The core issue was that the geometry batch cache (e.g. `MeshBatchCache` or
`PointCloudBatchCache`) was dependent on the object. This is problematic when
the the same geometry is used with multiple different objects because the cache
can't be consistent with all of them.
Fortunately, the only thing that was retrieved from the object was the number of
material slots, so if that can be avoided we should be fine. We can't just use
the number of material slots stored on the geometry because that may have no
material slots but still has material indices which are overridden on the object
level.
The solution is to take make the number of materials for a geometry only
dependent on the actual `material_index` attribute and not on the number of
available slots. More specifically, we find the maximal referenced material
index and handle that many materials. This number does not depend on how many
material slots there are on the object, but it still allows the object to
override materials slots that the mesh references.
A downside is that the maximum material index has to be computed which often
requires an iteration over the mesh. Fortunately, we can cache that quite easily
and the computation can be done in parallel. Also we are probably able to
eagerly update the material index in many cases when it's set instead of
computing it lazily. That is not implemented in this patch though.
The largest part of the patch is making the maximal material index easily
available on all the geometry types. Besides that, the material API is slightly
replaced and the drawing code now makes use of the updated API.
Pull Request: https://projects.blender.org/blender/blender/pulls/133498
This would change the image type in image_save_post, which would make the
logic in BKE_image_acquire/release_renderresult mismatched.
Also consistently call BKE_image_release_renderresult when the render result
is null, rather than only doing it half the time.
Pull Request: https://projects.blender.org/blender/blender/pulls/133503
* Ensure valid bit depth is set along with file type
* Guard against invalid inputs in stereo imbuf creation
* Remove some unused code
Thanks Yiming Wu for finding the cause.
Pull Request: https://projects.blender.org/blender/blender/pulls/133499
Before, it was only possible to apply modifiers through a multires modifier if
they were deform-only. The Geometry Nodes modifier is of course not deform-only.
However, often one can build a node setup, that only deforms and does nothing
else.
To make it possible to apply the Geometry Nodes modifier in such cases the
following things had to be done:
* Update `BKE_modifier_deform_verts` to work with modifiers that implement
`modify_geometry_set` instead of `deform_verts`.
* Add error handling for the case when `modify_geometry_set` does more than just
deformation.
* Allow the Geometry Nodes modifier to be applied through a multi-res modifier.
Two new utility types (`ArrayState` and `MeshTopologyState`) have been
introduced to allow for efficient and accurate checking whether the topology has
been modified. In common cases, they can detect that the topology has not been
changed in constant time, but they fall back to linear time checking if it's not
immediately obvious that the topology has not been changed.
This works with the example files from #98559 and #97603.
Pull Request: https://projects.blender.org/blender/blender/pulls/131904
This makes the `material_index` attribute built-in on `CurvesGeometry` which
means that it's built-in on `Curves` and `GreasePencil`. While only Grease
Pencil can use it so far during rendering, we already depend on them on curves
when converting between curves and Grease Pencil.
Making the attribute built-in implies that it is locked to one domain and type
(`Curve` domain and `int32` type in this case).
This makes curves more consistent with meshes.
Pull Request: https://projects.blender.org/blender/blender/pulls/133488
These geometry types don't work properly with attributes currently,
so the persistent base isn't really "persistent" and doesn't last after
exiting sculpt mode, but it's still useful for it to work within a
sculpt session. Enable that by adding temporary array storage in
`SculptSession`. During the sculpt refactor project I mistakenly
assumed that this didn't work well enough that anyone would use it.
Pull Request: https://projects.blender.org/blender/blender/pulls/133410
Since 4cf7fc3b3a (where a storage handling being added) a render layer node never has persist storage data.
And even more, some code of node update do use storage field as temporal variables for a callback so no any data can live forever in this place.
No one copy/initialize/free callback of the node also does not deal with node storage.
And compositor code also don't know about node storage since use different data from a node (id and custom1 fields).
Pull Request: https://projects.blender.org/blender/blender/pulls/132717
Color to grayscale conversions should take into account the colorspace,
and these are considered to be in scene linear colorspace.
No change with the default configuration.
Pull Request: https://projects.blender.org/blender/blender/pulls/133368
Both Image and Sequencer vectorscopes used YUV U,V coordinates
as-is, but their possible ranges are different: +-0.436 and
+-0.615 respectively.
It looks like more other software (ffmpeg, shotcut, etc.)
re-scales the vectorscope UV to fill up a square, i.e. streches
out the horizontal values, so do the same.
This fixes the "skin tone indicator" line, which at 123 degrees
(90 + 33 degrees to match positive I axis of YIQ color space) was
placed assuming the UV values fill up a square. So it was at the
wrong angle before.
The vectorscope horizontal scaling and skin tone indicator line now
matches other open source (ffmpeg, shotcut) and commercial (davinci
resolve, final cut pro) software packages.
Images in the PR.
Pull Request: https://projects.blender.org/blender/blender/pulls/133316
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
Previously, when evaluated on the face corner domain, the normal input
node just returned the face normals, as if the mesh was completely flat
shaded. This ignores face and edge smoothness, and custom face corner
normals. In the past couple years the expected behavior of accessing
normals has become much clearer and this behavior is clearly a mistake
in retrospect.
This commit exposes the same face corner normals used everywhere else
in Blender when the node is evaluated on the corner domain. The old
behavior is accessible with a node property in the sidebar. There is
versioning so old files have the property set and get the same results.
This is split from !132583.
Pull Request: https://projects.blender.org/blender/blender/pulls/133340
With an asset library configured in the Preferences that had an empty
path, the catalog tree of the "Current File" asset library would be
empty. That is because the path-less library would incorrectly be
considered as containing the current file's path (`BLI_path_contains()`
returned true for the empty path).
Before slotted actions, it was only useful to stash an action once on a
data-block: that was all that was needed to keep the action's channels
associated with the data-block, and stashing it multiple times was redundant.
Therefore Blender would simply refuse to stash an action if it was already
stashed.
However, now with slotted actions the user may legitimately want to stash an
action more than one time: if the action has multiple slots, the user may want
to keep the channels of multiple slots all associated with a data-block. Doing
that requires stashing each action+slot combination separately.
This PR addresses this use case by now allowing each action+slot *combo* to be
stashed once. Stashing a given action+slot combo more than once per data-block
is still prevented, however, since that is still redundant.
Pull Request: https://projects.blender.org/blender/blender/pulls/133307
Did a pass through the code base to make sure the uses of `Action.idroot`
are correct in the context of layered actions.
I found and fixed one issue in Grease Pencil v2 -> v3 conversion code
where the idroot was being set to a non-zero value without checking if
the action was layered or not first. Other than that, I also cleaned up
a couple of dodgy-looking things I found.
Pull Request: https://projects.blender.org/blender/blender/pulls/133205
This error would occur because the quadratic equation solving code didn't
check for `a` being non-zero. Fixed by adding the additional check. This
can be tested and reproduced using the `ShrinkwrapTargetNormalProject2`
`modifiers` test.
Pull Request: https://projects.blender.org/blender/blender/pulls/133273
Compilers do not seem to agree on valid printf format for `int64_t`,
recent clang 19 on linux requires `%ld`, while older compilers like the
buildbot ones of clang 15 on OSX ask for `%lld`.
So instead, cast the value to `int32_t`. Other solution may have been to
use `PRId64`, but this is fairly bad for readability.
Check done by PartiaWriteContext writing code to ensure there is no
library ID written which filepath is the same as the destination
blendfile path of the context, was flacky in case there would be
multiple library IDs with that same path.
While this is not expected situation currently, it will likely change in
the future, so handle that properly, and generate a CLOG warning.
The handling of library for IDs added to a PartiaWriteContext in 'make
local' mode (i.e. to make them local in the written blendfile) was
flacky, leading to invalid removal of the ID name from the library ID
namemap in G_MAIN.
Now simply esure there is a local copy of the library too when adding an
ID to the context, even if it will be made local there.
As established in issue #129849 , there are references to older versions of Grease Pencil that aren't needed.
This PR is meant to remove such references or, in places where otherwise `GreasePencil` is referenced
but `GPencil` is not, it changes the reference as for it to reference the v3 implementations in PaintMode.
It also removes `GreasePencil` from the list of options when declaring `PaintMode`.
Pull Request: https://projects.blender.org/blender/blender/pulls/131065
This uses the following accessor methods in more places in more places:
`is_group()`, `is_group_input()`, `is_group_output()`, `is_muted()`,
`is_frame()` and `is_reroute()`.
This results in simpler code and reduces the use of `bNode.type_legacy`.
Pull Request: https://projects.blender.org/blender/blender/pulls/132899
Using the `BKE_id_owner_get` during the node tree update can avoid having to
iterate over lots of data-blocks in common cases. Previously, the code had to
iterate over all potential owners of node group to find the correct one.
Pull Request: https://projects.blender.org/blender/blender/pulls/132903
Using `CLOG_INFO` and then reporting critical error is not very coherent,
nor practical, since INFO messages are not even displayed by default in
the console.
- Support a null "subfolder_name" argument
in keeping with the rest of BLI_appdir functions.
- Use "test_path" utility to de-duplicate path joinging & checks,
this also means the paths will be logged.
- De-duplicate loop body for the last item.
This patch adds support to apply a pose asset to multiple armatures in pose mode at once.
In case there is a suitable slot for the armature, it will be used to read from, otherwise
the system falls back to slot 0.
Part of #131840
Pull Request: https://projects.blender.org/blender/blender/pulls/132601
Regression in [0] that clamped the shape-key index when exiting
edit-mode but not when entering, causing the "Basis" shape not to
be treated as active when entering edit-mode, then treated as active
when exiting - overwriting it's values from the mesh.
Resolve by clamping on edit-mode enter & exit.
[0]: 225e3460d0
The hardcoded value doesn't work well with real scale human heads for
example (was already adjusted once in a76b5d3a07).
The result for too high values is a complete "freeze" of the whole curve
(since the solution from e7606139ba has the problem that it keeps
running into max iterations of the collision solver).
As long as no better solver is implemented, it is better to have an
adjustable value (to work on differently sizes objects) to not run into
the above issue (same as the old particle hair system had) and show it
in sculptmode next to the button which enables collision.
This is done per `Curves` (same as the flag
`CV_SCULPT_COLLISION_ENABLED`), similar to symmetry settings
[alternatively, it could be part of `BrushCurvesSculptSettings` but I
think it makes more sense in Curves] and then passed on to the
`CurvesConstraintSolver`.
Includes versioning code (to set the default for old files).
Pull Request: https://projects.blender.org/blender/blender/pulls/132997
* BLENDER_SYSTEM_SCRIPTS support for multiple script paths, separated by
; on Windows and : on other platforms.
* New BLENDER_CUSTOM_SPLASH to replace the splash screen image.
* New BLENDER_CUSTOM_SPLASH_BANNER to overlay an image on the splash.
Contributed by Sony Interactive Entertainment: https://github.com/PlayStation-OpenSource
`BKE_main_ensure_invariants` was added in 1fae5fd8f6. The older
`ED_node_tree_propagate_change` was already implemented as a thin wrapper around
`BKE_main_ensure_invariants`. This patch removes the wrapper and calls the more
general function directly.
A new overload of `BKE_main_ensure_invariants` is added for the common case when
only a single data-block has been modified.
Pull Request: https://projects.blender.org/blender/blender/pulls/133048
This makes specifying a legacy type for new node types optional (e.g.
`GEO_NODE_MESH_TO_CURVE`). Instead, only the `idname` is used as a stable
identifier for node types. So there is less redundancy for now.
This change helps with the decentralized definition of nodes and reduces the
number minimum number of files that need to be changed for a new node from 5 to
4. It especially helps when multiple nodes are worked on at the same time,
because the legacy type definition was very prone to have merge conflicts.
For compatibility reasons and because it's still used by existing code, the
`legacy_type` is not removed. All existing nodes keep their current
`legacy_type`. New nodes will receive an auto-incremented legacy type. It's
still necessary to give nodes unique legacy types, because some code checks if
two nodes have the same type by comparing their `legacy_type`. These types only
have to be unique at run-time though. Some randomness is used to avoid depending
on stable generated legacy types accidentally.
Pull Request: https://projects.blender.org/blender/blender/pulls/133044
The passed `StringRef` can be empty, so we shouldn't pass it to
`BLI_strdupn` in this case.
If an empty string is passed to either `set` function
makes sure the previous `char *` is freed and set to `nullptr`.
Pull Request: https://projects.blender.org/blender/blender/pulls/133042
ID names of images that included slashes would attempt to write
save the image in a sub-directory based on those slashes.
Also add notes to clarify the intended use of "safe" path functions.
Capitalize the default filename used for .blend files and other savable
and exportable file formats (like images, 3D formats, etc.) from
"untitled" to "Untitled".
Pull Request: https://projects.blender.org/blender/blender/pulls/132424
Previously, the owner-id on localized node trees, like the one created in
`GPU_material_from_nodetree` would have `bNodeTree::owner_id` set, even though
they were independent. The leads to asserts when using `BKE_id_owner_get` on the
localized node tree.
The `node_tree_localize` already took a `ID *new_owner_id` parameter. However,
if it was `nullptr` it was ignored instead actually clearing the `owner_id`.
This patch changes the parameter to `std::optional<ID *> new_owner_id`. Now,
`std::nullopt` means that the value is ignored and when `nullptr` is passed in,
the `owner_id` will be cleared on the localized tree.
Since `node_tree_localize` indirectly uses other functions, the same function
signature change is done in `BKE_libblock_copy_in_lib` and `BKE_id_copy_in_lib`.
The callers have been updated to pass in `nullopt` instead of `nullptr`.
Pull Request: https://projects.blender.org/blender/blender/pulls/133002
This patch adds a new `BKE_main_ensure_invariants` function. For now it only
ensures node-tree related invariants, but more may be added over time.
The already existing `ED_node_tree_propagate_change` now internally calls
`BKE_main_ensure_invariants`. We can probably remove this indirection at some
point and call the new function directly, but for now it is kept to keep this
patch small.
This is based on a recent discussion in the Core module meeting:
https://devtalk.blender.org/t/2024-12-12-core-meeting/38074
```cpp
/**
* Makes sure that invariants in original DNA data are maintained after changes.
*
* This function has to be idempotent, i.e. after calling it once, additional calls should not
* modify DNA data further. If it would, it would imply that this function does more than
* maintaining invariants.
*
* This has to be called after any kind of change to original DNA data that may be involved in some
* of the maintained invariants. It's possible to do multiple changes in a row and then fixing all
* invariants with a single call in the end. Obviously, the invariants are not maintained in the
* meantime then and functions relying on them might not work.
*
* If nothing is changed, this function does nothing and it should not be slower than checking a
* flag on every data-block in the given bmain.
*
* Sometimes, it is known that only a single or very few data-blocks have been changed (e.g. when a
* node has been inserted in a node tree). Passing in #modified_ids can speed up the function
* because it may avoid the need to iterate over all data-blocks to find modified data-blocks.
*
* Examples of maintained invariants:
* - Group nodes need to have the correct sockets based on the referenced node group.
* - The geometry nodes modifier needs to have the correct inputs based on the referenced group.
*/
void BKE_main_ensure_invariants(Main &bmain,
std::optional<blender::Span<ID *>> modified_ids = std::nullopt);
```
This also adds `windowmanager` as a dependency of `blenkernel` to be able to
send notifiers.
Pull Request: https://projects.blender.org/blender/blender/pulls/132023
In order to copy vertex group weights when subdividing strokes the
groups (`bDeformGroup`) must be created in advance, so that the attribute
wrappers correctly write to `CD_MDEFORMVERT` layer instead of creating simple
`CD_PROP_FLOAT` layers.
In addition the subdivision function must take care to fully write the
destination arrays, since initial values must be considered uninitialized (this
is obfuscated for simple CustomData arrays but breaks with more complex
attributes that use a buffer). In order to ensure fully defined destination
buffers without additional copies, the `finish` call to attribute writers is
postponed until the unselected attribute values have also been copied from input
buffers and all the values are properly defined.
Pull Request: https://projects.blender.org/blender/blender/pulls/132854
