In case the process creashes, the prints about blendfiles being
processed could fail to be captured by the test framework.
And split these tests in 32 slices now, 8 was becomming way too slow to
complete for each test.
These have been removed since 4.0, this is not a pertinent information
to single out anymore.
Also downgrade report about missing obdata to `WARNING`, `ERROR` is a
bit strong for this.
Armature::collection_array was indexed with -1,
returning an invalid collection.
When the collection index isn't usable, create a new collection
with the default name.
Ref !129608
This reverts commit 0a70360eb6. That's
because it caused issue #129366 which is much more serious, and the
cause is not obvious so far. Another fix will be submitted for the
original issue.
When using AMD official driver clearing depth attachments can fail.
I assume it is related to previous pipeline states and dynamic rendering
that don't require the correct pipeline state for clearing depth
attachments.
Co-authored-by: jeroen@blender.org <Jeroen Bakker>
Pull Request: https://projects.blender.org/blender/blender/pulls/129852
`set_active_node` call from `grease_pencil_copy_data` resets inactive
layer again as locked before executing the drawing code. Since property
is only operated on layers, it is possible to remove the code from
`set_active_node()` and only have it in `set_active_layer()`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129465
This was likely never an issue so far, as no ID type has ever been
removed (that was written in blendfiles at least), and no new BHead type
has been added since ages.
However, current behavior of the main loop in the readfile code (in
`blo_read_file_internal`) is to consider any BHead type which is not
explicitely known and handled, to be an ID BHead, and directly try to
load it as such.
This commit addresses that potential issue by validating that a BHead
code actually matches a known ID type before trying to use it to read an
ID from the file.
Issue identified while working on designs for future 5.0 forward
compatibility breaking changes in blendfile format itself.
Pull Request: https://projects.blender.org/blender/blender/pulls/129693
Previously the `W` menu was only available for brush tools (brush and
eraser), it should be available to all tools, thus putting the key map
registering function into `km_grease_pencil_paint_mode`.
Pull Request: https://projects.blender.org/blender/blender/pulls/129837
The sample index node specifies the supported geometry types and Grease
Pencil was missing. Note that only the layer domain will work since it's
the only domain accessible by index on GP data directly.
Pull Request: https://projects.blender.org/blender/blender/pulls/129758
When doing action baking, the option "Clear Constraints" mentions
it's doing visual keying while not actually using that option.
Just fixing the description, even though a better solution
would be to revisit the design of the options to see if they
even make sense in some configurations.
Pull Request: https://projects.blender.org/blender/blender/pulls/129052
The issue is that the action line may be there, but containing no action,
`bAnimListElem->data` pointer is a nullptr.
caused by b952782a44
In this refactor the function call was changed from
`BKE_action_frame_range_get` which could handle
the case of a nullptr passed to it.
Pull Request: https://projects.blender.org/blender/blender/pulls/129648
Vulkan backend selection will be hidden behind developer extras. There
are some issues which would not be inviting for users to start testing.
- Some basic functionality are not working (object selection, depth
testing)
- NVIDIA UI artifacts are annoying and unsure where the error is.
- AMD artifacts
- Memory leaks
The backend will still be part of Blender 4.3 only hidden behind an
additional flag.
Ref: #129265, #129262, #127225, #128624, #127768, #127223
Pull Request: https://projects.blender.org/blender/blender/pulls/129799
Qualcomm driver can not handle the optimized SPIR-V that Blender
creates. According to their driver engineers this is an issue in
shaderc. As we are late in the release process they asked us to disable
the optimizations.
In Blender 4.4 we are planning to update shaderc what might fix the
issue. This should be retested after the update.
Pull Request: https://projects.blender.org/blender/blender/pulls/129775
RNA update function called from `rna_property_override_apply_default`
does not have access to active `Scene`, so it is explicitly set to null.
Use `ptr->owner_id` instead of scene provided in argument.
Pull Request: https://projects.blender.org/blender/blender/pulls/129562
In March 2022 this was added as a provision to check/guard the build if
external folks were compiling without USD "Imaging" support [1]
In Feb 2023 code was added which ended up depending on "Imaging"
unconditionally (and probably more at this point) [2]
Since we haven't heard this being a problem so far, and because the USD
build option is enabled by default, let's just remove it.
The test which was added is also removed because the main import/export
code serves as the compile test, and the python test code serves as a
much more sufficient check of actual functionality.
[1] https://archive.blender.org/developer/D14456
[2] Commit `72a85d976a5781a21166356b24668b8c48d51690`
Pull Request: https://projects.blender.org/blender/blender/pulls/129733
Interpolating these attributes as integer values isn't meaningful or
helpful and is potentially problematic. So far I'd guess this is unlikely
to happen in practice which is why it probably hasn't been noticed yet.
Fixes part of #129691.
Pull Request: https://projects.blender.org/blender/blender/pulls/129809
Only retrieve a mutable copy of the attribute if we're actually able to change it.
If topology changes and there are no IDs, we can't mix the attribute and we
should avoid retrieving it in case it's shared. This is more of a hypothetical
change, I didn't actually observe a real world performance change.
Pull Request: https://projects.blender.org/blender/blender/pulls/129811
When using the Voxel Remesh operator, there are two possible sources of
crashes currently:
* TBB task stealing in nested parallel loops with thread local data
* Null dereference of PBVH data when pushing sculpt undo steps
To fix the former issue, this commit guards the
`threading::parallel_for` internal function with
`threading::isolate_task` to prevent possible task stealing and
corruption of the thread local data.
Additionally, it has the effect of fixing debug asserts inside
`array_utils::gather` due to the this task stealing.
To fix the latter issue, this commit adds a call to
`BKE_sculpt_update_object_for_edit` to ensure that this data is
populated.
Pull Request: https://projects.blender.org/blender/blender/pulls/129704
The issue is caused by the new way to compute the projection
matrix that tries to recreate it from view3d properties.
This is needed for proper overscan support.
However, this breaks for VR as the view3d is only partially
setup.
Skip the setup in VR and remove overscan for VR to avoid any
possible issues with it.
This fix avoid changing the behavior of the
`ED_view3d_draw_offscreen` which is exposed from our pyGPU
API. The pyGPU API is kind of broken if using custom
projection matrix + overscan, but that's a different issue.
Fix#125456
Pull Request: https://projects.blender.org/blender/blender/pulls/129813
