The root cause is that viewport can draw cached version of themself but
the scene can have been updated and the pointed curvemapping could have
been freed.
To workaround this we just keep a copy of the curvemap at the viewport
level.
- Change the default cache method to replay
- Change the default resolution to 32 (The same as old smoke)
which have a speedup of about 4x (~4 FPS vs. ~16 FPS on initial
playback)
Peformance was tested with 3700x and RTX 2070
Differential Revision: https://developer.blender.org/D6853
Fixes T73799
The Layer brush was in Blender before 2.81, when the sculpt API was
introduced. It had a huge amount of bugs and glitches which made it
almost unusable for anything but the most trivial cases. Also, it needed
some hacks in the code just to support the persistent base.
The brush was completely rewritten using the Sculpt API. It fulfills the
same use case as the old one, but it has:
- All previous artifacts fixed
- Simpler code
- Persistent base now works with multires thanks to the sculpt API
- Small cursor widget to preview the layer height
- More controllable and smoother strength and deformation
- More correct masking support
- More predictable invert support. When using persistent base, the brush invert mode resets to layer height 0, instead of jumping from +1 to -1. The brush can still be inverted in the brush direction property.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D7147
Issue was with setting of frame to load from an image sequence,
synchronization was not done properly at some point, leading to
generation of an invalid final filepath to be read.
Add method to invalidate strip cache in range of non-overlapping strip.
Invalidate original strip in range of new strip created by cutting.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D7313
`seq_free_animdata()` removes fcurve pointers belonging to strips from
`Scene` CoW datablock's `AnimData` during `BKE_scene_graph_update_for_newframe`.
This causes problems with updating animation.
This worked before rBbe2e41c397ba, because `AnimData` was freed by `BKE_animdata_free()`
before `seq_free_animdata()` was executed, so it had no data to operate on and returned
on precondition `if (scene->adt == NULL || scene->adt->action == NULL)`
Reviewed By: mont29, brecht
Maniphest Tasks: T74897
Differential Revision: https://developer.blender.org/D7264
Due to floating point approximations, the weights for interpolating
the mdeformvert layer could add up to a tiny bit more than 1.0.
This was not a problem in practice, but the mesh validation routine
used in regression tests was testing for this and therefore failing.
Just changed interpolation of mdeformverts to clamp max to 1.0f.
Logic to handle shepkeys datablocks in helper in 'make local' code that
checks which ID should be copied, and which can be directly made local,
was wrong.
One fewer coordinate needs to be calculated when extrusion is zero to
avoid corrupting the end of the memory chunk.
Differential Revision: https://developer.blender.org/D7368
Excluding a collection also changes the exclude setting on all child collections
so that it is possible to selectively enable some children without the parent
being enabled.
This change makes it so that if you unexclude the parent, the exclude setting
of children are restored again instead of being permanently lost.
Original patch by Szymon with modifications by Brecht.
Differential Revision: https://developer.blender.org/D7016
Tasks: move priority from task to task pool {rBf7c18df4f599fe39ffc914e645e504fcdbee8636}
Tasks: split task.c into task_pool.cc and task_iterator.c {rB4ada1d267749931ca934a74b14a82479bcaa92e0}
Differential Revision: https://developer.blender.org/D7385
rB10162d68e385 introduced a difference in computing the matrix, as well
as a better way to compute the inverse matrix. This commit reverts the
former, while keeping the latter and some other improvements.
Allows to open newer files in older Blender after new distortion model
has been added.
It will behave as if this is a polynomial model with all 0 coefficients
which are then being refined and assigned explicitly after solving the
motion.
Removal of 'Translation' checkbox. Enable translation options when selecting non-English languages.
Differential Revision: https://developer.blender.org/D7210
Reviewed by Brecht Van Lommel
These changes only have an effect when the experimental Undo Speedup preference
is enabled.
* For DEG_id_tag_update, accumulate recalc flags immediately before the undo
push happens instead of afterwards. Otherwise the undo state does not
contain enough flags, and the current state may contain too many flags.
This also means we call DEG_id_tag_update after undo with the accumulated
flags to ensure they are flushed to other datablocks.
* For undo, accumulate recalc flags in id->recalc and clear accumulated flags
immediately. Not clearing would cause circular behavior where accumulated
flags may never end up being cleared.
This matches what happens after an undo push where these are also cleared,
indicating that the undo state and current in-memory state match exactly.
* Don't change id->recalc of identical datablocks, it should not be needed.
There is one exception for armatures where pointers across datablocks
exist which otherwise would cause problems. There may be a better solution
to this but it seems to work in agent 327 production files.
* This contains a change in undofile.c to avoid detecting all datablocks as
changed for the first of the two undo steps, where we restore to the state
of the last undo push before going to the one before.
Without this the whole system is much less efficient. However this is unsafe
in the sense that if an app handler or operators edits a datablock after an
undo push, that change will not be undone.
It can be argued that this is acceptable behavior, since a following undo push
will include that change and this may already have unexpected side effects.
Ref T60695
Differential Revision: https://developer.blender.org/D7339
Issue was that the noise simulation was trying to bake with a minimized domain size (adaptive domain initializes domain with size (1,1,1)). Similarly to the base resolution bake, there should be no noise baking happening at those domain sizes - a domain at this size is considered empty.
The symmetrize operator now tries to make sure that the armature
constraints are correctly mirrored.
Before it would only mirror the subtargets for the constraints (and that
failed too in some cases).
Reviewed By: Sybren
Differential Revision: http://developer.blender.org/D6009
The root of the issue was caused by mesh which was a result of to_mesh()
had the same edit_mesh pointer as the input object, causing double-free
error.
This fix makes it so result mesh does not have edit mesh pointer.
Motivation part behind this is to make the result of to_mesh() to be
somewhat independent from the input.
Differential Revision: https://developer.blender.org/D7361
Need to communicate available number of threads to the camera
intrinsics implementation, otherwise default value of 1 is used.
Must have been single-threaded for a very long time.
When setting the current playback time in BKE_sound_play_scene we didn't
account for the frame length. So the current frame/time would be wrong
when we asked the audio playback what time it was.
This would lead to playback being offset when using time remapping and
AV sync.
Reviewed By: Richard Antalik and Sybren A. Stüvel
Differential Revision: http://developer.blender.org/D7248