK_RGBA:
switch (from->type) {
case SOCK_RGBA:
return 4;
case SOCK_FLOAT:
return 3;
case SOCK_INT:
return 2;
case SOCK_BOOLEAN:
return 1;
}
return -1;
case SOCK_VECTOR:
switch (from->type) {
case SOCK_VECTOR:
return 4;
case SOCK_FLOAT:
return 3;
case SOCK_INT:
return 2;
case SOCK_BOOLEAN:
return 1;
}
return -1;
case SOCK_FLOAT:
switch (from->type) {
case SOCK_FLOAT:
return 5;
case SOCK_INT:
return 4;
case SOCK_BOOLEAN:
return 3;
case SOCK_RGBA:
return 2;
case SOCK_VECTOR:
return 1;
}
return -1;
case SOCK_INT:
switch (from->type) {
case SOCK_INT:
return 5;
case SOCK_FLOAT:
return 4;
case SOCK_BOOLEAN:
return 3;
case SOCK_RGBA:
return 2;
case SOCK_VECTOR:
return 1;
}
return -1;
case SOCK_BOOLEAN:
switch (from->type) {
case SOCK_BOOLEAN:
return 5;
case SOCK_INT:
return 4;
case SOCK_FLOAT:
return 3;
case SOCK_RGBA:
return 2;
case SOCK_VECTOR:
return 1;
}
return -1;
case SOCK_ROTATION:
switch (from->type) {
case SOCK_ROTATION:
return 3;
case SOCK_VECTOR:
return 2;
case SOCK_FLOAT:
return 1;
}
return -1;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
/* The rest of the socket types only allow an internal link if both the input and output socket
* have the same type. If the sockets are custom, we check the idname instead. */
if (to->type == from->type && (to->type != SOCK_CUSTOM || to->idname == from->idname)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return 1;
}
return -1;
}
/* Check both the tree's own tags and the interface tags. */
static bool is_tree_changed(const bNodeTree &tree)
{
return tree.runtime->changed_flag != NTREE_CHANGED_NOTHING || tree.tree_interface.is_changed();
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
using TreeNodePair = std::pair<bNodeTree *, bNode *>;
using ObjectModifierPair = std::pair<Object *, ModifierData *>;
using NodeSocketPair = std::pair<bNode *, bNodeSocket *>;
/**
* Cache common data about node trees from the #Main database that is expensive to retrieve on
* demand every time.
*/
struct NodeTreeRelations {
private:
Main *bmain_;
std::optional<Vector<bNodeTree *>> all_trees_;
std::optional<MultiValueMap<bNodeTree *, TreeNodePair>> group_node_users_;
std::optional<MultiValueMap<bNodeTree *, ObjectModifierPair>> modifiers_users_;
public:
NodeTreeRelations(Main *bmain) : bmain_(bmain) {}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void ensure_all_trees()
{
if (all_trees_.has_value()) {
return;
}
all_trees_.emplace();
if (bmain_ == nullptr) {
return;
}
FOREACH_NODETREE_BEGIN (bmain_, ntree, id) {
all_trees_->append(ntree);
}
FOREACH_NODETREE_END;
}
void ensure_group_node_users()
{
if (group_node_users_.has_value()) {
return;
}
group_node_users_.emplace();
if (bmain_ == nullptr) {
return;
}
this->ensure_all_trees();
for (bNodeTree *ntree : *all_trees_) {
for (bNode *node : ntree->all_nodes()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (node->id == nullptr) {
continue;
}
ID *id = node->id;
if (GS(id->name) == ID_NT) {
bNodeTree *group = (bNodeTree *)id;
group_node_users_->add(group, {ntree, node});
}
}
}
}
void ensure_modifier_users()
{
if (modifiers_users_.has_value()) {
return;
}
modifiers_users_.emplace();
if (bmain_ == nullptr) {
return;
}
LISTBASE_FOREACH (Object *, object, &bmain_->objects) {
LISTBASE_FOREACH (ModifierData *, md, &object->modifiers) {
if (md->type == eModifierType_Nodes) {
NodesModifierData *nmd = (NodesModifierData *)md;
if (nmd->node_group != nullptr) {
modifiers_users_->add(nmd->node_group, {object, md});
}
}
}
}
}
Span<ObjectModifierPair> get_modifier_users(bNodeTree *ntree)
{
BLI_assert(modifiers_users_.has_value());
return modifiers_users_->lookup(ntree);
}
Span<TreeNodePair> get_group_node_users(bNodeTree *ntree)
{
BLI_assert(group_node_users_.has_value());
return group_node_users_->lookup(ntree);
}
};
struct TreeUpdateResult {
bool interface_changed = false;
bool output_changed = false;
};
class NodeTreeMainUpdater {
private:
Main *bmain_;
const NodeTreeUpdateExtraParams &params_;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
Map<bNodeTree *, TreeUpdateResult> update_result_by_tree_;
NodeTreeRelations relations_;
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
bool needs_relations_update_ = false;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
public:
NodeTreeMainUpdater(Main *bmain, const NodeTreeUpdateExtraParams &params)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
: bmain_(bmain), params_(params), relations_(bmain)
{
}
void update()
{
Vector<bNodeTree *> changed_ntrees;
FOREACH_NODETREE_BEGIN (bmain_, ntree, id) {
if (is_tree_changed(*ntree)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
changed_ntrees.append(ntree);
}
}
FOREACH_NODETREE_END;
this->update_rooted(changed_ntrees);
}
void update_rooted(Span<bNodeTree *> root_ntrees)
{
if (root_ntrees.is_empty()) {
return;
}
bool is_single_tree_update = false;
if (root_ntrees.size() == 1) {
bNodeTree *ntree = root_ntrees[0];
if (!is_tree_changed(*ntree)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return;
}
const TreeUpdateResult result = this->update_tree(*ntree);
update_result_by_tree_.add_new(ntree, result);
if (!result.interface_changed && !result.output_changed) {
is_single_tree_update = true;
}
}
if (!is_single_tree_update) {
Vector<bNodeTree *> ntrees_in_order = this->get_tree_update_order(root_ntrees);
for (bNodeTree *ntree : ntrees_in_order) {
if (!is_tree_changed(*ntree)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
continue;
}
if (!update_result_by_tree_.contains(ntree)) {
const TreeUpdateResult result = this->update_tree(*ntree);
update_result_by_tree_.add_new(ntree, result);
}
const TreeUpdateResult result = update_result_by_tree_.lookup(ntree);
Span<TreeNodePair> dependent_trees = relations_.get_group_node_users(ntree);
if (result.output_changed) {
for (const TreeNodePair &pair : dependent_trees) {
add_node_tag(pair.first, pair.second, NTREE_CHANGED_NODE_OUTPUT);
}
}
if (result.interface_changed) {
for (const TreeNodePair &pair : dependent_trees) {
add_node_tag(pair.first, pair.second, NTREE_CHANGED_NODE_PROPERTY);
}
}
}
}
for (const auto item : update_result_by_tree_.items()) {
bNodeTree *ntree = item.key;
const TreeUpdateResult &result = item.value;
this->reset_changed_flags(*ntree);
if (result.interface_changed) {
if (ntree->type == NTREE_GEOMETRY) {
relations_.ensure_modifier_users();
for (const ObjectModifierPair &pair : relations_.get_modifier_users(ntree)) {
Object *object = pair.first;
ModifierData *md = pair.second;
if (md->type == eModifierType_Nodes) {
MOD_nodes_update_interface(object, (NodesModifierData *)md);
}
}
}
}
if (result.output_changed) {
ntree->runtime->geometry_nodes_lazy_function_graph_info.reset();
}
ID *owner_id = BKE_id_owner_get(&ntree->id);
ID &owner_or_self_id = owner_id ? *owner_id : ntree->id;
if (params_.tree_changed_fn) {
params_.tree_changed_fn(*ntree, owner_or_self_id);
}
if (params_.tree_output_changed_fn && result.output_changed) {
params_.tree_output_changed_fn(*ntree, owner_or_self_id);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
if (needs_relations_update_) {
if (bmain_) {
DEG_relations_tag_update(bmain_);
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
private:
enum class ToposortMark {
None,
Temporary,
Permanent,
};
using ToposortMarkMap = Map<bNodeTree *, ToposortMark>;
/**
* Finds all trees that depend on the given trees (through node groups). Then those trees are
* ordered such that all trees used by one tree come before it.
*/
Vector<bNodeTree *> get_tree_update_order(Span<bNodeTree *> root_ntrees)
{
relations_.ensure_group_node_users();
Set<bNodeTree *> trees_to_update = get_trees_to_update(root_ntrees);
Vector<bNodeTree *> sorted_ntrees;
ToposortMarkMap marks;
for (bNodeTree *ntree : trees_to_update) {
marks.add_new(ntree, ToposortMark::None);
}
for (bNodeTree *ntree : trees_to_update) {
if (marks.lookup(ntree) == ToposortMark::None) {
const bool cycle_detected = !this->get_tree_update_order__visit_recursive(
ntree, marks, sorted_ntrees);
/* This should be prevented by higher level operators. */
BLI_assert(!cycle_detected);
UNUSED_VARS_NDEBUG(cycle_detected);
}
}
std::reverse(sorted_ntrees.begin(), sorted_ntrees.end());
return sorted_ntrees;
}
bool get_tree_update_order__visit_recursive(bNodeTree *ntree,
ToposortMarkMap &marks,
Vector<bNodeTree *> &sorted_ntrees)
{
ToposortMark &mark = marks.lookup(ntree);
if (mark == ToposortMark::Permanent) {
return true;
}
if (mark == ToposortMark::Temporary) {
/* There is a dependency cycle. */
return false;
}
mark = ToposortMark::Temporary;
for (const TreeNodePair &pair : relations_.get_group_node_users(ntree)) {
this->get_tree_update_order__visit_recursive(pair.first, marks, sorted_ntrees);
}
sorted_ntrees.append(ntree);
mark = ToposortMark::Permanent;
return true;
}
Set<bNodeTree *> get_trees_to_update(Span<bNodeTree *> root_ntrees)
{
relations_.ensure_group_node_users();
Set<bNodeTree *> reachable_trees;
VectorSet<bNodeTree *> trees_to_check = root_ntrees;
while (!trees_to_check.is_empty()) {
bNodeTree *ntree = trees_to_check.pop();
if (reachable_trees.add(ntree)) {
for (const TreeNodePair &pair : relations_.get_group_node_users(ntree)) {
trees_to_check.add(pair.first);
}
}
}
return reachable_trees;
}
TreeUpdateResult update_tree(bNodeTree &ntree)
{
TreeUpdateResult result;
ntree.runtime->link_errors.clear();
ntree.runtime->invalid_zone_output_node_ids.clear();
if (this->update_panel_toggle_names(ntree)) {
result.interface_changed = true;
}
this->update_socket_link_and_use(ntree);
this->update_individual_nodes(ntree);
this->update_internal_links(ntree);
this->update_generic_callback(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
this->remove_unused_previews_when_necessary(ntree);
Nodes: experimental node previews in the shader editor First implementation of node previews in the shader node editor. Using the same user interface as compositor node previews, most shader nodes can now be previewed (except group in/output and material output). This is currently still an experimental feature, as polishing of the user experience and performance improvements are planned. These will be easier to do as incremental changes on this implementation. See #110353 for details on the work that remains to be done and known limitations. Implementation notes: We take advantage of the `RenderResult` available as `ImBuf` images to store a `Render` for every viewed nested node tree present in a `SpaceNode`. The computation is initiated at the moment of drawing nodes overlays. One render is started for the current nodetree, having a `ViewLayer` associated with each previewed node. We separate the previewed nodes in two categories: the shader ones and the non-shader ones. - For non-shader nodes, we use AOVs which highly speed up the rendering process by rendering every non-shader nodes at the same time. They are rendered in the first `ViewLayer`. - For shader nodes, we render them each in a different `ViewLayer`, by rerouting the node to the output of the material in the preview scene. The preview scene takes the same aspect as the Material preview scene, and the same preview object is used. At the moment of drawing the node overlay, we take the `Render` of the viewed node tree and extract the `ImBuf` of the wanted viewlayer/pass for each previewed node. Pull Request: https://projects.blender.org/blender/blender/pulls/110065
2023-08-08 17:36:06 +02:00
this->make_node_previews_dirty(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
this->propagate_runtime_flags(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (ntree.type == NTREE_GEOMETRY) {
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
if (this->propagate_enum_definitions(ntree)) {
result.interface_changed = true;
}
if (node_field_inferencing::update_field_inferencing(ntree)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
result.interface_changed = true;
}
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
this->update_from_field_inference(ntree);
Refactor: Geometry Nodes: improve lifetime analysis for anonymous attributes This refactors the lifetime analysis of anonymous attributes in geometry nodes. The refactor has a couple of goals: * Use a better and simpler abstraction that can be used when building the lazy-function graph. We currently have a bunch of duplicate code to handle "field source" and "caller propagation" attributes. This is now unified so that one only has to worry about one kind of "reference sets". * Make the abstraction compatible with handling bundles and closures in case we want to support them in the future. Both types can contain geometries and fields so they need to be taken into account when determining lifetimes. * Make more parts independent of the concept of "anonymous attributes". In theory, there could be more kinds of referenced data whose lifetimes need to be managed. I don't have any concrete plans for adding any though. At its core, deterministic anonymous attributes still work the same they have been since they became deterministic [0]. Even the generated lazy-function graph is still pretty much or even exactly the same as before. The patch renames `AnonymousAttributeSet` to the more general `GeometryNodesReferenceSet` which is more. This also makes more places independent of the concept of anonymous attributes. Functionally, this still the same though. It's only used in the internals of geometry nodes nowadays. Most code just gets an `AttributeFilter` that is based on it. [0]: https://archive.blender.org/developer/D16858 Pull Request: https://projects.blender.org/blender/blender/pulls/128667
2024-10-07 12:59:39 +02:00
if (node_tree_reference_lifetimes::analyse_reference_lifetimes(ntree)) {
result.interface_changed = true;
}
Geometry Nodes: support attaching gizmos to input values This adds support for attaching gizmos for input values. The goal is to make it easier for users to set input values intuitively in the 3D viewport. We went through multiple different possible designs until we settled on the one implemented here. We picked it for it's flexibility and ease of use when using geometry node assets. The core principle in the design is that **gizmos are attached to existing input values instead of being the input value themselves**. This actually fits the existing concept of gizmos in Blender well, but may be a bit unintutitive in a node setup at first. The attachment is done using links in the node editor. The most basic usage of the node is to link a Value node to the new Linear Gizmo node. This attaches the gizmo to the input value and allows you to change it from the 3D view. The attachment is indicated by the gizmo icon in the sockets which are controlled by a gizmo as well as the back-link (notice the double link) when the gizmo is active. The core principle makes it straight forward to control the same node setup from the 3D view with gizmos, or by manually changing input values, or by driving the input values procedurally. If the input value is controlled indirectly by other inputs, it's often possible to **automatically propagate** the gizmo to the actual input. Backpropagation does not work for all nodes, although more nodes can be supported over time. This patch adds the first three gizmo nodes which cover common use cases: * **Linear Gizmo**: Creates a gizmo that controls a float or integer value using a linear movement of e.g. an arrow in the 3D viewport. * **Dial Gizmo**: Creates a circular gizmo in the 3D viewport that can be rotated to change the attached angle input. * **Transform Gizmo**: Creates a simple gizmo for location, rotation and scale. In the future, more built-in gizmos and potentially the ability for custom gizmos could be added. All gizmo nodes have a **Transform** geometry output. Using it is optional but it is recommended when the gizmo is used to control inputs that affect a geometry. When it is used, Blender will automatically transform the gizmos together with the geometry that they control. To achieve this, the output should be merged with the generated geometry using the *Join Geometry* node. The data contained in *Transform* output is not visible geometry, but just internal information that helps Blender to give a better user experience when using gizmos. The gizmo nodes have a multi-input socket. This allows **controlling multiple values** with the same gizmo. Only a small set of **gizmo shapes** is supported initially. It might be extended in the future but one goal is to give the gizmos used by different node group assets a familiar look and feel. A similar constraint exists for **colors**. Currently, one can choose from a fixed set of colors which can be modified in the theme settings. The set of **visible gizmos** is determined by a multiple factors because it's not really feasible to show all possible gizmos at all times. To see any of the geometry nodes gizmos, the "Active Modifier" option has to be enabled in the "Viewport Gizmos" popover. Then all gizmos are drawn for which at least one of the following is true: * The gizmo controls an input of the active modifier of the active object. * The gizmo controls a value in a selected node in an open node editor. * The gizmo controls a pinned value in an open node editor. Pinning works by clicking the gizmo icon next to the value. Pull Request: https://projects.blender.org/blender/blender/pulls/112677
2024-07-10 16:18:47 +02:00
if (nodes::gizmos::update_tree_gizmo_propagation(ntree)) {
result.interface_changed = true;
}
this->update_socket_shapes(ntree);
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
this->update_eval_dependencies(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
result.output_changed = this->check_if_output_changed(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
this->update_socket_link_and_use(ntree);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
this->update_link_validation(ntree);
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
if (this->update_nested_node_refs(ntree)) {
result.interface_changed = true;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (ntree.type == NTREE_TEXTURE) {
ntreeTexCheckCyclics(&ntree);
}
if (ntree.tree_interface.is_changed()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
result.interface_changed = true;
}
#ifndef NDEBUG
/* Check the uniqueness of node identifiers. */
Set<int32_t> node_identifiers;
const Span<const bNode *> nodes = ntree.all_nodes();
for (const int i : nodes.index_range()) {
const bNode &node = *nodes[i];
BLI_assert(node.identifier > 0);
node_identifiers.add_new(node.identifier);
BLI_assert(node.runtime->index_in_tree == i);
}
#endif
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return result;
}
void update_socket_link_and_use(bNodeTree &tree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
tree.ensure_topology_cache();
for (bNodeSocket *socket : tree.all_input_sockets()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (socket->directly_linked_links().is_empty()) {
socket->link = nullptr;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
else {
socket->link = socket->directly_linked_links()[0];
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
this->update_socket_used_tags(tree);
}
void update_socket_used_tags(bNodeTree &tree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
tree.ensure_topology_cache();
for (bNodeSocket *socket : tree.all_sockets()) {
const bool socket_is_linked = !socket->directly_linked_links().is_empty();
SET_FLAG_FROM_TEST(socket->flag, socket_is_linked, SOCK_IS_LINKED);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
void update_individual_nodes(bNodeTree &ntree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
for (bNode *node : ntree.all_nodes()) {
bke::node_declaration_ensure(ntree, *node);
if (this->should_update_individual_node(ntree, *node)) {
bke::bNodeType &ntype = *node->typeinfo;
Nodes: unify static and dynamic declarations This helps solving the problem encountered in #113553. The problem is that we currently can't support link-drag-search for nodes which have a dynamic declaration. With this patch, there is only a single `declare` function per node type, instead of the separate `declare` and `declare_dynamic` functions. The new `declare` function has access to the node and tree. However, both are allowed to be null. The final node declaration has a flag for whether it depends on the node context or not. Nodes that previously had a dynamic declaration should now create as much of the declaration as possible that does not depend on the node. This allows code like for link-drag-search to take those sockets into account even if the other sockets are dynamic. For node declarations that have dynamic types (e.g. Switch node), we can also add extra information to the static node declaration, like the identifier of the socket with the dynamic type. This is not part of this patch though. I can think of two main alternatives to the approach implemented here: * Define two separate functions for dynamic nodes. One that creates the "static declaration" without node context, and on that creates the actual declaration with node context. * Have a single declare function that generates "build instructions" for the actual node declaration. So instead of building the final declaration directly, one can for example add a socket whose type depends on a specific rna path in the node. The actual node declaration is then automatically generated based on the build instructions. This becomes quite a bit more tricky with dynamic amounts of sockets and introduces another indirection between declarations and what sockets the node actually has. I found the approach implemented in this patch to lead to the least amount of boilerplate (doesn't require a seperate "build instructions" data structure) and code duplication (socket properties are still only defined in one place). At the same time, it offers more flexibility to how nodes can be dynamic. Pull Request: https://projects.blender.org/blender/blender/pulls/113742
2023-10-15 20:28:23 +02:00
if (ntype.declare) {
/* Should have been created when the node was registered. */
BLI_assert(ntype.static_declaration != nullptr);
if (ntype.static_declaration->is_context_dependent) {
nodes::update_node_declaration_and_sockets(ntree, *node);
}
Nodes: Use dynamic declarations for group nodes Since a year and a half ago we've been switching to a new way to represent what sockets a node should have called "declarations" that's easier to use, clearer, and more flexible for upcoming features like dynamic socket counts or generic type sockets. All builtin nodes with a static set of sockets have switched, but one missing area has been group nodes and group input/output nodes. These nodes have **dynamic** declarations which change based on their properties or the group they're inside of. This patch addresses that, in preparation for using the same dynamic declaration feature for simulation nodes. Generally there shouldn't be user-visible differences, but one benefit is that user-created socket descriptions are now visible directly in the node editor for group nodes and group input/output nodes. The commit contains a few changes: - Add a node type callback for building dynamic declarations with different arguments - Add an `Extend` socket declaration for the "virtual" sockets used for connecting new links - A similar `Custom` socket declaration is used for addon-defined socket - Simplify the node update loop to use the declaration to build update sockets - Replace the "group update" functions with the declaration building - Move the node group input/output link creation to link drag operator - Make the field status part of group node declarations (not for group input/output nodes though) - Some fixes for declarations to make them update and build properly Differential Revision: https://developer.blender.org/D16850
2023-01-16 15:47:10 -06:00
}
2025-02-11 17:25:10 +01:00
else if (node->is_undefined()) {
/* If a node has become undefined (it generally was unregistered from Python), it does
* not have a declaration anymore. */
delete node->runtime->declaration;
node->runtime->declaration = nullptr;
LISTBASE_FOREACH (bNodeSocket *, socket, &node->inputs) {
socket->runtime->declaration = nullptr;
}
LISTBASE_FOREACH (bNodeSocket *, socket, &node->outputs) {
socket->runtime->declaration = nullptr;
}
2025-02-11 17:25:10 +01:00
}
if (ntype.updatefunc) {
ntype.updatefunc(&ntree, node);
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
bool should_update_individual_node(const bNodeTree &ntree, const bNode &node)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
if (ntree.runtime->changed_flag & NTREE_CHANGED_ANY) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
if (node.runtime->changed_flag & NTREE_CHANGED_NODE_PROPERTY) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
if (ntree.runtime->changed_flag & NTREE_CHANGED_LINK) {
Nodes: Use dynamic declarations for group nodes Since a year and a half ago we've been switching to a new way to represent what sockets a node should have called "declarations" that's easier to use, clearer, and more flexible for upcoming features like dynamic socket counts or generic type sockets. All builtin nodes with a static set of sockets have switched, but one missing area has been group nodes and group input/output nodes. These nodes have **dynamic** declarations which change based on their properties or the group they're inside of. This patch addresses that, in preparation for using the same dynamic declaration feature for simulation nodes. Generally there shouldn't be user-visible differences, but one benefit is that user-created socket descriptions are now visible directly in the node editor for group nodes and group input/output nodes. The commit contains a few changes: - Add a node type callback for building dynamic declarations with different arguments - Add an `Extend` socket declaration for the "virtual" sockets used for connecting new links - A similar `Custom` socket declaration is used for addon-defined socket - Simplify the node update loop to use the declaration to build update sockets - Replace the "group update" functions with the declaration building - Move the node group input/output link creation to link drag operator - Make the field status part of group node declarations (not for group input/output nodes though) - Some fixes for declarations to make them update and build properly Differential Revision: https://developer.blender.org/D16850
2023-01-16 15:47:10 -06:00
/* Currently we have no way to tell if a node needs to be updated when a link changed. */
return true;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
if (ntree.tree_interface.is_changed()) {
if (node.is_group_input() || node.is_group_output()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
Geometry Nodes: add simulation support This adds support for building simulations with geometry nodes. A new `Simulation Input` and `Simulation Output` node allow maintaining a simulation state across multiple frames. Together these two nodes form a `simulation zone` which contains all the nodes that update the simulation state from one frame to the next. A new simulation zone can be added via the menu (`Simulation > Simulation Zone`) or with the node add search. The simulation state contains a geometry by default. However, it is possible to add multiple geometry sockets as well as other socket types. Currently, field inputs are evaluated and stored for the preceding geometry socket in the order that the sockets are shown. Simulation state items can be added by linking one of the empty sockets to something else. In the sidebar, there is a new panel that allows adding, removing and reordering these sockets. The simulation nodes behave as follows: * On the first frame, the inputs of the `Simulation Input` node are evaluated to initialize the simulation state. In later frames these sockets are not evaluated anymore. The `Delta Time` at the first frame is zero, but the simulation zone is still evaluated. * On every next frame, the `Simulation Input` node outputs the simulation state of the previous frame. Nodes in the simulation zone can edit that data in arbitrary ways, also taking into account the `Delta Time`. The new simulation state has to be passed to the `Simulation Output` node where it is cached and forwarded. * On a frame that is already cached or baked, the nodes in the simulation zone are not evaluated, because the `Simulation Output` node can return the previously cached data directly. It is not allowed to connect sockets from inside the simulation zone to the outside without going through the `Simulation Output` node. This is a necessary restriction to make caching and sub-frame interpolation work. Links can go into the simulation zone without problems though. Anonymous attributes are not propagated by the simulation nodes unless they are explicitly stored in the simulation state. This is unfortunate, but currently there is no practical and reliable alternative. The core problem is detecting which anonymous attributes will be required for the simulation and afterwards. While we can detect this for the current evaluation, we can't look into the future in time to see what data will be necessary. We intend to make it easier to explicitly pass data through a simulation in the future, even if the simulation is in a nested node group. There is a new `Simulation Nodes` panel in the physics tab in the properties editor. It allows baking all simulation zones on the selected objects. The baking options are intentially kept at a minimum for this MVP. More features for simulation baking as well as baking in general can be expected to be added separately. All baked data is stored on disk in a folder next to the .blend file. #106937 describes how baking is implemented in more detail. Volumes can not be baked yet and materials are lost during baking for now. Packing the baked data into the .blend file is not yet supported. The timeline indicates which frames are currently cached, baked or cached but invalidated by user-changes. Simulation input and output nodes are internally linked together by their `bNode.identifier` which stays the same even if the node name changes. They are generally added and removed together. However, there are still cases where "dangling" simulation nodes can be created currently. Those generally don't cause harm, but would be nice to avoid this in more cases in the future. Co-authored-by: Hans Goudey <h.goudey@me.com> Co-authored-by: Lukas Tönne <lukas@blender.org> Pull Request: https://projects.blender.org/blender/blender/pulls/104924
2023-05-03 13:18:51 +02:00
}
/* Check paired simulation zone nodes. */
if (all_zone_input_node_types().contains(node.type_legacy)) {
const bNodeZoneType &zone_type = *zone_type_by_node_type(node.type_legacy);
if (const bNode *output_node = zone_type.get_corresponding_output(ntree, node)) {
Geometry Nodes: new Repeat Zone This adds support for running a set of nodes repeatedly. The number of iterations can be controlled dynamically as an input of the repeat zone. The repeat zone can be added in via the search or from the Add > Utilities menu. The main use case is to replace long repetitive node chains with a more flexible alternative. Technically, repeat zones can also be used for many other use cases. However, due to their serial nature, performance is very sub-optimal when they are used to solve problems that could be processed in parallel. Better solutions for such use cases will be worked on separately. Repeat zones are similar to simulation zones. The major difference is that they have no concept of time and are always evaluated entirely in the current frame, while in simulations only a single iteration is evaluated per frame. Stopping the repetition early using a dynamic condition is not yet supported. "Break" functionality can be implemented manually using Switch nodes in the loop for now. It's likely that this functionality will be built into the repeat zone in the future. For now, things are kept more simple. Remaining Todos after this first version: * Improve socket inspection and viewer node support. Currently, only the first iteration is taken into account for socket inspection and the viewer. * Make loop evaluation more lazy. Currently, the evaluation is eager, meaning that it evaluates some nodes even though their output may not be required. Pull Request: https://projects.blender.org/blender/blender/pulls/109164
2023-07-11 22:36:10 +02:00
if (output_node->runtime->changed_flag & NTREE_CHANGED_NODE_PROPERTY) {
return true;
}
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return false;
}
struct InternalLink {
bNodeSocket *from;
bNodeSocket *to;
int multi_input_sort_id = 0;
BLI_STRUCT_EQUALITY_OPERATORS_3(InternalLink, from, to, multi_input_sort_id);
};
const bNodeLink *first_non_dangling_link(const bNodeTree & /*ntree*/,
const Span<const bNodeLink *> links) const
{
for (const bNodeLink *link : links) {
if (!link->fromnode->is_dangling_reroute()) {
return link;
}
}
return nullptr;
}
void update_internal_links(bNodeTree &ntree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
bke::node_tree_runtime::AllowUsingOutdatedInfo allow_outdated_info{ntree};
ntree.ensure_topology_cache();
for (bNode *node : ntree.all_nodes()) {
if (!this->should_update_individual_node(ntree, *node)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
continue;
}
/* Find all expected internal links. */
Vector<InternalLink> expected_internal_links;
for (const bNodeSocket *output_socket : node->output_sockets()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (!output_socket->is_available()) {
continue;
}
if (output_socket->flag & SOCK_NO_INTERNAL_LINK) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
continue;
}
const bNodeSocket *input_socket = this->find_internally_linked_input(ntree, output_socket);
if (input_socket == nullptr) {
continue;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
const Span<const bNodeLink *> connected_links = input_socket->directly_linked_links();
const bNodeLink *connected_link = first_non_dangling_link(ntree, connected_links);
const int index = connected_link ? connected_link->multi_input_sort_id :
std::max<int>(0, connected_links.size() - 1);
expected_internal_links.append(InternalLink{const_cast<bNodeSocket *>(input_socket),
const_cast<bNodeSocket *>(output_socket),
index});
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
/* Rebuilt internal links if they have changed. */
if (node->runtime->internal_links.size() != expected_internal_links.size()) {
this->update_internal_links_in_node(ntree, *node, expected_internal_links);
continue;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
const bool all_expected_internal_links_exist = std::all_of(
node->runtime->internal_links.begin(),
node->runtime->internal_links.end(),
[&](const bNodeLink &link) {
const InternalLink internal_link{link.fromsock, link.tosock, link.multi_input_sort_id};
return expected_internal_links.as_span().contains(internal_link);
});
if (all_expected_internal_links_exist) {
continue;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
this->update_internal_links_in_node(ntree, *node, expected_internal_links);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
const bNodeSocket *find_internally_linked_input(const bNodeTree &ntree,
const bNodeSocket *output_socket)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
const bNode &node = output_socket->owner_node();
if (node.typeinfo->internally_linked_input) {
return node.typeinfo->internally_linked_input(ntree, node, *output_socket);
}
const bNodeSocket *selected_socket = nullptr;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
int selected_priority = -1;
bool selected_is_linked = false;
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
for (const bNodeSocket *input_socket : node.input_sockets()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (!input_socket->is_available()) {
continue;
}
if (input_socket->flag & SOCK_NO_INTERNAL_LINK) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
continue;
}
const int priority = get_internal_link_type_priority(input_socket->typeinfo,
output_socket->typeinfo);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (priority < 0) {
continue;
}
const bool is_linked = input_socket->is_directly_linked();
const bool is_preferred = priority > selected_priority || (is_linked && !selected_is_linked);
if (!is_preferred) {
continue;
}
selected_socket = input_socket;
selected_priority = priority;
selected_is_linked = is_linked;
}
return selected_socket;
}
void update_internal_links_in_node(bNodeTree &ntree,
bNode &node,
Span<InternalLink> internal_links)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
node.runtime->internal_links.clear();
node.runtime->internal_links.reserve(internal_links.size());
for (const InternalLink &internal_link : internal_links) {
bNodeLink link{};
link.fromnode = &node;
link.fromsock = internal_link.from;
link.tonode = &node;
link.tosock = internal_link.to;
link.multi_input_sort_id = internal_link.multi_input_sort_id;
link.flag |= NODE_LINK_VALID;
node.runtime->internal_links.append(link);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
BKE_ntree_update_tag_node_internal_link(&ntree, &node);
}
void update_generic_callback(bNodeTree &ntree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
if (ntree.typeinfo->update == nullptr) {
return;
}
ntree.typeinfo->update(&ntree);
}
void remove_unused_previews_when_necessary(bNodeTree &ntree)
{
/* Don't trigger preview removal when only those flags are set. */
const uint32_t allowed_flags = NTREE_CHANGED_LINK | NTREE_CHANGED_SOCKET_PROPERTY |
NTREE_CHANGED_NODE_PROPERTY | NTREE_CHANGED_NODE_OUTPUT;
if ((ntree.runtime->changed_flag & allowed_flags) == ntree.runtime->changed_flag) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return;
}
blender::bke::node_preview_remove_unused(&ntree);
Nodes: experimental node previews in the shader editor First implementation of node previews in the shader node editor. Using the same user interface as compositor node previews, most shader nodes can now be previewed (except group in/output and material output). This is currently still an experimental feature, as polishing of the user experience and performance improvements are planned. These will be easier to do as incremental changes on this implementation. See #110353 for details on the work that remains to be done and known limitations. Implementation notes: We take advantage of the `RenderResult` available as `ImBuf` images to store a `Render` for every viewed nested node tree present in a `SpaceNode`. The computation is initiated at the moment of drawing nodes overlays. One render is started for the current nodetree, having a `ViewLayer` associated with each previewed node. We separate the previewed nodes in two categories: the shader ones and the non-shader ones. - For non-shader nodes, we use AOVs which highly speed up the rendering process by rendering every non-shader nodes at the same time. They are rendered in the first `ViewLayer`. - For shader nodes, we render them each in a different `ViewLayer`, by rerouting the node to the output of the material in the preview scene. The preview scene takes the same aspect as the Material preview scene, and the same preview object is used. At the moment of drawing the node overlay, we take the `Render` of the viewed node tree and extract the `ImBuf` of the wanted viewlayer/pass for each previewed node. Pull Request: https://projects.blender.org/blender/blender/pulls/110065
2023-08-08 17:36:06 +02:00
}
void make_node_previews_dirty(bNodeTree &ntree)
{
ntree.runtime->previews_refresh_state++;
for (bNode *node : ntree.all_nodes()) {
if (!node->is_group()) {
Nodes: experimental node previews in the shader editor First implementation of node previews in the shader node editor. Using the same user interface as compositor node previews, most shader nodes can now be previewed (except group in/output and material output). This is currently still an experimental feature, as polishing of the user experience and performance improvements are planned. These will be easier to do as incremental changes on this implementation. See #110353 for details on the work that remains to be done and known limitations. Implementation notes: We take advantage of the `RenderResult` available as `ImBuf` images to store a `Render` for every viewed nested node tree present in a `SpaceNode`. The computation is initiated at the moment of drawing nodes overlays. One render is started for the current nodetree, having a `ViewLayer` associated with each previewed node. We separate the previewed nodes in two categories: the shader ones and the non-shader ones. - For non-shader nodes, we use AOVs which highly speed up the rendering process by rendering every non-shader nodes at the same time. They are rendered in the first `ViewLayer`. - For shader nodes, we render them each in a different `ViewLayer`, by rerouting the node to the output of the material in the preview scene. The preview scene takes the same aspect as the Material preview scene, and the same preview object is used. At the moment of drawing the node overlay, we take the `Render` of the viewed node tree and extract the `ImBuf` of the wanted viewlayer/pass for each previewed node. Pull Request: https://projects.blender.org/blender/blender/pulls/110065
2023-08-08 17:36:06 +02:00
continue;
}
if (bNodeTree *nested_tree = reinterpret_cast<bNodeTree *>(node->id)) {
this->make_node_previews_dirty(*nested_tree);
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
void propagate_runtime_flags(const bNodeTree &ntree)
{
ntree.ensure_topology_cache();
ntree.runtime->runtime_flag = 0;
for (const bNode *group_node : ntree.group_nodes()) {
const bNodeTree *group = reinterpret_cast<bNodeTree *>(group_node->id);
if (group != nullptr) {
ntree.runtime->runtime_flag |= group->runtime->runtime_flag;
}
}
if (ntree.type == NTREE_SHADER) {
/* Check if the tree itself has an animated image. */
for (const StringRefNull idname : {"ShaderNodeTexImage", "ShaderNodeTexEnvironment"}) {
for (const bNode *node : ntree.nodes_by_type(idname)) {
Image *image = reinterpret_cast<Image *>(node->id);
if (image != nullptr && BKE_image_is_animated(image)) {
ntree.runtime->runtime_flag |= NTREE_RUNTIME_FLAG_HAS_IMAGE_ANIMATION;
break;
}
}
}
/* Check if the tree has a material output. */
for (const StringRefNull idname : {"ShaderNodeOutputMaterial",
"ShaderNodeOutputLight",
"ShaderNodeOutputWorld",
"ShaderNodeOutputAOV"})
{
const Span<const bNode *> nodes = ntree.nodes_by_type(idname);
if (!nodes.is_empty()) {
ntree.runtime->runtime_flag |= NTREE_RUNTIME_FLAG_HAS_MATERIAL_OUTPUT;
break;
}
}
}
if (ntree.type == NTREE_GEOMETRY) {
/* Check if there is a simulation zone. */
if (!ntree.nodes_by_type("GeometryNodeSimulationOutput").is_empty()) {
ntree.runtime->runtime_flag |= NTREE_RUNTIME_FLAG_HAS_SIMULATION_ZONE;
}
}
}
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
void update_from_field_inference(bNodeTree &ntree)
{
/* Automatically tag a bake item as attribute when the input is a field. The flag should not be
* removed automatically even when the field input is disconnected because the baked data may
* still contain attribute data instead of a single value. */
const Span<bke::FieldSocketState> field_states = ntree.runtime->field_states;
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
for (bNode *node : ntree.nodes_by_type("GeometryNodeBake")) {
NodeGeometryBake &storage = *static_cast<NodeGeometryBake *>(node->storage);
for (const int i : IndexRange(storage.items_num)) {
const bNodeSocket &socket = node->input_socket(i);
NodeGeometryBakeItem &item = storage.items[i];
if (field_states[socket.index_in_tree()] == FieldSocketState::IsField) {
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
item.flag |= GEO_NODE_BAKE_ITEM_IS_ATTRIBUTE;
}
}
}
}
void update_socket_shapes(bNodeTree &ntree)
{
ntree.ensure_topology_cache();
const Span<bke::FieldSocketState> field_states = ntree.runtime->field_states;
for (bNodeSocket *socket : ntree.all_sockets()) {
switch (field_states[socket->index_in_tree()]) {
case bke::FieldSocketState::RequiresSingle:
socket->display_shape = SOCK_DISPLAY_SHAPE_CIRCLE;
break;
case bke::FieldSocketState::CanBeField:
socket->display_shape = SOCK_DISPLAY_SHAPE_DIAMOND_DOT;
break;
case bke::FieldSocketState::IsField:
socket->display_shape = SOCK_DISPLAY_SHAPE_DIAMOND;
break;
}
}
}
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
void update_eval_dependencies(bNodeTree &ntree)
{
ntree.ensure_topology_cache();
nodes::GeometryNodesEvalDependencies new_deps =
nodes::gather_geometry_nodes_eval_dependencies_with_cache(ntree);
Geometry Nodes: improve detecting data-block dependencies Previously, the data-block dependencies were always detected in `update_depsgraph` in `MOD_nodes.cc`. This would only be called when something called `DEG_relations_tag_update` before. We don't want to trigger a depsgraph rebuild after each operation in the node editor, as that would be expensive. However, that also meant that we often had to add data-block dependencies that are not actually used, but might be used if the user changed e.g. a link. A typical example for that is a object socket that has a default value, but the socket is also linked. Now, the dependencies referenced by the node tree are collected by the node tree update code which runs after all changes. This way we can detect whether the dependencies have changed. Only if they have changed, a depsgraph rebuild is triggered. This now allows also taking into account the mute status of nodes and whether an input is linked. There are still more things that could be taken into account. Most obviously whether a node is connected to an output. This can be done later. The most tricky aspect here is probably that we also have to consider all viewer nodes as output, because at the time the node runs, it's not known which viewer will actually be used (which depends on other editors). This also cleans up some special cases we had for e.g. the scene time node where we always had to trigger a depsgraph rebuild when it was added/removed because of its time dependence. This is now part of a more general system. This fixes #109219. Pull Request: https://projects.blender.org/blender/blender/pulls/131446
2024-12-05 18:02:14 +01:00
/* Check if the dependencies have changed. */
if (!ntree.runtime->geometry_nodes_eval_dependencies ||
new_deps != *ntree.runtime->geometry_nodes_eval_dependencies)
{
needs_relations_update_ = true;
ntree.runtime->geometry_nodes_eval_dependencies =
std::make_unique<nodes::GeometryNodesEvalDependencies>(std::move(new_deps));
}
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
bool propagate_enum_definitions(bNodeTree &ntree)
{
ntree.ensure_interface_cache();
/* Propagation from right to left to determine which enum
* definition to use for menu sockets. */
for (bNode *node : ntree.toposort_right_to_left()) {
const bool node_updated = this->should_update_individual_node(ntree, *node);
Vector<bNodeSocket *> locally_defined_enums;
if (node->is_type("GeometryNodeMenuSwitch")) {
bNodeSocket &enum_input = node->input_socket(0);
BLI_assert(enum_input.is_available() && enum_input.type == SOCK_MENU);
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* Generate new enum items when the node has changed, otherwise keep existing items. */
if (node_updated) {
const NodeMenuSwitch &storage = *static_cast<NodeMenuSwitch *>(node->storage);
const RuntimeNodeEnumItems *enum_items = this->create_runtime_enum_items(
storage.enum_definition);
this->set_enum_ptr(*enum_input.default_value_typed<bNodeSocketValueMenu>(), enum_items);
/* Remove initial user. */
enum_items->remove_user_and_delete_if_last();
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
locally_defined_enums.append(&enum_input);
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
/* Clear current enum references. */
for (bNodeSocket *socket : node->input_sockets()) {
if (socket->is_available() && socket->type == SOCK_MENU &&
!locally_defined_enums.contains(socket))
{
clear_enum_reference(*socket);
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
}
for (bNodeSocket *socket : node->output_sockets()) {
if (socket->is_available() && socket->type == SOCK_MENU) {
clear_enum_reference(*socket);
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
}
/* Propagate enum references from output links. */
for (bNodeSocket *output : node->output_sockets()) {
if (!output->is_available() || output->type != SOCK_MENU) {
continue;
}
for (const bNodeSocket *input : output->directly_linked_sockets()) {
if (!input->is_available() || input->type != SOCK_MENU) {
continue;
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
this->update_socket_enum_definition(*output->default_value_typed<bNodeSocketValueMenu>(),
*input->default_value_typed<bNodeSocketValueMenu>());
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
}
if (node->is_group()) {
/* Node groups expose internal enum definitions. */
if (node->id == nullptr) {
continue;
}
const bNodeTree *group_tree = reinterpret_cast<bNodeTree *>(node->id);
group_tree->ensure_interface_cache();
for (const int socket_i : group_tree->interface_inputs().index_range()) {
bNodeSocket &input = *node->input_sockets()[socket_i];
const bNodeTreeInterfaceSocket &iosocket = *group_tree->interface_inputs()[socket_i];
BLI_assert(STREQ(input.identifier, iosocket.identifier));
if (input.is_available() && input.type == SOCK_MENU) {
BLI_assert(STREQ(iosocket.socket_type, "NodeSocketMenu"));
this->update_socket_enum_definition(
*input.default_value_typed<bNodeSocketValueMenu>(),
*static_cast<bNodeSocketValueMenu *>(iosocket.socket_data));
}
}
}
else if (node->is_type("GeometryNodeMenuSwitch")) {
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* First input is always the node's own menu, propagate only to the enum case inputs. */
const bNodeSocket *output = node->output_sockets().first();
for (bNodeSocket *input : node->input_sockets().drop_front(1)) {
if (input->is_available() && input->type == SOCK_MENU) {
this->update_socket_enum_definition(
*input->default_value_typed<bNodeSocketValueMenu>(),
*output->default_value_typed<bNodeSocketValueMenu>());
}
}
}
else if (node->is_type("GeometryNodeForeachGeometryElementInput")) {
/* Propagate menu from element inputs to field inputs. */
BLI_assert(node->input_sockets().size() == node->output_sockets().size());
/* Inputs Geometry, Selection and outputs Index, Element are ignored. */
const IndexRange sockets = node->input_sockets().index_range().drop_front(2);
for (const int socket_i : sockets) {
bNodeSocket *input = node->input_sockets()[socket_i];
bNodeSocket *output = node->output_sockets()[socket_i];
if (input->is_available() && input->type == SOCK_MENU && output->is_available() &&
output->type == SOCK_MENU)
{
this->update_socket_enum_definition(
*input->default_value_typed<bNodeSocketValueMenu>(),
*output->default_value_typed<bNodeSocketValueMenu>());
}
}
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
else {
/* Propagate over internal relations. */
/* XXX Placeholder implementation just propagates all outputs
* to all inputs for built-in nodes This could perhaps use
* input/output relations to handle propagation generically? */
for (bNodeSocket *input : node->input_sockets()) {
if (input->is_available() && input->type == SOCK_MENU) {
for (const bNodeSocket *output : node->output_sockets()) {
if (output->is_available() && output->type == SOCK_MENU) {
this->update_socket_enum_definition(
*input->default_value_typed<bNodeSocketValueMenu>(),
*output->default_value_typed<bNodeSocketValueMenu>());
}
}
}
}
}
}
/* Find conflicts between on corresponding menu sockets on different group input nodes. */
const Span<bNode *> group_input_nodes = ntree.group_input_nodes();
for (const int interface_input_i : ntree.interface_inputs().index_range()) {
const bNodeTreeInterfaceSocket &interface_socket =
*ntree.interface_inputs()[interface_input_i];
if (interface_socket.socket_type != StringRef("NodeSocketMenu")) {
continue;
}
const RuntimeNodeEnumItems *found_enum_items = nullptr;
bool found_conflict = false;
for (bNode *input_node : group_input_nodes) {
const bNodeSocket &socket = input_node->output_socket(interface_input_i);
const auto &socket_value = *socket.default_value_typed<bNodeSocketValueMenu>();
if (socket_value.has_conflict()) {
found_conflict = true;
break;
}
if (found_enum_items == nullptr) {
found_enum_items = socket_value.enum_items;
}
else if (socket_value.enum_items != nullptr) {
if (found_enum_items != socket_value.enum_items) {
found_conflict = true;
break;
}
}
}
if (found_conflict) {
/* Make sure that all group input sockets know that there is a socket. */
for (bNode *input_node : group_input_nodes) {
bNodeSocket &socket = input_node->output_socket(interface_input_i);
auto &socket_value = *socket.default_value_typed<bNodeSocketValueMenu>();
if (socket_value.enum_items) {
socket_value.enum_items->remove_user_and_delete_if_last();
socket_value.enum_items = nullptr;
}
socket_value.runtime_flag |= NodeSocketValueMenuRuntimeFlag::NODE_MENU_ITEMS_CONFLICT;
}
}
else if (found_enum_items != nullptr) {
/* Make sure all corresponding menu sockets have the same menu reference. */
for (bNode *input_node : group_input_nodes) {
bNodeSocket &socket = input_node->output_socket(interface_input_i);
auto &socket_value = *socket.default_value_typed<bNodeSocketValueMenu>();
if (socket_value.enum_items == nullptr) {
found_enum_items->add_user();
socket_value.enum_items = found_enum_items;
}
}
}
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* Build list of new enum items for the node tree interface. */
Vector<bNodeSocketValueMenu> interface_enum_items(ntree.interface_inputs().size(), {0});
for (const bNode *group_input_node : ntree.group_input_nodes()) {
for (const int socket_i : ntree.interface_inputs().index_range()) {
const bNodeSocket &output = *group_input_node->output_sockets()[socket_i];
if (output.is_available() && output.type == SOCK_MENU) {
this->update_socket_enum_definition(interface_enum_items[socket_i],
*output.default_value_typed<bNodeSocketValueMenu>());
}
}
}
/* Move enum items to the interface and detect if anything changed. */
bool changed = false;
for (const int socket_i : ntree.interface_inputs().index_range()) {
bNodeTreeInterfaceSocket &iosocket = *ntree.interface_inputs()[socket_i];
if (STREQ(iosocket.socket_type, "NodeSocketMenu")) {
bNodeSocketValueMenu &dst = *static_cast<bNodeSocketValueMenu *>(iosocket.socket_data);
const bNodeSocketValueMenu &src = interface_enum_items[socket_i];
if (dst.enum_items != src.enum_items || dst.has_conflict() != src.has_conflict()) {
changed = true;
if (dst.enum_items) {
dst.enum_items->remove_user_and_delete_if_last();
}
/* Items are moved, no need to change user count. */
dst.enum_items = src.enum_items;
SET_FLAG_FROM_TEST(dst.runtime_flag, src.has_conflict(), NODE_MENU_ITEMS_CONFLICT);
}
else {
/* If the item isn't move make sure it gets released again. */
if (src.enum_items) {
src.enum_items->remove_user_and_delete_if_last();
}
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
}
}
return changed;
}
/**
* Make a runtime copy of the DNA enum items.
* The runtime items list is shared by sockets.
*/
const RuntimeNodeEnumItems *create_runtime_enum_items(const NodeEnumDefinition &enum_def)
{
RuntimeNodeEnumItems *enum_items = new RuntimeNodeEnumItems();
enum_items->items.reinitialize(enum_def.items_num);
for (const int i : enum_def.items().index_range()) {
const NodeEnumItem &src = enum_def.items()[i];
RuntimeNodeEnumItem &dst = enum_items->items[i];
dst.identifier = src.identifier;
dst.name = src.name ? src.name : "";
dst.description = src.description ? src.description : "";
}
return enum_items;
}
void clear_enum_reference(bNodeSocket &socket)
{
BLI_assert(socket.is_available() && socket.type == SOCK_MENU);
bNodeSocketValueMenu &default_value = *socket.default_value_typed<bNodeSocketValueMenu>();
this->reset_enum_ptr(default_value);
default_value.runtime_flag &= ~NODE_MENU_ITEMS_CONFLICT;
}
void update_socket_enum_definition(bNodeSocketValueMenu &dst, const bNodeSocketValueMenu &src)
{
if (dst.has_conflict()) {
/* Target enum already has a conflict. */
BLI_assert(dst.enum_items == nullptr);
return;
}
if (src.has_conflict()) {
/* Target conflict if any source enum has a conflict. */
this->reset_enum_ptr(dst);
dst.runtime_flag |= NODE_MENU_ITEMS_CONFLICT;
}
else if (!dst.enum_items) {
/* First connection, set the reference. */
this->set_enum_ptr(dst, src.enum_items);
}
else if (src.enum_items && dst.enum_items != src.enum_items) {
/* Error if enum ref does not match other connections. */
this->reset_enum_ptr(dst);
dst.runtime_flag |= NODE_MENU_ITEMS_CONFLICT;
}
}
void reset_enum_ptr(bNodeSocketValueMenu &dst)
{
if (dst.enum_items) {
dst.enum_items->remove_user_and_delete_if_last();
dst.enum_items = nullptr;
}
}
void set_enum_ptr(bNodeSocketValueMenu &dst, const RuntimeNodeEnumItems *enum_items)
{
if (dst.enum_items) {
dst.enum_items->remove_user_and_delete_if_last();
dst.enum_items = nullptr;
}
if (enum_items) {
enum_items->add_user();
dst.enum_items = enum_items;
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void update_link_validation(bNodeTree &ntree)
{
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* Tests if enum references are undefined. */
const auto is_invalid_enum_ref = [](const bNodeSocket &socket) -> bool {
if (socket.type == SOCK_MENU) {
return socket.default_value_typed<bNodeSocketValueMenu>()->enum_items == nullptr;
}
return false;
};
const bNodeTreeZones *fallback_zones = nullptr;
if (ntree.type == NTREE_GEOMETRY && !ntree.zones() && ntree.runtime->last_valid_zones) {
fallback_zones = ntree.runtime->last_valid_zones.get();
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
LISTBASE_FOREACH (bNodeLink *, link, &ntree.links) {
link->flag |= NODE_LINK_VALID;
if (!link->fromsock->is_available() || !link->tosock->is_available()) {
link->flag &= ~NODE_LINK_VALID;
continue;
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
if (is_invalid_enum_ref(*link->fromsock) || is_invalid_enum_ref(*link->tosock)) {
link->flag &= ~NODE_LINK_VALID;
ntree.runtime->link_errors.add(
NodeLinkKey{*link},
NodeLinkError{TIP_("Use node groups to reuse the same menu multiple times")});
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
continue;
}
if (ntree.type == NTREE_GEOMETRY) {
const Span<FieldSocketState> field_states = ntree.runtime->field_states;
if (field_states[link->fromsock->index_in_tree()] == FieldSocketState::IsField &&
field_states[link->tosock->index_in_tree()] != FieldSocketState::IsField)
{
link->flag &= ~NODE_LINK_VALID;
ntree.runtime->link_errors.add(
NodeLinkKey{*link}, NodeLinkError{TIP_("The node input does not support fields")});
continue;
}
}
const bNode &from_node = *link->fromnode;
const bNode &to_node = *link->tonode;
if (from_node.runtime->toposort_left_to_right_index >
to_node.runtime->toposort_left_to_right_index)
{
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
link->flag &= ~NODE_LINK_VALID;
ntree.runtime->link_errors.add(
NodeLinkKey{*link},
NodeLinkError{TIP_("The links form a cycle which is not supported")});
continue;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
if (ntree.typeinfo->validate_link) {
const eNodeSocketDatatype from_type = eNodeSocketDatatype(link->fromsock->type);
const eNodeSocketDatatype to_type = eNodeSocketDatatype(link->tosock->type);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (!ntree.typeinfo->validate_link(from_type, to_type)) {
link->flag &= ~NODE_LINK_VALID;
ntree.runtime->link_errors.add(
NodeLinkKey{*link},
I18n: extract and disambiguate a few messages Extract - Add to Quick Favorites tooltip. - "Mask", the name of a newly created mask (DATA_). - "New" in the context of the new mask ID button. - A few strings using BLI_STR_UTF8_ defines were not extracted. Take the special characters out of the translation macros. - "External" menu items from the filebrowser's Files context menu (right-click on a file). These items were already extracted, but not translated. Improve - Separate formatted error message "%s is not compatible with ["the specified", "any"] 'refresh' options" into two messages. Disambiguate - Use Action context for new F-modifiers' names. This is already used for the "type" operator prop. - Translate ImportHelper's default confirmation text using the Operator context, as it uses the operator name which is extracted with this context. - "Scale" can be a noun, the scale of something, or a verb, to scale something. The latter mostly uses the Operator context, so apply this context to verbs, and the default contexts to nouns. - "Scale Influence" can mean "Influence on Scale" (tracking stabilization) and "to Scale the Influence" (dynamic paint canvas). - "Object Line Art" as type of Line Art to add, as opposed to the active object's Line Art settings. - Float to Integer node: use NodeTree context for the node label, as this is already extracted and used for the enum. Do not translate - Sequencer labels containing only a string formatting field. Some issues reported by Gabriel Gazzán and Ye Gui. Pull Request: https://projects.blender.org/blender/blender/pulls/122283
2024-05-27 19:33:35 +02:00
NodeLinkError{fmt::format("{}: {} " BLI_STR_UTF8_BLACK_RIGHT_POINTING_SMALL_TRIANGLE
" {}",
TIP_("Conversion is not supported"),
TIP_(link->fromsock->typeinfo->label),
TIP_(link->tosock->typeinfo->label))});
continue;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
if (fallback_zones) {
if (!fallback_zones->link_between_sockets_is_allowed(*link->fromsock, *link->tosock)) {
if (const bNodeTreeZone *from_zone = fallback_zones->get_zone_by_socket(*link->fromsock))
{
ntree.runtime->invalid_zone_output_node_ids.add(*from_zone->output_node_id);
}
link->flag &= ~NODE_LINK_VALID;
ntree.runtime->link_errors.add(
NodeLinkKey{*link},
NodeLinkError{TIP_("Links can only go into a zone but not out")});
continue;
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
bool check_if_output_changed(const bNodeTree &tree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
tree.ensure_topology_cache();
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* Compute a hash that represents the node topology connected to the output. This always has
* to be updated even if it is not used to detect changes right now. Otherwise
* #btree.runtime.output_topology_hash will go out of date. */
const Vector<const bNodeSocket *> tree_output_sockets = this->find_output_sockets(tree);
const uint32_t old_topology_hash = tree.runtime->output_topology_hash;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
const uint32_t new_topology_hash = this->get_combined_socket_topology_hash(
tree, tree_output_sockets);
tree.runtime->output_topology_hash = new_topology_hash;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (const AnimData *adt = BKE_animdata_from_id(&tree.id)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
/* Drivers may copy values in the node tree around arbitrarily and may cause the output to
* change even if it wouldn't without drivers. Only some special drivers like `frame/5` can
* be used without causing updates all the time currently. In the future we could try to
* handle other drivers better as well.
* Note that this optimization only works in practice when the depsgraph didn't also get a
* copy-on-evaluation tag for the node tree (which happens when changing node properties). It
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
* does work in a few situations like adding reroutes and duplicating nodes though. */
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
LISTBASE_FOREACH (const FCurve *, fcurve, &adt->drivers) {
const ChannelDriver *driver = fcurve->driver;
const StringRef expression = driver->expression;
if (expression.startswith("frame")) {
const StringRef remaining_expression = expression.drop_known_prefix("frame");
if (remaining_expression.find_first_not_of(" */+-0123456789.") == StringRef::not_found) {
continue;
}
}
/* Unrecognized driver, assume that the output always changes. */
return true;
}
}
if (tree.runtime->changed_flag & NTREE_CHANGED_ANY) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
if (old_topology_hash != new_topology_hash) {
return true;
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* The topology hash can only be used when only topology-changing operations have been done.
*/
if (tree.runtime->changed_flag ==
(tree.runtime->changed_flag & (NTREE_CHANGED_LINK | NTREE_CHANGED_REMOVED_NODE)))
{
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (old_topology_hash == new_topology_hash) {
return false;
}
}
if (!this->check_if_socket_outputs_changed_based_on_flags(tree, tree_output_sockets)) {
return false;
}
return true;
}
Vector<const bNodeSocket *> find_output_sockets(const bNodeTree &tree)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
Vector<const bNodeSocket *> sockets;
for (const bNode *node : tree.all_nodes()) {
if (!this->is_output_node(*node)) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
continue;
}
for (const bNodeSocket *socket : node->input_sockets()) {
if (!STREQ(socket->idname, "NodeSocketVirtual")) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
sockets.append(socket);
}
}
}
return sockets;
}
bool is_output_node(const bNode &node) const
{
if (node.typeinfo->nclass == NODE_CLASS_OUTPUT) {
return true;
}
if (node.is_group_output()) {
return true;
}
if (node.is_type("GeometryNodeWarning")) {
return true;
}
Geometry Nodes: support attaching gizmos to input values This adds support for attaching gizmos for input values. The goal is to make it easier for users to set input values intuitively in the 3D viewport. We went through multiple different possible designs until we settled on the one implemented here. We picked it for it's flexibility and ease of use when using geometry node assets. The core principle in the design is that **gizmos are attached to existing input values instead of being the input value themselves**. This actually fits the existing concept of gizmos in Blender well, but may be a bit unintutitive in a node setup at first. The attachment is done using links in the node editor. The most basic usage of the node is to link a Value node to the new Linear Gizmo node. This attaches the gizmo to the input value and allows you to change it from the 3D view. The attachment is indicated by the gizmo icon in the sockets which are controlled by a gizmo as well as the back-link (notice the double link) when the gizmo is active. The core principle makes it straight forward to control the same node setup from the 3D view with gizmos, or by manually changing input values, or by driving the input values procedurally. If the input value is controlled indirectly by other inputs, it's often possible to **automatically propagate** the gizmo to the actual input. Backpropagation does not work for all nodes, although more nodes can be supported over time. This patch adds the first three gizmo nodes which cover common use cases: * **Linear Gizmo**: Creates a gizmo that controls a float or integer value using a linear movement of e.g. an arrow in the 3D viewport. * **Dial Gizmo**: Creates a circular gizmo in the 3D viewport that can be rotated to change the attached angle input. * **Transform Gizmo**: Creates a simple gizmo for location, rotation and scale. In the future, more built-in gizmos and potentially the ability for custom gizmos could be added. All gizmo nodes have a **Transform** geometry output. Using it is optional but it is recommended when the gizmo is used to control inputs that affect a geometry. When it is used, Blender will automatically transform the gizmos together with the geometry that they control. To achieve this, the output should be merged with the generated geometry using the *Join Geometry* node. The data contained in *Transform* output is not visible geometry, but just internal information that helps Blender to give a better user experience when using gizmos. The gizmo nodes have a multi-input socket. This allows **controlling multiple values** with the same gizmo. Only a small set of **gizmo shapes** is supported initially. It might be extended in the future but one goal is to give the gizmos used by different node group assets a familiar look and feel. A similar constraint exists for **colors**. Currently, one can choose from a fixed set of colors which can be modified in the theme settings. The set of **visible gizmos** is determined by a multiple factors because it's not really feasible to show all possible gizmos at all times. To see any of the geometry nodes gizmos, the "Active Modifier" option has to be enabled in the "Viewport Gizmos" popover. Then all gizmos are drawn for which at least one of the following is true: * The gizmo controls an input of the active modifier of the active object. * The gizmo controls a value in a selected node in an open node editor. * The gizmo controls a pinned value in an open node editor. Pinning works by clicking the gizmo icon next to the value. Pull Request: https://projects.blender.org/blender/blender/pulls/112677
2024-07-10 16:18:47 +02:00
if (nodes::gizmos::is_builtin_gizmo_node(node)) {
return true;
}
/* Assume node groups without output sockets are outputs. */
if (node.is_group()) {
const bNodeTree *node_group = reinterpret_cast<const bNodeTree *>(node.id);
if (node_group != nullptr &&
node_group->runtime->runtime_flag & NTREE_RUNTIME_FLAG_HAS_MATERIAL_OUTPUT)
{
return true;
}
}
return false;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
/**
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
* Computes a hash that changes when the node tree topology connected to an output node
* changes. Adding reroutes does not have an effect on the hash.
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
*/
uint32_t get_combined_socket_topology_hash(const bNodeTree &tree,
Span<const bNodeSocket *> sockets)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
if (tree.has_available_link_cycle()) {
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* Return dummy value when the link has any cycles. The algorithm below could be improved
* to handle cycles more gracefully. */
return 0;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
Array<uint32_t> hashes = this->get_socket_topology_hashes(tree, sockets);
uint32_t combined_hash = 0;
for (uint32_t hash : hashes) {
combined_hash = noise::hash(combined_hash, hash);
}
return combined_hash;
}
Array<uint32_t> get_socket_topology_hashes(const bNodeTree &tree,
const Span<const bNodeSocket *> sockets)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
BLI_assert(!tree.has_available_link_cycle());
Array<std::optional<uint32_t>> hash_by_socket_id(tree.all_sockets().size());
Stack<const bNodeSocket *> sockets_to_check = sockets;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
auto get_socket_ptr_hash = [&](const bNodeSocket &socket) {
const uint64_t socket_ptr = uintptr_t(&socket);
return noise::hash(socket_ptr, socket_ptr >> 32);
};
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
while (!sockets_to_check.is_empty()) {
const bNodeSocket &socket = *sockets_to_check.peek();
const bNode &node = socket.owner_node();
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (hash_by_socket_id[socket.index_in_tree()].has_value()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
sockets_to_check.pop();
/* Socket is handled already. */
continue;
}
uint32_t socket_hash = 0;
if (socket.is_input()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
/* For input sockets, first compute the hashes of all linked sockets. */
bool all_origins_computed = true;
bool get_value_from_origin = false;
for (const bNodeLink *link : socket.directly_linked_links()) {
if (link->is_muted()) {
continue;
}
if (!link->is_available()) {
continue;
}
const bNodeSocket &origin_socket = *link->fromsock;
const std::optional<uint32_t> origin_hash =
hash_by_socket_id[origin_socket.index_in_tree()];
if (origin_hash.has_value()) {
if (get_value_from_origin || socket.type != origin_socket.type) {
socket_hash = noise::hash(socket_hash, *origin_hash);
}
else {
/* Copy the socket hash because the link did not change it. */
socket_hash = *origin_hash;
}
get_value_from_origin = true;
}
else {
sockets_to_check.push(&origin_socket);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
all_origins_computed = false;
}
}
if (!all_origins_computed) {
continue;
}
if (!get_value_from_origin) {
socket_hash = get_socket_ptr_hash(socket);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
else {
bool all_available_inputs_computed = true;
for (const bNodeSocket *input_socket : node.input_sockets()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (input_socket->is_available()) {
if (!hash_by_socket_id[input_socket->index_in_tree()].has_value()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
sockets_to_check.push(input_socket);
all_available_inputs_computed = false;
}
}
}
if (!all_available_inputs_computed) {
continue;
}
if (node.is_reroute()) {
socket_hash = *hash_by_socket_id[node.input_socket(0).index_in_tree()];
}
else if (node.is_muted()) {
const bNodeSocket *internal_input = socket.internal_link_input();
if (internal_input == nullptr) {
socket_hash = get_socket_ptr_hash(socket);
}
else {
if (internal_input->type == socket.type) {
socket_hash = *hash_by_socket_id[internal_input->index_in_tree()];
}
else {
socket_hash = get_socket_ptr_hash(socket);
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
else {
socket_hash = get_socket_ptr_hash(socket);
for (const bNodeSocket *input_socket : node.input_sockets()) {
if (input_socket->is_available()) {
const uint32_t input_socket_hash = *hash_by_socket_id[input_socket->index_in_tree()];
socket_hash = noise::hash(socket_hash, input_socket_hash);
}
}
/* The Image Texture node has a special case. The behavior of the color output changes
* depending on whether the Alpha output is linked. */
if (node.is_type("ShaderNodeTexImage") && socket.index() == 0) {
BLI_assert(STREQ(socket.name, "Color"));
const bNodeSocket &alpha_socket = node.output_socket(1);
BLI_assert(STREQ(alpha_socket.name, "Alpha"));
if (alpha_socket.is_directly_linked()) {
socket_hash = noise::hash(socket_hash);
}
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
hash_by_socket_id[socket.index_in_tree()] = socket_hash;
/* Check that nothing has been pushed in the meantime. */
BLI_assert(sockets_to_check.peek() == &socket);
sockets_to_check.pop();
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
/* Create output array. */
Array<uint32_t> hashes(sockets.size());
for (const int i : sockets.index_range()) {
hashes[i] = *hash_by_socket_id[sockets[i]->index_in_tree()];
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
return hashes;
}
/**
* Returns true when any of the provided sockets changed its values. A change is detected by
* checking the #changed_flag on connected sockets and nodes.
*/
bool check_if_socket_outputs_changed_based_on_flags(const bNodeTree &tree,
Span<const bNodeSocket *> sockets)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
/* Avoid visiting the same socket twice when multiple links point to the same socket. */
Array<bool> pushed_by_socket_id(tree.all_sockets().size(), false);
Stack<const bNodeSocket *> sockets_to_check = sockets;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
for (const bNodeSocket *socket : sockets) {
pushed_by_socket_id[socket->index_in_tree()] = true;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
while (!sockets_to_check.is_empty()) {
const bNodeSocket &socket = *sockets_to_check.pop();
const bNode &node = socket.owner_node();
if (socket.runtime->changed_flag != NTREE_CHANGED_NOTHING) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
if (node.runtime->changed_flag != NTREE_CHANGED_NOTHING) {
const bool only_unused_internal_link_changed = !node.is_muted() &&
node.runtime->changed_flag ==
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
NTREE_CHANGED_INTERNAL_LINK;
const bool only_parent_changed = node.runtime->changed_flag == NTREE_CHANGED_PARENT;
const bool change_affects_output = !(only_unused_internal_link_changed ||
only_parent_changed);
if (change_affects_output) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
return true;
}
}
if (socket.is_input()) {
for (const bNodeSocket *origin_socket : socket.directly_linked_sockets()) {
bool &pushed = pushed_by_socket_id[origin_socket->index_in_tree()];
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (!pushed) {
sockets_to_check.push(origin_socket);
pushed = true;
}
}
}
else {
for (const bNodeSocket *input_socket : node.input_sockets()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (input_socket->is_available()) {
bool &pushed = pushed_by_socket_id[input_socket->index_in_tree()];
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (!pushed) {
sockets_to_check.push(input_socket);
pushed = true;
}
}
}
Geometry Nodes: new For Each Geometry Element zone This adds a new type of zone to Geometry Nodes that allows executing some nodes for each element in a geometry. ## Features * The `Selection` input allows iterating over a subset of elements on the set domain. * Fields passed into the input node are available as single values inside of the zone. * The input geometry can be split up into separate (completely independent) geometries for each element (on all domains except face corner). * New attributes can be created on the input geometry by outputting a single value from each iteration. * New geometries can be generated in each iteration. * All of these geometries are joined to form the final output. * Attributes from the input geometry are propagated to the output geometries. ## Evaluation The evaluation strategy is similar to the one used for repeat zones. Namely, it dynamically builds a `lazy_function::Graph` once it knows how many iterations are necessary. It contains a separate node for each iteration. The inputs for each iteration are hardcoded into the graph. The outputs of each iteration a passed to a separate lazy-function that reduces all the values down to the final outputs. This final output can have a huge number of inputs and that is not ideal for multi-threading yet, but that can still be improved in the future. ## Performance There is a non-neglilible amount of overhead for each iteration. The overhead is way larger than the per-element overhead when just doing field evaluation. Therefore, normal field evaluation should be preferred when possible. That can partially still be optimized if there is only some number crunching going on in the zone but that optimization is not implemented yet. However, processing many small geometries (e.g. each hair of a character separately) will likely **always be slower** than working on fewer larger geoemtries. The additional flexibility you get by processing each element separately comes at the cost that Blender can't optimize the operation as well. For node groups that need to handle lots of geometry elements, we recommend trying to design the node setup so that iteration over tiny sub-geometries is not required. An opposite point is true as well though. It can be faster to process more medium sized geometries in parallel than fewer very large geometries because of more multi-threading opportunities. The exact threshold between tiny, medium and large geometries depends on a lot of factors though. Overall, this initial version of the new zone does not implement all optimization opportunities yet, but the points mentioned above will still hold true later. Pull Request: https://projects.blender.org/blender/blender/pulls/127331
2024-09-24 11:52:02 +02:00
/* Zones may propagate changes from the input node to the output node even though there is
* no explicit link. */
switch (node.type_legacy) {
Geometry Nodes: new For Each Geometry Element zone This adds a new type of zone to Geometry Nodes that allows executing some nodes for each element in a geometry. ## Features * The `Selection` input allows iterating over a subset of elements on the set domain. * Fields passed into the input node are available as single values inside of the zone. * The input geometry can be split up into separate (completely independent) geometries for each element (on all domains except face corner). * New attributes can be created on the input geometry by outputting a single value from each iteration. * New geometries can be generated in each iteration. * All of these geometries are joined to form the final output. * Attributes from the input geometry are propagated to the output geometries. ## Evaluation The evaluation strategy is similar to the one used for repeat zones. Namely, it dynamically builds a `lazy_function::Graph` once it knows how many iterations are necessary. It contains a separate node for each iteration. The inputs for each iteration are hardcoded into the graph. The outputs of each iteration a passed to a separate lazy-function that reduces all the values down to the final outputs. This final output can have a huge number of inputs and that is not ideal for multi-threading yet, but that can still be improved in the future. ## Performance There is a non-neglilible amount of overhead for each iteration. The overhead is way larger than the per-element overhead when just doing field evaluation. Therefore, normal field evaluation should be preferred when possible. That can partially still be optimized if there is only some number crunching going on in the zone but that optimization is not implemented yet. However, processing many small geometries (e.g. each hair of a character separately) will likely **always be slower** than working on fewer larger geoemtries. The additional flexibility you get by processing each element separately comes at the cost that Blender can't optimize the operation as well. For node groups that need to handle lots of geometry elements, we recommend trying to design the node setup so that iteration over tiny sub-geometries is not required. An opposite point is true as well though. It can be faster to process more medium sized geometries in parallel than fewer very large geometries because of more multi-threading opportunities. The exact threshold between tiny, medium and large geometries depends on a lot of factors though. Overall, this initial version of the new zone does not implement all optimization opportunities yet, but the points mentioned above will still hold true later. Pull Request: https://projects.blender.org/blender/blender/pulls/127331
2024-09-24 11:52:02 +02:00
case GEO_NODE_REPEAT_OUTPUT:
case GEO_NODE_SIMULATION_OUTPUT:
case GEO_NODE_FOREACH_GEOMETRY_ELEMENT_OUTPUT: {
const bNodeTreeZones *zones = tree.zones();
if (!zones) {
break;
}
const bNodeTreeZone *zone = zones->get_zone_by_node(node.identifier);
if (!zone->input_node()) {
Geometry Nodes: new For Each Geometry Element zone This adds a new type of zone to Geometry Nodes that allows executing some nodes for each element in a geometry. ## Features * The `Selection` input allows iterating over a subset of elements on the set domain. * Fields passed into the input node are available as single values inside of the zone. * The input geometry can be split up into separate (completely independent) geometries for each element (on all domains except face corner). * New attributes can be created on the input geometry by outputting a single value from each iteration. * New geometries can be generated in each iteration. * All of these geometries are joined to form the final output. * Attributes from the input geometry are propagated to the output geometries. ## Evaluation The evaluation strategy is similar to the one used for repeat zones. Namely, it dynamically builds a `lazy_function::Graph` once it knows how many iterations are necessary. It contains a separate node for each iteration. The inputs for each iteration are hardcoded into the graph. The outputs of each iteration a passed to a separate lazy-function that reduces all the values down to the final outputs. This final output can have a huge number of inputs and that is not ideal for multi-threading yet, but that can still be improved in the future. ## Performance There is a non-neglilible amount of overhead for each iteration. The overhead is way larger than the per-element overhead when just doing field evaluation. Therefore, normal field evaluation should be preferred when possible. That can partially still be optimized if there is only some number crunching going on in the zone but that optimization is not implemented yet. However, processing many small geometries (e.g. each hair of a character separately) will likely **always be slower** than working on fewer larger geoemtries. The additional flexibility you get by processing each element separately comes at the cost that Blender can't optimize the operation as well. For node groups that need to handle lots of geometry elements, we recommend trying to design the node setup so that iteration over tiny sub-geometries is not required. An opposite point is true as well though. It can be faster to process more medium sized geometries in parallel than fewer very large geometries because of more multi-threading opportunities. The exact threshold between tiny, medium and large geometries depends on a lot of factors though. Overall, this initial version of the new zone does not implement all optimization opportunities yet, but the points mentioned above will still hold true later. Pull Request: https://projects.blender.org/blender/blender/pulls/127331
2024-09-24 11:52:02 +02:00
break;
}
for (const bNodeSocket *input_socket : zone->input_node()->input_sockets()) {
Geometry Nodes: new For Each Geometry Element zone This adds a new type of zone to Geometry Nodes that allows executing some nodes for each element in a geometry. ## Features * The `Selection` input allows iterating over a subset of elements on the set domain. * Fields passed into the input node are available as single values inside of the zone. * The input geometry can be split up into separate (completely independent) geometries for each element (on all domains except face corner). * New attributes can be created on the input geometry by outputting a single value from each iteration. * New geometries can be generated in each iteration. * All of these geometries are joined to form the final output. * Attributes from the input geometry are propagated to the output geometries. ## Evaluation The evaluation strategy is similar to the one used for repeat zones. Namely, it dynamically builds a `lazy_function::Graph` once it knows how many iterations are necessary. It contains a separate node for each iteration. The inputs for each iteration are hardcoded into the graph. The outputs of each iteration a passed to a separate lazy-function that reduces all the values down to the final outputs. This final output can have a huge number of inputs and that is not ideal for multi-threading yet, but that can still be improved in the future. ## Performance There is a non-neglilible amount of overhead for each iteration. The overhead is way larger than the per-element overhead when just doing field evaluation. Therefore, normal field evaluation should be preferred when possible. That can partially still be optimized if there is only some number crunching going on in the zone but that optimization is not implemented yet. However, processing many small geometries (e.g. each hair of a character separately) will likely **always be slower** than working on fewer larger geoemtries. The additional flexibility you get by processing each element separately comes at the cost that Blender can't optimize the operation as well. For node groups that need to handle lots of geometry elements, we recommend trying to design the node setup so that iteration over tiny sub-geometries is not required. An opposite point is true as well though. It can be faster to process more medium sized geometries in parallel than fewer very large geometries because of more multi-threading opportunities. The exact threshold between tiny, medium and large geometries depends on a lot of factors though. Overall, this initial version of the new zone does not implement all optimization opportunities yet, but the points mentioned above will still hold true later. Pull Request: https://projects.blender.org/blender/blender/pulls/127331
2024-09-24 11:52:02 +02:00
if (input_socket->is_available()) {
bool &pushed = pushed_by_socket_id[input_socket->index_in_tree()];
if (!pushed) {
sockets_to_check.push(input_socket);
pushed = true;
}
}
}
break;
}
}
Geometry Nodes: Menu Switch Node This patch adds support for _Menu Switch_ nodes and enum definitions in node trees more generally. The design is based on the outcome of the [2022 Nodes Workshop](https://code.blender.org/2022/11/geometry-nodes-workshop-2022/#menu-switch). The _Menu Switch_ node is an advanced version of the _Switch_ node which has a customizable **menu input socket** instead of a simple boolean. The _items_ of this menu are owned by the node itself. Each item has a name and description and unique identifier that is used internally. A menu _socket_ represents a concrete value out of the list of items. To enable selection of an enum value for unconnected sockets the menu is presented as a dropdown list like built-in enums. When the socket is connected a shared pointer to the enum definition is propagated along links and stored in socket default values. This allows node groups to expose a menu from an internal menu switch as a parameter. The enum definition is a runtime copy of the enum items in DNA that allows sharing. A menu socket can have multiple connections, which can lead to ambiguity. If two or more different menu source nodes are connected to a socket it gets marked as _undefined_. Any connection to an undefined menu socket is invalid as a hint to users that there is a problem. A warning/error is also shown on nodes with undefined menu sockets. At runtime the value of a menu socket is the simple integer identifier. This can also be a field in geometry nodes. The identifier is unique within each enum definition, and it is persistent even when items are added, removed, or changed. Changing the name of an item does not affect the internal identifier, so users can rename enum items without breaking existing input values. This also persists if, for example, a linked node group is temporarily unavailable. Pull Request: https://projects.blender.org/blender/blender/pulls/113445
2024-01-26 12:40:01 +01:00
/* The Normal node has a special case, because the value stored in the first output
* socket is used as input in the node. */
if ((node.is_type("ShaderNodeNormal") || node.is_type("CompositorNodeNormal")) &&
socket.index() == 1)
{
BLI_assert(STREQ(socket.name, "Dot"));
const bNodeSocket &normal_output = node.output_socket(0);
BLI_assert(STREQ(normal_output.name, "Normal"));
bool &pushed = pushed_by_socket_id[normal_output.index_in_tree()];
if (!pushed) {
sockets_to_check.push(&normal_output);
pushed = true;
}
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
return false;
}
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
/**
* Make sure that the #bNodeTree::nested_node_refs is up to date. It's supposed to contain a
* reference to all (nested) simulation zones.
*/
bool update_nested_node_refs(bNodeTree &ntree)
{
ntree.ensure_topology_cache();
/* Simplify lookup of old ids. */
Map<bNestedNodePath, int32_t> old_id_by_path;
Set<int32_t> old_ids;
for (const bNestedNodeRef &ref : ntree.nested_node_refs_span()) {
old_id_by_path.add(ref.path, ref.id);
old_ids.add(ref.id);
}
Vector<bNestedNodePath> nested_node_paths;
/* Don't forget nested node refs just because the linked file is not available right now. */
for (const bNestedNodePath &path : old_id_by_path.keys()) {
const bNode *node = ntree.node_by_id(path.node_id);
if (node && node->is_group() && node->id) {
if (node->id->tag & ID_TAG_MISSING) {
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
nested_node_paths.append(path);
}
}
}
if (ntree.type == NTREE_GEOMETRY) {
Geometry Nodes: new Bake node This adds a new `Bake` node which allows saving and loading intermediate geometries. Typical use cases we want address with this currently are: * Bake some data for use with a render engine. * Bake parts of the node tree explicitly for better performance. For now, the format that is written to disk is not considered to be an import/export format. It's not guaranteed that data written with one Blender version can be read by another Blender version. For that it's better to use proper interchange formats. Better support for those will be added eventually as well. We also plan an `Import Bake` node that allows reading the blender-specific baked data independent of the Bake node and at different frames. The baking works very similar to the baking in the simulation zone (UI and implementation wise). Major differences are: * The Bake node has a `Bake Still` and `Bake Animation` mode. * The Bake node doesn't do automatic caching. Implementation details: * Refactored how we create the Python operators for moving socket items so that it also makes sense for non-zones. * The `ModifierCache` stores an independent map of `SimulationNodeCache` and `BakeNodeCache`, but both share a common data structure for the actually baked data. * For baking, the `Bake` node is added as a side-effect-node in the modifier. This will make sure that the node is baked even if it's currently not connected to the output. * Had to add a new `DEG_id_tag_update_for_side_effect_request` function that is used during baking. It's necessary because I want to evaluate the object again even though none of its inputs changed. The reevaluation is necessary to create the baked data. Using `DEG_id_tag_update` technically works as well, but has the problem that it also uses the `DEG_UPDATE_SOURCE_USER_EDIT` flag which (rightly) invalidates simulation caches which shouldn't happen here. * Slightly refactored the timeline drawing so that it can also show the baked ranges of Bake nodes. It does not show anything for baked nodes with a in Still mode though. * The bake operator is refactored to bake a list of `NodeBakeRequest` which makes the code easier to follow compared to the previous nested `ObjectBakeData > ModifierBakeData > NodeBakeData` data structure. * The bake operators are disabled when the .blend file is not yet saved. This is technically only necessary when the bake path depends on the .blend file path but seems ok to force the user anyway (otherwise the bake path may be lost as well if it's set explicitly). * The same operators are used to bake and delete single bakes in `Bake` nodes and `Simulation Zones`. On top of that, there are separate operators of baking and deleting all simulation bakes (those ignore bake nodes). * The `Bake` node remembers which inputs have been fields and thus may be baked as attributes. For that it uses an `Is Attribute` flag on the socket item. This is needed because the baked data may still contain attribute data, even if the inputs to the bake node are disconnected. * Similar to simulation zones, the behavior of `Bake` nodes is passed into the geometry nodes evaluation from the outside (from the modifier only currently). This is done by providing the new `GeoNodesBakeParams` in `GeoNodesCallData` when executing geometry nodes. Next Steps (mostly because they also involve simulations): * Visualize nodes that have not been evaluated in the last evaluation. * Fix issue with seemingly loosing baked data after undo. * Improve error handling when baked data is not found. * Show bake node in link drag search. * Higher level tools for managing bakes. Pull Request: https://projects.blender.org/blender/blender/pulls/115466
2023-12-18 13:01:06 +01:00
/* Create references for simulations and bake nodes in geometry nodes.
* Those are the nodes that we want to store settings for at a higher level. */
for (StringRefNull idname : {"GeometryNodeSimulationOutput", "GeometryNodeBake"}) {
for (const bNode *node : ntree.nodes_by_type(idname)) {
nested_node_paths.append({node->identifier, -1});
}
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
}
}
/* Propagate references to nested nodes in group nodes. */
for (const bNode *node : ntree.group_nodes()) {
const bNodeTree *group = reinterpret_cast<const bNodeTree *>(node->id);
if (group == nullptr) {
continue;
}
for (const int i : group->nested_node_refs_span().index_range()) {
const bNestedNodeRef &child_ref = group->nested_node_refs[i];
nested_node_paths.append({node->identifier, child_ref.id});
}
}
/* Used to generate new unique IDs if necessary. */
RandomNumberGenerator rng = RandomNumberGenerator::from_random_seed();
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
Map<int32_t, bNestedNodePath> new_path_by_id;
for (const bNestedNodePath &path : nested_node_paths) {
const int32_t old_id = old_id_by_path.lookup_default(path, -1);
if (old_id != -1) {
/* The same path existed before, it should keep the same ID as before. */
new_path_by_id.add(old_id, path);
continue;
}
int32_t new_id;
while (true) {
new_id = rng.get_int32(INT32_MAX);
if (!old_ids.contains(new_id) && !new_path_by_id.contains(new_id)) {
break;
}
}
/* The path is new, it should get a new ID that does not collide with any existing IDs. */
new_path_by_id.add(new_id, path);
}
/* Check if the old and new references are identical. */
if (!this->nested_node_refs_changed(ntree, new_path_by_id)) {
return false;
}
MEM_SAFE_FREE(ntree.nested_node_refs);
if (new_path_by_id.is_empty()) {
ntree.nested_node_refs_num = 0;
return true;
}
/* Allocate new array for the nested node references contained in the node tree. */
bNestedNodeRef *new_refs = MEM_malloc_arrayN<bNestedNodeRef>(size_t(new_path_by_id.size()),
__func__);
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
int index = 0;
for (const auto item : new_path_by_id.items()) {
bNestedNodeRef &ref = new_refs[index];
ref.id = item.key;
ref.path = item.value;
index++;
}
ntree.nested_node_refs = new_refs;
ntree.nested_node_refs_num = new_path_by_id.size();
return true;
}
bool nested_node_refs_changed(const bNodeTree &ntree,
const Map<int32_t, bNestedNodePath> &new_path_by_id)
{
if (ntree.nested_node_refs_num != new_path_by_id.size()) {
return true;
}
for (const bNestedNodeRef &ref : ntree.nested_node_refs_span()) {
if (!new_path_by_id.contains(ref.id)) {
return true;
}
}
return false;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void reset_changed_flags(bNodeTree &ntree)
{
ntree.runtime->changed_flag = NTREE_CHANGED_NOTHING;
for (bNode *node : ntree.all_nodes()) {
node->runtime->changed_flag = NTREE_CHANGED_NOTHING;
node->runtime->update = 0;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
LISTBASE_FOREACH (bNodeSocket *, socket, &node->inputs) {
socket->runtime->changed_flag = NTREE_CHANGED_NOTHING;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
LISTBASE_FOREACH (bNodeSocket *, socket, &node->outputs) {
socket->runtime->changed_flag = NTREE_CHANGED_NOTHING;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
}
ntree.tree_interface.reset_changed_flags();
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
/**
* Update the panel toggle sockets to use the same name as the panel.
*/
bool update_panel_toggle_names(bNodeTree &ntree)
{
bool changed = false;
ntree.ensure_interface_cache();
for (bNodeTreeInterfaceItem *item : ntree.interface_items()) {
if (item->item_type != NODE_INTERFACE_PANEL) {
continue;
}
bNodeTreeInterfacePanel *panel = reinterpret_cast<bNodeTreeInterfacePanel *>(item);
if (bNodeTreeInterfaceSocket *toggle_socket = panel->header_toggle_socket()) {
if (!STREQ(panel->name, toggle_socket->name)) {
MEM_SAFE_FREE(toggle_socket->name);
toggle_socket->name = BLI_strdup_null(panel->name);
changed = true;
}
}
}
return changed;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
};
} // namespace blender::bke
void BKE_ntree_update_tag_all(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_ANY);
}
void BKE_ntree_update_tag_node_property(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_NODE_PROPERTY);
}
void BKE_ntree_update_tag_node_new(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_NODE_PROPERTY);
2025-02-11 17:25:10 +01:00
}
void BKE_ntree_update_tag_node_type(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_NODE_PROPERTY);
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
}
void BKE_ntree_update_tag_socket_property(bNodeTree *ntree, bNodeSocket *socket)
{
add_socket_tag(ntree, socket, NTREE_CHANGED_SOCKET_PROPERTY);
}
void BKE_ntree_update_tag_socket_new(bNodeTree *ntree, bNodeSocket *socket)
{
add_socket_tag(ntree, socket, NTREE_CHANGED_SOCKET_PROPERTY);
}
void BKE_ntree_update_tag_socket_removed(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_REMOVED_SOCKET);
}
void BKE_ntree_update_tag_socket_type(bNodeTree *ntree, bNodeSocket *socket)
{
add_socket_tag(ntree, socket, NTREE_CHANGED_SOCKET_PROPERTY);
}
void BKE_ntree_update_tag_socket_availability(bNodeTree *ntree, bNodeSocket *socket)
{
add_socket_tag(ntree, socket, NTREE_CHANGED_SOCKET_PROPERTY);
}
void BKE_ntree_update_tag_node_removed(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_REMOVED_NODE);
}
void BKE_ntree_update_tag_node_mute(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_NODE_PROPERTY);
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void BKE_ntree_update_tag_node_internal_link(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_INTERNAL_LINK);
}
void BKE_ntree_update_tag_link_changed(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_LINK);
}
void BKE_ntree_update_tag_link_removed(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_LINK);
}
void BKE_ntree_update_tag_link_added(bNodeTree *ntree, bNodeLink * /*link*/)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
add_tree_tag(ntree, NTREE_CHANGED_LINK);
}
void BKE_ntree_update_tag_link_mute(bNodeTree *ntree, bNodeLink * /*link*/)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
add_tree_tag(ntree, NTREE_CHANGED_LINK);
}
void BKE_ntree_update_tag_active_output_changed(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_ANY);
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void BKE_ntree_update_tag_missing_runtime_data(bNodeTree *ntree)
{
add_tree_tag(ntree, NTREE_CHANGED_ALL);
}
void BKE_ntree_update_tag_parent_change(bNodeTree *ntree, bNode *node)
{
add_node_tag(ntree, node, NTREE_CHANGED_PARENT);
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void BKE_ntree_update_tag_id_changed(Main *bmain, ID *id)
{
FOREACH_NODETREE_BEGIN (bmain, ntree, ntree_id) {
for (bNode *node : ntree->all_nodes()) {
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (node->id == id) {
node->runtime->update |= NODE_UPDATE_ID;
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
add_node_tag(ntree, node, NTREE_CHANGED_NODE_PROPERTY);
}
}
}
FOREACH_NODETREE_END;
}
void BKE_ntree_update_tag_image_user_changed(bNodeTree *ntree, ImageUser * /*iuser*/)
{
/* Would have to search for the node that uses the image user for a more detailed tag. */
add_tree_tag(ntree, NTREE_CHANGED_ANY);
}
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
uint64_t bNestedNodePath::hash() const
{
return blender::get_default_hash(this->node_id, this->id_in_node);
Nodes: add nested node ids and use them for simulation state The simulation state used by simulation nodes is owned by the modifier. Since a geometry nodes setup can contain an arbitrary number of simulations, the modifier has a mapping from `SimulationZoneID` to `SimulationZoneState`. This patch changes what is used as `SimulationZoneID`. Previously, the `SimulationZoneID` contained a list of `bNode::identifier` that described the path from the root node tree to the simulation output node. This works ok in many cases, but also has a significant problem: The `SimulationZoneID` changes when moving the simulation zone into or out of a node group. This implies that any of these operations loses the mapping from zone to simulation state, invalidating the cache or even baked data. The goal of this patch is to introduce a single-integer ID that identifies a (nested) simulation zone and is stable even when grouping and un-grouping. The ID should be stable even if the node group containing the (nested) simulation zone is in a separate linked .blend file and that linked file is changed. In the future, the same kind of ID can be used to store e.g. checkpoint/baked/frozen data in the modifier. To achieve the described goal, node trees can now store an arbitrary number of nested node references (an array of `bNestedNodeRef`). Each nested node reference has an ID that is unique within the current node tree. The node tree does not store the entire path to the nested node. Instead it only know which group node the nested node is in, and what the nested node ID of the node is within that group. Grouping and un-grouping operations have to update the nested node references to keep the IDs stable. Importantly though, these operations only have to care about the two node groups that are affected. IDs in higher level node groups remain unchanged by design. A consequence of this design is that every `bNodeTree` now has a `bNestedNodeRef` for every (nested) simulation zone. Two instances of the same simulation zone (because a node group is reused) are referenced by two separate `bNestedNodeRef`. This is important to keep in mind, because it also means that this solution doesn't scale well if we wanted to use it to keep stable references to *all* nested nodes. I can't think of a solution that fulfills the described requirements but scales better with more nodes. For that reason, this solution should only be used when we want to store data for each referenced nested node at the top level (like we do for simulations). This is not a replacement for `ViewerPath` which can store a path to data in a node tree without changing the node tree. Also `ViewerPath` can contain information like the loop iteration that should be viewed (#109164). `bNestedNodeRef` can't differentiate between different iterations of a loop. This also means that simulations can't be used inside of a loop (loops inside of a simulation work fine though). When baking, the new stable ID is now written to disk, which means that baked data is not invalidated by grouping/un-grouping operations. Backward compatibility for baked data is provided, but only works as long as the simulation zone has not been moved to a different node group yet. Forward compatibility for the baked data is not provided (so older versions can't load the data baked with a newer version of Blender). Pull Request: https://projects.blender.org/blender/blender/pulls/109444
2023-07-01 11:54:32 +02:00
}
bool operator==(const bNestedNodePath &a, const bNestedNodePath &b)
{
return a.node_id == b.node_id && a.id_in_node == b.id_in_node;
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
/**
* Protect from recursive calls into the updating function. Some node update functions might
* trigger this from Python or in other cases.
*
* This could be added to #Main, but given that there is generally only one #Main, that's not
* really worth it now.
*/
static bool is_updating = false;
void BKE_ntree_update(Main &bmain,
const std::optional<blender::Span<bNodeTree *>> modified_trees,
const NodeTreeUpdateExtraParams &params)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
if (is_updating) {
return;
}
is_updating = true;
blender::bke::NodeTreeMainUpdater updater{&bmain, params};
if (modified_trees.has_value()) {
updater.update_rooted(*modified_trees);
}
else {
updater.update();
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
is_updating = false;
}
void BKE_ntree_update_after_single_tree_change(Main &bmain,
bNodeTree &modified_tree,
const NodeTreeUpdateExtraParams &params)
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
{
BKE_ntree_update(bmain, blender::Span{&modified_tree}, params);
}
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
void BKE_ntree_update_without_main(bNodeTree &tree)
{
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
if (is_updating) {
return;
}
is_updating = true;
NodeTreeUpdateExtraParams params;
blender::bke::NodeTreeMainUpdater updater{nullptr, params};
updater.update_rooted({&tree});
Nodes: refactor node tree update handling Goals of this refactor: * More unified approach to updating everything that needs to be updated after a change in a node tree. * The updates should happen in the correct order and quadratic or worse algorithms should be avoided. * Improve detection of changes to the output to avoid tagging the depsgraph when it's not necessary. * Move towards a more declarative style of defining nodes by having a more centralized update procedure. The refactor consists of two main parts: * Node tree tagging and update refactor. * Generally, when changes are done to a node tree, it is tagged dirty until a global update function is called that updates everything in the correct order. * The tagging is more fine-grained compared to before, to allow for more precise depsgraph update tagging. * Depsgraph changes. * The shading specific depsgraph node for node trees as been removed. * Instead, there is a new `NTREE_OUTPUT` depsgrap node, which is only tagged when the output of the node tree changed (e.g. the Group Output or Material Output node). * The copy-on-write relation from node trees to the data block they are embedded in is now non-flushing. This avoids e.g. triggering a material update after the shader node tree changed in unrelated ways. Instead the material has a flushing relation to the new `NTREE_OUTPUT` node now. * The depsgraph no longer reports data block changes through to cycles through `Depsgraph.updates` when only the node tree changed in ways that do not affect the output. Avoiding unnecessary updates seems to work well for geometry nodes and cycles. The situation is a bit worse when there are drivers on the node tree, but that could potentially be improved separately in the future. Avoiding updates in eevee and the compositor is more tricky, but also less urgent. * Eevee updates are triggered by calling `DRW_notify_view_update` in `ED_render_view3d_update` indirectly from `DEG_editors_update`. * Compositor updates are triggered by `ED_node_composite_job` in `node_area_refresh`. This is triggered by calling `ED_area_tag_refresh` in `node_area_listener`. Removing updates always has the risk of breaking some dependency that no one was aware of. It's not unlikely that this will happen here as well. Adding back missing updates should be quite a bit easier than getting rid of unnecessary updates though. Differential Revision: https://developer.blender.org/D13246
2021-12-21 15:18:56 +01:00
is_updating = false;
}