When using Metal backend on ATI/Intel GPUs the stencil buffer doesn't
contain the correct information, skipping diffuse and glossy PBR components.
Failing behavior was introduced by e97e06ea2c
This PR works around this by reverting the code when using ATI/Intel GPUs.
The root cause is somewhere the combination of stencil export support, clear
operations and subpass.
**TODO**
- [ ] Validate that this fixes the ATI issue as well
Pull Request: https://projects.blender.org/blender/blender/pulls/122993
Having the sun extracted is mandatory to keep
the same look and avoid too much light
leaking compared to EEVEE-Legacy. But adding
shadows might create performance overhead and
change the result in a very different way.
So we disable shadows in older file.
This should not happen and any failure here should be considered a bug.
But for end users better not to hang Blender, and to have a better
diagnostic for developers in bug reports.
Ref #82483
Pull Request: https://projects.blender.org/blender/blender/pulls/123023
CTX_data_ensure_evaluated_depsgraph should not be used in drawing code.
This triggered a depsgraph updated which in turn called
DRW_notify_view_update. This would bind the GPU context a second time
and hang due to recursive mutex calls.
Issue wasn't directly related to material assets or the shader editor.
Simpler steps to reproduce:
- Open Asset Browser
- Change Asset Browser to different editor type
- Open new file (Ctrl+N)
The asset browser would remain in storage as inactive editor, including
pointers to the asset system. When opening a new file, the asset system
would get freed before the asset browser, which would then access
dangling pointers as part of its own freeing process.
Part of the issue is that `SpaceType.exit()` doesn't get called in this
case, which would remove the asset system references before the asset
system is freed. Will address this in a follow up in main, but best to
not depend on the `exit()` callback too much. Easy to do here.
Add the color for the new keytype 'generated' to the VSE theme settings
as well.
This includes the versioning code that loads the default from the scene.
Without this, the preference would be initialised to black.
Note that the 'Blender Light' theme does not have any colors for any of
the key types, so I also didn't add the 'generated' type there.
Pull Request: https://projects.blender.org/blender/blender/pulls/123005
The following code in py console would return a python string object,
instead of the expected bytes one.
```python
value = b"Hello World"
key = "a"
C.object[key] = value
C.object[key]
>>> b"Hello World"
C.object.path_resolve('["%s"]' % key)
>>> "Hello World"
```
Now it will return a byte object as it should.
Found while investigating #122843 .
- "Show the result of running commands in the main interface...": add
punctuation.
- "Re-Installed" -> "Reinstalled" seems more common.
- "increase_" and "decrease_subdivision" in the Grease Pencil
modal...: use title case instead of snake case for label.
- "Use a automatic number..." -> "an", typo.
- "... points,{} splines": missing space.
- "End Frame is larger than Start Frame": wrong order in error message.
- "Approximate the object as sphere, which diameter is equal to
thickness the defined...": grammar.
- "Show Seconds" -> "Use Timecode": this option is not necessarily about
seconds. Suggested by Pablo Vasquez.
- Brightness and contrast node input descriptions: add missing spaces.
- "Gamma controls the relative intensity... full white.": remove
trailing ".".
Pull Request: https://projects.blender.org/blender/blender/pulls/122972
Fixes#116567
The issue was with a flag that skips a particular post-evaluation step
specially for the Child Of constraint. The flag wasn't getting properly
unset when the constraint target was removed, and therefore that
post-evaluation step was getting erroneously skipped in some cases.
This commit fixes the issue by always setting the flag appropriately in
the Child Of evaluation function itself. This is admittedly rather
hacky, but no more hacky than the existence of the flag in the first
place.
Pull Request: https://projects.blender.org/blender/blender/pulls/122881
This patch makes sure that when both left/right handles on a strip are
selected, they are given individal snap points in TransSeqSnapData.
Prior to this change, when both handles were selected, snapping only
worked on the left handle.
Pull Request: https://projects.blender.org/blender/blender/pulls/122931
Snapping to hold offsets was broken by 76043bc, this patch restores the
functionality by simply calculating the image data start/end points and
letting existing clamping code take over to make sure this snap target
does not exist outside of the bounds of the strip.
Pull Request: https://projects.blender.org/blender/blender/pulls/122934
`BMVert.copy_from_vert_interp()`'s description and the name `vert_pair`
implies two verts, but the argument itself was described and typed as a
single `BMVert`.
Ref: !122974
Light volume probe world shader doesn't read from the irradiance atlas
but its create info registered it as a read/write texture. This
raised a validation error on Metal/Intel as the combination
RGBA16F FLOAT3D read write isn't supported.
Pull Request: https://projects.blender.org/blender/blender/pulls/122646
All add-ons were being scanned at startup, while this didn't cause
errors it was noticeable with extensions where any errors in the
manifest were being reported at startup, even when running with
factory-startup (including blender's own tests).
Address two issues:
- The logic to "reset" add-ons, so as to match the preferences when
reverting or resetting preferences always ran on startup.
This occurred because a check for Python being initialized was
incorrectly used to detect that this wasn't the first time preferences
were being loaded (regression in [0]).
- Resetting add-ons scanned all add-ons (including disabled add-ons) to
ensure their module cache is up to date. Since this the cache is
lazily initialized, it's simpler to set it as uninitialized as
resetting the add-ons doesn't require the cached meta-data.
[0]: 497bc4d199
Although c9fa73a379 made the problem worse. The incorrect to snap to
edges when one of their vertices is behind the viewer is much older.
The problem occurred because, in perspective mode, when we project a
vertex with negative "zfac", its projected value is undefined as it
depends on the position of the other vertex.
Thus, both the distance test and the lambda in these cases were wrong.
The solution was to calculate the snap point directly in 3D, using
`closest_ray_to_segment_v3` and avoiding projecting the edge.
This was caused by the base page tagging.
Since that the base page is now required for fixing
soft shadows, it would be the first level to be rendered
which, combined with the update throttle, lead to only
the min LOD to be available when the shadow was updated
in viewport.
This was affecting the jittered shadows, animation update
and viewport transform.
The fix is to only request the base page after throttling
the updates and only if there is no other LODs to be
rendered. This mean jittered shadow in viewport might
become affected by the missing base page artifacts, but
this is less noticeable than huge light leaking.
only if
The following operations have been removed for system repositories:
- Installing from disk.
- Uninstall extensions.
- Delete files (when removing the repository).
Also change the operator to remove a repository so the option to
remove files is now a boolean instead of an enum. While a dynamic
enum can be made to work, this option is logically a boolean.
The `Snap to Grid` and `Snap to Vertex` modes were being mixed if
either mode was enabled.
Support for mixed snap modes doesn't yet fully work for UVs.
When starting a navigation operation that uses `Auto Depth` or
`Orbit Around Selection` the value of `#RegionView3D::dist` is modified
to match the distance from the new pivot.
This value was incorrect when the pivot was behind the viewer.
In these cases `dist` is negative.
This commit, in addition to fixing `dist`, also limits the execution of
this code to only when `Auto Depth` is used.
Pull Request: https://projects.blender.org/blender/blender/pulls/122795
Allow precompiling specialization constants variations in parallel.
Only supported in OpenGL as the rest of the batch compilation API,
on the other backends the function is a no-op.
This also moves the `SpecializationConstant` from
`gpu_shader_create_info` (private API) into`GPU_common_types`
(public API).
Pull Request: https://projects.blender.org/blender/blender/pulls/122796
This changes fixes a wrong mesh as a result of apply base mesh from the Multires modifier
In the previous code some corner vertices where incorrectly assigned to null and skipped form processing as a safety guard
that was only relevant when the mesh had hidden faces in edit mode. The hidden corner vertices bug is handled now when checking
> if (corner_x_index < 0 || corner_y_index < 0) {
> continue;
> }
Since the patch that caused this error was a fix to a previous infinite loop bug,
I also tested and made sure this updated patch still fixes the original issue it was trying to fix.
Note: This PR replaces #122845 due to a large amount of git rebase issues.
Pull Request: https://projects.blender.org/blender/blender/pulls/122859
Previously selected strips in VSE timeline were drawing their outline
1px outside of the strip boundaries. This makes outlines of the strips
overlap each other when neighboring strips are selected.
Now the selected outline is fully within regular strip shape.
Pull Request: https://projects.blender.org/blender/blender/pulls/122890
The File Output node ignores color space overrides for EXR images. To
fix this, we save the images using save_as_render set to true. We don't
need to provide this as an option similar to other image types because
even when save_as_render is set to true, it will not have an effect
unless the user chooses to override the color space explicitly, since it
is not affected by view transforms and the like.
Pull Request: https://projects.blender.org/blender/blender/pulls/122791
The File Output node forces all inputs to have the same size, which
should only be the case for multilayer files. This is a regression in
931c188ce5. To fix this, we allow inputs to have any size, except for
multilayer files, which are realized on the automatic operation domain
of the operation.
Pull Request: https://projects.blender.org/blender/blender/pulls/122824
The CPU compositor does not recognize viewer groups inside node groups
unless a viewer node exists at the top level node group. This is caused
by bad logic when registering viewer node at the node operation builder.
And the fix is just to correct the logic to always register viewers if
no active viewer already exists, which can then later be overwritten if
a viewer node that takes precedence was discovered.
Pull Request: https://projects.blender.org/blender/blender/pulls/122869
Right now probe passes are compiled one by one, since passes are only
requested if `materials.queued_shaders_count` is 0 and requesting a
pass will increase the number.
This splits the logic into 2 functions, one for checking if sync is needed
and another to check if pass shaders are needed.
This allows compiling the shaders in parallel.
Pull Request: https://projects.blender.org/blender/blender/pulls/122799
PR #122576 added rounded corners to VSE timeline strips, but they were not
"snapped" to pixel grid so the outline that is normally 1px was sometimes
falling in between pixels and was blurred out.
Fix by rounding all SDF related coordinates inside the shader to the pixel
grid.
Pull Request: https://projects.blender.org/blender/blender/pulls/122764
