Age | Commit message (Collapse) | Author |
|
|
|
Render datablocks were copying state from original datablocks, where the frame
number and loaded volume grids were out of sync.
|
|
This adds support to the Link modifiers data. This was missing.
Also I did a small cleanup using LISTBASE_FOREACH macro.
Related to T76478
Reviewed By: campbellbarton
Differential Revision: https://developer.blender.org/D7643
|
|
With the Oculus runtime, the VR session would freeze when taking off the HMD
and putting it back on. This was caused by the deletion of graphics resources
too early in the OpenXR state machine, at least for Oculus.
The resources will now only be freed once the session is actually destroyed.
Also fixes an issue where it wasn't possible to stop the session via the UI
when the HMD was taken off.
Reviewed By: Julian Eisel
Differential Revision: https://developer.blender.org/D7635
|
|
Recent changes assumed OpenCL 2.0 platform. This adds a check to see if
we are compiling on an OpenCL 2.0 platform.
Patch was tested on:
* AMD Radeon Pro WX 7100 with amdgpu-pro-19.50-1011208-ubuntu-18.04 drivers
* AMD Vega 64 with amdgpu-pro-20.10-1048554-ubuntu-18.04 drivers
* AMD RX 5700 with amdgpu-pro-20.10-1048554-ubuntu-18.04 drivers
Reviewed By: Brecht van Lommel
Differential Revision: https://developer.blender.org/D7637
|
|
Caused by rBe07b245fe1f4 (new 'Select Linked' posemode op).
Above commit took CTRL+L (and this is kept for consistency with other
'Select Linked' operators).
So now keymap for poselib is:
poselib browse: ALT+L (was CTRL+L -- conflict here)
poselib add pose: SHIFT+L (same as before)
poselib remove pose: SHIFT+ALT+L (was ALT+L)
poselib rename pose: SHIFT+CTRL+L (same as before)
ref. D7542
|
|
Was flipping around the 0-1 range, now (optionally) flip around each tile.
Also added this option for BMesh bmo_mirror.
Reviewed By: campbellbarton
Maniphest Tasks: T75793
Differential Revision: https://developer.blender.org/D7460
|
|
pencil text
|
|
|
|
Differential Revision: https://developer.blender.org/D7630
|
|
Temporary fix that will skip cache invalidation from flow / effector objects.
|
|
objects
If we parent with type `PAR_ARMATURE` (where vertexgroups are already
set up and named correctly according to the corresponding bones), we
still need an armature modifier. This just wasnt added.
In contrast to meshes [which add their armature modifier early in
`ED_object_parent_set`], grease pencil used to do this (adding the
armature modifier) in `ED_gpencil_add_armature_weights`.
Now split ED_gpencil_add_armature_weights in two:
- ED_gpencil_add_armature
- ED_gpencil_add_armature_weights (which calls ED_gpencil_add_armature)
- use ED_gpencil_add_armature for the PAR_ARMATURE case
Maniphest Tasks: T76416
Differential Revision: https://developer.blender.org/D7625
|
|
If try to interpolate a layer without an active a next frame, the interpolation failed.
Now, check for both frames before start interpolation.
|
|
In 2.79, selecting an object would cycle past the first object
even if selection cycling wasn't in use.
Restore this behavior as it wasn't intentionally removed
and it's useful to be able to select an object behind the
current active object.
|
|
|
|
|
|
This implements the restore function for Draw Face Sets and Layer, which
don't affect coordinates or masks directly. This is needed for the
anchored and dot brush strokes.
Layer frees the current displacement and a new one is created on each
stroke sample. Draw Face Sets copies the data back from the first undo
node to the mesh datalayer.
Also fixes T75727
Reviewed By: jbakker
Maniphest Tasks: T75727
Differential Revision: https://developer.blender.org/D7442
|
|
Hide the nvrtc* symbols just like we did for cu* already.
|
|
|
|
It appears to work fine after a recent bugfix and testing for the past few
weeks.
|
|
|
|
Regression in own recent commit b1037aa88fedb.
|
|
Regression after recent changes. The precise cause is unclear to me, but
we do not need to update the size right after creating a new window.
|
|
This may fix or help diagnose T76378.
|
|
with OpenVDB in 2.83
Config files always use the .uni extenstion.
|
|
Thanks @LazyDodo for the help!
No functional changes.
|
|
Dependancy missing while building depsgraph for particle systems.
fix: adding a relation texture->particles when texture has animation data.
Reviewed By: sergey
Maniphest Tasks: T76251
Differential Revision: https://developer.blender.org/D7573
|
|
subdivide"
The fix was actually missing few bits:
- Firstly, the operator itself is not to do UNDO push
- Secondly, multires sculpt/top level are not pushed to the undo
node, so undo can not happen reliably.
It should be possible to incorporate some of the work from WIP patch
for propagation undo, but it needs more work.
Reverting code base to the previous state, since currently things
are actually a bit more confusing then they used to be.
This reverts commit 0c928087a3c8a4f758a907fe3096346f88d678c9.
|
|
Differential Revision: https://developer.blender.org/D7620
|
|
Differential Revision: https://developer.blender.org/D7523
|
|
We set the default buffer size for strict buffers to 0. This increased
the performance on intel but reduced the performance on old AMD cards.
This patch sets back the default buffer size for strict buffers. After
testing I didn't detect a slow-down anymore.
Note that today new firmwares were pushed to the device what might fix
some issues.
|
|
This fixes an issue where an animated modifier property that's used as
variable in a driver wouldn't animate that driver's value.
Building the relations for the driver target creates the relation
`PARAMETERS_EVAL` → `DRIVER(variable)`. Building the relations for the
FCurve targeting the modifier property creates the relation
`ANIMATION_EXIT` → `GEOMETRY_EVAL_INIT`.
This means that there is NOT a relation `ANIMATION_EXIT` →
`PARAMETERS_EVAL`, and as a result, the driver is not properly updated
when its variable reads animated data. This is resolved in this commit
by adding the missing relation.
Differential Revision: https://developer.blender.org/D7615
|
|
In previous commit the pressure was disabled by error.
|
|
This was introduced by error during refactor. The grid must not be visible in these modes.
|
|
During the refactor, the location of the grid was broken. Now, the grid is in Object origin or 3D cursor origin as it was in 2.82.
|
|
More small tweaks to get the right "feeling" when drawing.
The defaults has been tested by @pepeland.
|
|
|
|
With the previous commit, the angle effect was too subtle. Now the effect is more visible,
|
|
Add missing outliner selection sync tagging for various non-outliner
operators.
* Curve separate
* Grease Pencil separate
* Mesh separate
* Make instances real
* 3D view paste
* Sequencer paste
* Armature delete, dissolve, separate, duplicate, subdivide, extrude,
click extrude, primitive add
* Pose Group select, delete
Resolves T71404
|
|
This is a workaround, but a proper solution requires significant changes to
ray intersection in the kernel.
|
|
|
|
This may help with T68052, crashes with Intel NEO OpenCL driver.
|
|
|
|
Differential Revision: https://developer.blender.org/D7587
|
|
`BKE_sequencer_prefetch_get_original_sequence()` didn't look in metas
and returned NULL. This caused crash in disk cache that was trying to
read seq->name.
Add function that will look in meta strips recursively and condition
that seq must not be NULL.
Reviewed By: brecht
Maniphest Tasks: T76033
Differential Revision: https://developer.blender.org/D7597
|
|
An unintentional side-effect of rBfe7528ee919b was that when
extend-selecting a selected element in the outliner, it would be
deselected and activated rather than selected and activated.
This commit restores the expected toggling behavior. Consistent behavior
for extend-selecting child datablocks is not resolvable without a much
larger cleanup of the outliner select functions.
|
|
The interpolated points recalculated the angle again and this added noise. The angle must not be calculated and must use the previous thickness.
|
|
For any modifier, the expected output when the input mesh is empty, is an
empty mesh. So this error message was useless, and could spam the
console in some usecases of the modifier stack...
Reviewed By: weasel, mont29
Differential Revision: https://developer.blender.org/D7571
|
|
This is to improve the case of T71055 where curves share the same batch
cache when they shouldn't.
This however, does not help to fix edit mode display.
The real fix would be to have a similar handling to what the mesh modifiers
do and duplicate the whole Curve data. But this is too much work/change for
the 2.83 release.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D7569
|
|
Looping over all primitives for every object is really slow, so this patch avoids that by moving
the necessary assignments inline with the primitive merging done for every geometry.
|