| Age | Commit message (Collapse) | Author |
|---|
|
Makes the following layout changes possible:
{F8473498} {F8473499} {F8473502}
The next commit will contain many layout changes to make good use of
these new possibilities. The result should be more consistent, easier to
read and should give a more organized impression. Additionally, it
should be possible to replace many sub-panels with compacter layouts.
Main changes:
* Checkboxes now respect the property split layouts
* Add support for row and column headers (i.e.
`uiLayout.column(heading="Foo")`, `uiLayout.row(heading="Bar")`). If the
first property added to this layout doesn't insert anything into the label
split column, the heading is inserted there. Otherwise, it's inserted as own
item.
* Add support for manually inserting decorators for an existing item
(`uiLayout.prop_decorator()`). That way layout creators can manually insert
this, which was the only way I saw to support property split layouts with a
checkbox before the actual property. {F8471883}
* Autogenerated layouts for operator properties look bad if there are only
checkboxes (which only use half the region width). So before creating the
layout, we iterate over visible properties and disable split layout if all
are booleans. I think this is fine, if needed we could also add layout hints
to operators.
* `uiTemplateOperatorPropertyButs()` now handles macros itself, the caller
used to be responsible for this. Code that didn't handle these so far never
used macros I think, so this change should be invisible.
* Remove manual property split layout from autogenerated operator properties
layout.
* Padding of checkboxes is tweaked to make their label visually more connected
to the checkboxes.
* Support split layout for menus (should work for `uiLayout.menu()`,
`.operator_menu_enum()`, `.prop_menu_enum()`, maybe more)
Maniphest Task: https://developer.blender.org/T65965
Differential Revision: https://developer.blender.org/D7427
Reviewed by: Brecht Van Lommel, William Reynish, Pablo Vazques
|
|
|
|
Scene audio volume changes require the scene to be tagged with
`ID_RECALC_AUDIO_VOLUME` (see `BKE_scene_update_sound()`). Tagging
happens in the RNA update function `rna_Scene_volume_update()`, but that
function is not called by the animation system. As a result, animated
volume changes are not sent to the audio system.
This commit adds a new depsgraph operation node that sets this tag when
necessary, so that the animated values are used in the rest of the
depsgraph evaluation.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D7429
|
|
In preparation of TBB we need to split the finalize function into reduce
and free. Reduce is used to combine results and free for freeing any
allocated memory.
The reduce function is called to join user data chunk into another, to reduce the
result to the original userdata_chunk memory. These functions should have no side
effects so that they can be run on any thread.
The free functions should free data created during execution (TaskParallelRangeFunc).
Original patch by Brecht van Lommel
{rB61f49db843cf5095203112226ae386f301be1e1a}.
Reviewed By: Brecht van Lommel, Bastien Montagne
Differential Revision: https://developer.blender.org/D7394
|
|
|
|
Batch files did not work when you ran them from a
different working directory.
|
|
Conflicts:
source/blender/makesdna/DNA_userdef_types.h
source/blender/makesrna/intern/rna_userdef.c
|
|
Note that given how experimental is working currently, I had to rename
and inverse the effect of the experimental undo flag, which will now
instead activate legacy code when set.
|
|
|
|
This would double draw the edit cage on top of the other overlays.
|
|
Was caused by uninitialized buffer.
|
|
To make it clear that's not in the 0..1, but more of a terrain height value
without a strict range.
|
|
|
|
|
|
'hash_b' was always zero as it's range was bit-shifted away.
|
|
|
|
|
|
|
|
Holding Ctrl while dragging the light point gizmo would use
uninitialized stack memory if the normal had not been
initialized by a surface.
Now holding Ctrl can be used to drag,
even when there is no surface to orient to.
|
|
|
|
|
|
|
|
The dopesheet needs to have a frame to display the channel, so an empty frame is created in the current frame.
See T66505 for details of why an empty channel cannot be displayed.
|
|
Converting menus to PIE's was removing convenient key accelerators.
|
|
|
|
- Allow 0.0..1.0 range, as even at 0.0 the selection is still visible.
- Correct versioning code, not to overwrite the value for new files.
|
|
This has been removed since clang-format now enforces code-style.
|
|
|
|
|
|
The layout of the new sequencer disk cache were not handled well with
large preference windows.
|
|
The dopesheet needs to have a frame to display the channel, so an empty frame is created in the current frame.
See T66505 for details of why an empty channel cannot be displayed.
|
|
Although we still dynamically hide scrollbars, they don't change the
region size anymore. They are simply drawn on top of the region content.
Because of this, some hacks introduced by fa28e50ac2a7 are no longer
necessary.
Without these hacks, the scrollbar visibility is evaluated much more
often (cheap operation) which should be more reliable and possibly solve
some glitches.
Also replaces integers passed as booleans.
Fixes T75782.
|
|
|
|
LIB_TAG_EXTRAUSER_SET flag
For example in the Image Editor, an assert would be triggered after
unlinking an image [with setting users to zero] and then setting the
image for the Image Editor again.
Whenever we set an Image for Image Editor, the Image ID is flagged
LIB_TAG_EXTRAUSER_SET, when we unlink [with setting users to zero] this
flag was not cleared.
quote @mont29: "a proper fix would be to move this to modern code, and
actually delete the ID..." but that is for later.
Maniphest Tasks: T75675
Differential Revision: https://developer.blender.org/D7452
|
|
|
|
This editor's code was a bit schizophrenic, some parts considering its
nodetree usages as real refcounted ones, others, as shallow 'user one'
ones...
Editors should not be real ID users anyway, unless there are *very* good
reasons for it, so swich it to fully 'shallow' usage now.
|
|
|
|
Caused by framebuffer initialized in the wrong context.
|
|
|
|
While code is supposed to handle gracefully invalid override operations,
it is much cleaner to avoid those completely.
|
|
|
|
operation.
|
|
While this should not happen, we still want to handle those errors
gracefully from user perspective (i.e. assert for devs, no crash for
users).
Actual fix of root cause of the issue will come later.
|
|
glAttributes also include `gl_` names. These don't have a location and
should be ignored during shader interface creation. Those internal names
received a location of -1 and therefore the bitmasking was undefined.
Users wouldn't notice this, but ASAN warned developers of this situation.
ASAN could quit making ASAN un-usable as most shaders have this issue.
Reviewed By: Clément Foucault`
Differential Revision: https://developer.blender.org/D7448
|
|
This goes along with the existing changes to ignore PYTHONPATH by default.
--python-use-system-env now controls both.
Differential Revision: https://developer.blender.org/D6962
|
|
|
|
The patching of brushes was not done.
|
|
|
|
|
|
This crashes with ASAN enabled.
```
==39366==ERROR: AddressSanitizer: memcpy-param-overlap: memory ranges [0x6230000ae848,0x6230000ae85a) and [0x6230000ae851, 0x6230000ae863) overlap
```
|
