Typically, we use this function to determine whether or not a subset of
data that has been collected with `gather` needs to be then persisted
into a larger array with `scatter`. As such, it makes sense to assert on
equality of the indices and smaller array size.
Pull Request: https://projects.blender.org/blender/blender/pulls/140752
We need to ensure that the compositor preview depsgraph gets updated
properly after undo, now that the compositing node tree is not embedded
in the scene anymore.
A test case will be committed separately in #140595
Pull Request: https://projects.blender.org/blender/blender/pulls/140596
This allows such functions' callbacks to receive the 'self' data as a
PointerRNA, instead of a pointer of the relevant DNA type.
This commit also adds some validation of flags when defining them for
a FunctionRNA.
Part of !135807.
Previously, Float2 images were internally stored as either Float3 or
Float4 images due to limitations in the implementation, which no longer
exists. So this patch refactors the compositor code to store Float2
images in actual Float2 containers, which gives better performance and
memory savings.
Some algorithms were adjusted to operate on Float2 instead of Float3 as
was previously the case.
Pull Request: https://projects.blender.org/blender/blender/pulls/140855
The code around view transform selection accidentally got too coupled
to the scene. However, the file output settings do not have scene, so
the enumerator getter was crashing.
Caused by 7ceb4495c5.
This patch brings the code closer to the Blender 4.4 behavior by
maintaining a global from view name to an ID.
There is still a bug in the code which displays views from the active
scene's display settings in the override panel. This is because the
itemf() callback is hardcoded to use the display settings from the
active scene (via bContext). However, this issue exists since the
initial commit of the color management override option.
Pull Request: https://projects.blender.org/blender/blender/pulls/140729
This removes the need to deal with void pointers and explicit casts in many
places. Also it simplifies possibly extending what kind of values can be stored
for each socket in the future.
Pull Request: https://projects.blender.org/blender/blender/pulls/140844
This new file can parse the file header (first few bytes) as well as the block
headers.
Right now, this is used by two places:
* `blendfile.py` which is used by `blend2json.py`
* `blend_render_info.py`
This new module is shipped with Blender because it's needed for
`blend_render_info.py` which is shipped with Blender too. This makes using it in
`blendfile.py` (which is not shipped with Blender) a bit more annoying. However,
this is already not ideal, because e.g. `blend2json` also has to add to
`sys.path` already to be able to import `blendfile.py`.
This new file could also be used by blender-asset-tracer (BAT).
The new `BlendFileHeader` and `BlockHeader` types may be subclassed by code
using it, because it wants to store additional derived data (`blendfile.py` and
BAT need this).
New tests have been added that check that the file and block header is parsed
correctly for different kinds of .blend files.
Pull Request: https://projects.blender.org/blender/blender/pulls/140341
The versionning was not inside `do_versions_after_linking_`
which made files created before 4.2 not switch back to
`BLENDER_EEVEE`.
# Conflicts:
# source/blender/blenloader/intern/versioning_500.cc
Previously, there was the issue that `SocketValueVariant` had a constructor that
took a forwarding reference as parameter. This was problematic, because this
could potentially hide copy/move constructors which is not intentional. This
patch makes sure that these special constructors are not overridden and adds two
static utility functions to make sure it's still straight forward to construct
the `SocketValueVariant` on a single line.
Clangd also warned about this case.
Pull Request: https://projects.blender.org/blender/blender/pulls/140842
When USD sends work to the GPU (dispatch) incorrect datatype conversions
can send incorrect vulkan commands. This happens on AMD GPUs. This PR
will patch USD to limit the max allowed limit. A better solution would
be to only use uint32_t in this function, but there is no `GfVec3u` data
type and I didn't want to do to many changes.
**TODO**
- [x] Validate with @ZedDB that this compiles and solves all the issues
- [x] Add other platform maintainers as reviewers. This is a Windows + Linux issue.
- [ ] Rebuild the USD library including this patch.
Pull Request: https://projects.blender.org/blender/blender/pulls/140102
This replaces API for accessing the uiLayout fixed_size,
red_alert, root_panel, search_weight and width properties
with methods, following uiLayout refactors and the
Python API naming
Pull Request: https://projects.blender.org/blender/blender/pulls/140673
When movie strip is added and redo panel is tweaked, operator failed to
re-execute.
Caused by c4eab49b9. This added `sequencer_generic_invoke_xy__internal`
call in operator exec function, but did not specify `SEQPROP_NOPATHS`
argument, which caused `directory` and `filepath` properties to be
overwritten.
This PR adds clarification for `SEQPROP_NOPATHS` use case, because it
isn't immediately obvious what this is doing.
Pull Request: https://projects.blender.org/blender/blender/pulls/140736
When the colored titlebar decoration style was implemented in
PR #123982, both the titlebar background and foreground were exposed
to the OS backends to be used when styling the Blender window titlebar.
In practice, both backends that implement this decoration style (Win32
and macOS/Cocoa) only use the background color, and they either rely on
the OS to automatically set the text color or use the color HSV value
component to switch between the OS dark/white color.
While it was thought to still keep this value around for a potential
future backend implementation, the color settings parsed to obtain the
titlebar color (`TH_BUTBACK_TEXT_HI`, which wasn't properly suited for
this feature to begin with) has been replaced by `TH_TEXT` in #140726,
which ends up being too bright to be used as titlebar text for most
themes/use cases.
As such, this PR removes this unused titlebar decoration style setting.
Future backends that wishes the implement the colored titlebar
decoration style should either use an existing OS/DE text color or use
a set white/dark text color, similarly to what the Cocoa backend does.
Pull Request: https://projects.blender.org/blender/blender/pulls/140823
Caused by changes in da4eda148b. Did not realized we have a few
duplicate blend-file names in `tests/files`, these ended up stepping on
each other's toes at random during testing.
Now ensure that the generated temporary 'save & reload' blend-file names
are unique, by adding a hash of the whole file path.
Previously in 308292b5d7 an error message
was added to notify that some selected objects are not valid for mesh
data transfer, that message didn't take account that src==dst is also a
case where the transfer is skipped, that doesn't mean the object is
invalid. So update the logic for this.
Pull Request: https://projects.blender.org/blender/blender/pulls/140463
Correct an oversight in df6d345bb4: without proportional editing, all
the transform data is "selected" by definition. This means that the
assumption "the selected data is sorted first in the array" is
trivially true, without calling any sorting function. So instead of
using the sorted index map in these cases, just fall back to regular
iteration.
To make the code handle this nicely, I made two "foreach" functions.
The first one transparently uses the sorted index map when available,
and performs regular iteration otherwise. The second function only
visits the selected items.
This makes the usage of these functions clearer, and the fact that
selected items are expected to be sorted first (either trivially or
explicitly) can be documented in a central place.
Pull Request: https://projects.blender.org/blender/blender/pulls/140720
This changes the naming of "Custom Split Normals" or "Split Normals" to be "Custom Normals"
This addresses #134744, only UI naming has been adjusted.
Internal/External APIs are left unchanged.
Pull Request: https://projects.blender.org/blender/blender/pulls/140440
Restore expected behavior of the NDOF, especially in the context of
non-3D editors. It addresses the following issues:
- Object mode, 3D viewport:
With "Lock Horizon" off, the rotation axes will invert unexpectedly,
making the camera behave similar to a "Fly mode".
- Fly mode, 2D editors:
Changing a navigation mode to "Fly" makes no effect.
Applies to UV and Nodes editors, Camera Preview etc.
Related to #140165
This includes some minor API changes back-ported from `main`
since the PR was created for `main`.
Co-authored-by: Patryk-Skowronski <patryk_skowronski@3dconnexion.com>
Ref !140537
