| Age | Commit message (Collapse) | Author |
|---|
|
Basic idea is to split first loop in two, and run checks before making
anything actually local, to detect data-blocks that we can directly make
local (because we are sure they are only used by already/future local
datablocks).
This allows to avoid a lot of overhead in later 'cleanup' steps of this
function, here with barbershop shot it's four times quicker (from 190s to 48s).
We are still far from the instantaneous results of MakeLocal in 2.77,
but in that version main characters lose their connection to their
armature and remain static after makelocal, so guess new code is still
better. ;)
There are probably more optimizations possible here, but would rather
polish this area of code once we get rid of proxies, those really
make it a nightmare to work on.
|
|
|
|
This *should* work, but do not hesitate to revert in case it creates new
crashes in append or makelocal processes.
|
|
empty
If the opengl render with grease pencil is run from VSE with the current
frame outside visible frames, the render pass is wrong and the render
must be canceled because nothing to render. Related to #T49975
|
|
Before this commit, the brush set was created with the first stroke
drawing, but if the user creates the datablock or the layer manually
(not drawing) the brush list was empty.
This commit complement the python fix by Sergey:
https://developer.blender.org/rB89c1f9db37cc1becdd437fcfdb1877306cc2b329
|
|
Culprit here was once more proxies. Think what was happening here was:
1) Both proxy and proxified armatures' PoseChannels were cleared
(needed after remapping due to Bone pointers being stored in pchans).
2) Proxy PoseChannels got rebuilt in `BKE_pose_rebuild_ex()`, which ends,
in proxy cases, by actually replacing rebuilt pchans by those from
the proxified object... which has not yet been rebuilt.
Fixed the issue by merely adding bone pointer to data copied from
original pchan into new 'from proxy' one... Sounds much, much safer and
sanier anyway, that way we can be sure bone pointer is actually pointing
to a bone of the object's armature (this is supposed to be the same
Armature datablock between proxy and proxified objects, but that may not
be always true especially during makelocal process).
|
|
Did similar trick to old dependency graph: tag invisible relations for update.
Might need some re-consideration, see the comment.
This should solve our issues with powerlib addon here in the studio.
|
|
|
|
Was causing relations be build twice in certain cases.
|
|
New dependency graph expects strict separation between nodes and relations builder,
meaning, if we try to create relation with an object which is not in the graph yet
we'll have an error in depsgraph.
Now, so far object nodes were created from bases of the current scene, which caused
missing objects in graph in certain cases.
Didn't find better approach than to simply ensure object nodes exists when we know
they'll be used by relation builder.
|
|
|
|
Might have caused nodes created multiple times for the same object.
|
|
|
|
Solves errors, but not sure interface is indeed what users will expect.
Will ask GP team to check on this.
|
|
|
|
|
|
|
|
`kernel_path.h` and `kernel_path_branched.h` have a lot of conditional code and
it was kind of hard to tell what code belonged to which directive. Should be
easier to read now.
|
|
It was annoying to only have regular Enter confirming input there.
|
|
Proxified objects can never be local, we can totally ignore them here.
This 'fixes' the asserts related to usercount when trying to remap poselib
of localized proxified objects (not sure what exactly was going on wrong here,
but proxies are a giant can of worms for sane data-blocks handling anyway :/).
|
|
Similar to a previous commit.
Doing separately for an easy of bisect.
|
|
Copy paste error...
How to avoid those?
|
|
Solution as suggested by Sergey Sharybin. Initial debugging by
Antonio Vazquez.
|
|
anyway even if we told cmake we want release
|
|
anyway even if we told cmake we want release
|
|
This new `bpy.types.ID.make_local(clear_proxies=True)` allows Python
code to press the "Make Local" button on any ID block. I chose
`clear_proxies=True` as the default, since it's the default behaviour
of `id_make_local()` (defined in `library.c`).
The caller does need to take care of ensuring that linked-in objects
don't refer to local data, and that proxies aren't broken.
Reviewers: sergey, mont29
Reviewed By: mont29
Subscribers: dfelinto
Differential Revision: https://developer.blender.org/D2346
|
|
|
|
|
|
Keep ifdef'd out for holes, this isn't needed currently.
|
|
For the new dependency graph we have to rebuild graph when
bones "topology" are changing.
|
|
|
|
|
|
Reverse loops in-place.
|
|
|
|
This is required since new dependency graph evaluates drivers in threads
so it was possible to have some partially written ID tag there.
|
|
|
|
Replace search with direct lookup.
|
|
|
|
Add subframe to the animated seed hash calculation.
Should be no difference for the regular files, only for cases when scene is
rendered from sequencer with a speed effect, which is not really a common thing.
|
|
This is a late follow-up commit to the light sample threshold changes which
caused difference in rendering all existing .blend files which is not something
we are happy about: it is fine to use new optimized defaults for new files, but
existing ones should always be rendering in the same way as they used to be.
Sorry for the inconveniece, but such thing should have been done to begin with.
If this setting was modified it will not be reset to zero.
Now all render tests should be passing again.
P.S. Also really annoying to bump subversion for such reasons, but currently we
don't have better way to achieve what we want.
|
|
Lamp Data node requires shadow sample array which is only enabled when
Shadows are enabled in the shading settings.
This commit prevents crash but might not give expected render results
in such a configuration.
|
|
Would always set both first1 and first2.
|
|
|
|
|
|
Correct unhelpful comment & some comment edits.
Rename 'disk_is_flagged' -> 'bm_vert_is_manifold_flagged',
since the check is quite specific.
|
|
|
|
After CUDA dynload changes having CUDA toolkit became required
in order to compile Cycles. This only happened due to wrong
default value to the option.
|
|
|
|
This was included in the commit by accident, it doesn't belong there.
|
|
|
