This adds a function that can turn an existing `bNodeTree` into an inlined one.
The new node tree has all node groups, repeat zones, closures and bundles
inlined. So it's just a flat tree that ideally can be consumed easily by render
engines. As part of the process, it also does constant folding.
The goal is to support more advanced features from geometry nodes (repeat zones,
etc.) in shader nodes which the evaluator is more limited because it has to be
able to run on the GPU. Creating an inlined `bNodeTree` is likely the most
direct way to get but may also be limiting in the future. Since this is a fairly
local change, it's likely still worth it to support these features in all render
engines without having to make their evaluators significantly more complex.
Some limitations apply here that do not apply in Geometry Nodes. For example,
the iterations count in a repeat zone has to be a constant after constant
folding.
There is also a `Test Inlining Shader Nodes` operator that creates the inlined
tree and creates a group node for it. This is just for testing purposes.
#145811 will make this functionality available to the Python API as well so that
external renderers can use it too.
Previously, when a node was muted, all automatically hidden sockets suddenly
became visible. This is because of a rule where inputs that are never ever used,
are not-hidden (only sockets where the usage depends on some condition are
allowed to be hidden automatically). Since the inputs without internal link in a
muted nodes are never used, they become visible when a node is muted.
The solution is to ignore whether a node is muted in the case when the socket
usage for node-editor-drawing is computed. In nested nodes, muting should not be
ignored.
Pull Request: https://projects.blender.org/blender/blender/pulls/145729
This adds a boolean output for each of the menu items. The output is true, if
the passed in menu value is that item. This avoids the need to compare the
output value to the input values to get a boolean for whether a specific menu
item was passed in.
Support is added for Geometry Nodes as well as the Compositor. Usage/Value
inferencing has been updated as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/145712
Previously, group input values were logged for each group input socket. This
means that each value might be logged many times currently when there are many
Group Input nodes.
This patch changes it so that group input values are only logged for a single
socket. In the provided file from #145385 this speeds up playback performance
from 23 to 39 fps. The next bottleneck there is node editor drawing. If I change
the Geometry Nodes editor to another editor type the speedup is 36 (`main`) to
110 fps (this patch).
Pull Request: https://projects.blender.org/blender/blender/pulls/145781
Functions for convert between the color types and ostream support are
now outside the classes.
Many files were changed to fix cases where direct includes for headers
were missing.
Pull Request: https://projects.blender.org/blender/blender/pulls/145756
Improve handling of runtime defined python RNA properties. Mainly:
* Add `get_transform` and `set_transform` new callbacks.
These allow to edit the value, while still using the default
(IDProperty-based) storage system.
* Read-only properties should now be defined using a new `options` flag,
`READ_ONLY`.
* `get`/`set` should only be used when storing data outside of the
default system now.
* Having a `get` without a `set` defined forces property to be
read-only (same behavior as before).
* Having a `set` without a `get` is now an error.
* Just like with existing `get/set` callbacks, `get_/set_transform`
callbacks must always generate values matching the constraints defined
by their `bpy.props` property definition (same type, within required
range, same dimensions/sizes for the `Vector` properties, etc.).
* To simplify handling of non-statically sized strings, the relevant
RNA API has been modified, to use `std::string` instead of
(allocated) char arrays.
Relevant unittests and benchmarking have been added or updated as part
of this project.
Note: From initial benchmarking, 'transform' versions of get/set are
several times faster than 'real' get/set.
Implements #141042.
Pull Request: https://projects.blender.org/blender/blender/pulls/141303
Currently, `InferenceValue` contains either a primitive value or is in the
"unknown" state. Previously, all code had the assumption that when the value is
not unknown, it is a primitive value. However, that may not be true anymore when
we support inferencing through e.g. bundles and closures as those are not
"primitive". This assumption is removed now, non-primitive values have not been
added yet though.
Pull Request: https://projects.blender.org/blender/blender/pulls/145532
Value and usage inferencing can be done independently. Usage-inferencing uses
the value-inferencing but not the other way around. Extracting value-inferencing
makes the separation more clear and also simplifies potentially reusing the
value-inferencing code later on.
Pull Request: https://projects.blender.org/blender/blender/pulls/145492
This patch turns node Menu options into menu inputs. This patch only
covers node operations like Filter, Distort, and so on. Pixel nodes like
Color Balance, Matte, and so on will be done in a separate patch.
Pull Request: https://projects.blender.org/blender/blender/pulls/144495
This is an extension of 131404a1bd. The inferencing time is now down
to 4-5ms from 24ms originally.
Further improvements are likely possible. Especially, caching the results
can be very beneficial, but cache invalidation may not be entirely trivial.
This was found to be a bottleneck in the file from #144756 which has 235
group inputs and 174 group input nodes, which leads to 40,890 sockets.
Most of those are never used, so it's wasteful to add them to the map
of input values.
The inferencing time improved from 24ms to 16ms. Can likely still be
improved more, but that's a good improvement for such a small change
already.
This patch generalizes node tree subtypes to be usable for node trees
other than Geometry Nodes. In particular, this:
- Renames SpaceNode.geometry_nodes_type to node_tree_sub_type, which now
store a tree type-specific enum.
- Renames SpaceNode.geometry_nodes_tool_tree to selected_node_group,
which now stores any context-less tree of any type.
This breaks the python API due to renaming.
Pull Request: https://projects.blender.org/blender/blender/pulls/144544
Previously, we always just used `int` when dealing with menu values on the C++
side. It's currently the only type where we have the same base type (`int`) for
two different socket types (integer and menu sockets). This has some downsides:
* It requires special cases in some places.
* There is no function from static base type to socket type (which is useful for
some utilities like `SocketValueVariant`).
* It implicitly allows operations on menu values that shouldn't be allowed such
as arithmetic operations and conversions to and from other types.
This patch adds a new `MenuValue` type that is used for menu sockets in Geometry
Nodes and the (CPU) Compositor, clarifying the distinction between integer and
menu values.
Pull Request: https://projects.blender.org/blender/blender/pulls/144476
This type was the same for every socket type supported by Geometry Nodes.
It's always `SocketValueVariant` now. Therefore, it was unnecessary to s
tore an explicit pointer to it.
Pull Request: https://projects.blender.org/blender/blender/pulls/144458
Now that every socket type uses `SocketValueVariant` since #144355, the
parameter access code can be simplified a bit.
This also adds a new `GeoNodesMultiInput<T>` type that's used to access the list
of inputs value multi-input sockets.
Pull Request: https://projects.blender.org/blender/blender/pulls/144409
This is similar to #144199. It needs a few more changes because more places
handle geometries in a special way compared to data-block sockets.
It might be that there are more usages of `GeometrySet` as value for geometry
sockets instead of `SocketValueVariant`. Unfortunately, there isn't really an
automated way to find these. So far I found all the places that needed fixing
through tests.
This is the last socket type that did not use `SocketValueVariant` yet. So
afterwards, it's likely possible to simplify a bunch of code to use
`SocketValueVariant` instead of `void *`.
Pull Request: https://projects.blender.org/blender/blender/pulls/144355
The fix is to detect when the compute context is recursive and to stop the
search early in that case. Currently, this does not generate a warning. There
will be a warning when trying to evaluate the recursive closure though.
Pull Request: https://projects.blender.org/blender/blender/pulls/144330
Use `SocketValueVariant` for all the data-block types as well. Before it was
used for e.g. float, integer and menu sockets.
The only remaining type afterwards is the geometry socket which will be moved
separately.
Using `SocketValueVariant` for all socket types will simplify code later on and
also removes the need for dealing with raw memory in more cases.
Pull Request: https://projects.blender.org/blender/blender/pulls/144199
Many nodes operate on all the instances that are passed into them. For example,
the Subdivision Surface node subdivides the mesh at the root but also instanced
meshes. This works well for most nodes, but there are a few nodes were the old
`modify_geometry_sets` function was not very well defined and it was tricky to
use correctly.
The fundamental problem was that the behavior is not obvious when a node creates
or modifies instances and how those are integrated with the already existing
instances.
This patch solves this with the following changes:
* Remove the old `GeometrySet::modify_geometry_sets` and related
`*_during_modify` methods.
* Add a new `blender::geometry::foreach_real_geometry` function that is similar
to the old `modify_geometry_sets` but has a more well-defined interface:
* It never passes instances into the callback. So existing instances can't be
modified with it.
* The callback is allowed to create new instances. This will automatically be
merged back with potentially already existing instances. The callback does
not have to worry about accidentally invalidating existing instances like
before.
* A few existing usages used `modify_geometry_sets` to actually modify existing
instances (usually just removing attributes). Those can't use the new
`foreach_real_geometry`, so they just get a custom simple recursive
implementation instead of using a generic function.
Pull Request: https://projects.blender.org/blender/blender/pulls/143898
This adds support for tracing bundle and closure structures through repeat zones,
simulation zones and bake nodes. Previously, syncing through these nodes just
didn't work.
Pull Request: https://projects.blender.org/blender/blender/pulls/143860
This adds support for creating Combine Bundle, Separate Bundle and Evaluate Closure
nodes using link drag search in some cases that were not previously supported.
Pull Request: https://projects.blender.org/blender/blender/pulls/143835
This removes the "Geometry" part from their name because we want to use them in
other node tree types too (see #141936).
Usually, we don't change idnames because they are the primary identifier of
nodes and is expected to stay the same. Since they are generally not shown to
users (just Python developers), it's also not urgent to change them. However, in
this specific case we have the opportunity to update the idname before the node
becomes an official feature, so it's not as bad to change it.
This patch has full backward compatibility, but no forward compatibility (for
files that used the experimental feature).
Pull Request: https://projects.blender.org/blender/blender/pulls/143823
The issue was that the parent node group was not necessarily updated already
when the tracing code ran. So use socket identifiers instead of indices to try
to find corresponding sockets between group nodes and their corresponding
node groups.
When using link-drag-search to create bundle or closure nodes, the newly created
nodes are already synced automatically. Now this automatic syncing also happens
when an empty node is first linked. If there are any sockets already, the
automatic syncing does not happen as it can be unintentional. In this case the
user can just click the sync icon in the node header to update the sockets.
Pull Request: https://projects.blender.org/blender/blender/pulls/143744
This moves code used for tracing bundles and closures to a separate file. This
code is used to e.g. detect which Separate Bundle node a Combine Bundle node is
linked to. This allows providing automatic socket update operators for these
nodes. Similarly for closures.
Pull Request: https://projects.blender.org/blender/blender/pulls/143734
This moves most of the code to deal with syncable nodes (such as
Combine/Separate Bundle) to the nodes module. Over time it might be possible to
decentralize it more.
This also changes the caching mechanism from storing a flag on the node to
storing a map on the node editor runtime data. This simplifies the code quite
significantly and also removes the need to store any of this data in DNA.
The node tree update code now always clears this cache because before it was
missing many cases, e.g. when creating links that would connect a Combine to a
Separate Bundle node.
Pull Request: https://projects.blender.org/blender/blender/pulls/143661
Currently, bundles can only store socket values of Geometry Nodes. However, it
can make sense to store other kinds of data too. Specifically, this patch adds
support for storing arbitrary internal data in a bundle. This is useful when
storing e.g. the physics world when implementing a proper physics solver for
Geometry Nodes (like in #143171).
One can still see that the data exists in Geometry Nodes in the tooltip, but one
can't extract it. Built-in nodes can still read that data.
Storing built-in data in bundles can also be done as an alternative to having a
new "internal data socket" as we talked in a workshop in the past:
https://code.blender.org/2024/11/geometry-nodes-workshop-october-2024/#internal-data-sockets
A bundle still has to be copyable. Internal data is expected to use implicit
sharing. That way copying it just requires increasing the user count of the
data.
Pull Request: https://projects.blender.org/blender/blender/pulls/143472
