(TODO: identify the reason why such a strange WVertex is generated.)
Problem report by Vicente Carro with a .blend for reproducing the issue. Thanks a lot!
Was silly mistake from rigidbody merge, base was used after it's been
freed.
Now don't free base in BKE_scene_base_remove() and rename it to
BKE_scene_base_unlink().
MEM_recallocN() doesn't allocate memory when used on a null pointer.
Just revert commit since there is no real benefit to using
MEM_recallocN() in this case.
previous commit (54102) actually reintroduces an old bug where Blender sigfaults when
the sensor and controllers are not from the active object.
The real fix for report #33746 is to clear the "object" property after the operator ran.
I'm not sure why when I call the operator from command line the property is cleared, but not
when I called it from C. Either way all should be working now.
Also now display nice "area" units (maybe using "length" units was a perf matter, but anyway, you can't have more than a few tens of values displayed at a time, after that they become unreadable). Easy to undo anyway if we really want to keep ugly "10m" as area display!
Transparent region drawing and blend broke by commit 53919 5 days ago.
This commit reverts the change.
The claim in previous commit "edited code for readability" is quite
disputable :) The error is hard to notice even.
I also like to emphasize that people should check with owners for code
before committing changes! Cleaning code is first a job for maintainers.
this patch optimizes the dilate/erode step method (hopefully without any functional change),
making its speed not depend on the distance anymore.
Couldn't detect funtional changes so committing. Haven't tested for speed gain.
* credits to erwin94 David M
Now all add/remove operators for rigid body objects and constraints
automatically add objects to the appropriate groups and create groups if they
don't exist yet.
This makes handling rigid bodies easier but doesn't take away functionality.
If users want to handle groups manually they just need to create them before
adding any objects.
The previous behaviour was confusing and was even considered to be a bug since
clicking on rigid body in the physics tab seemed to do nothing.
This is useful for gcc which does not define sqrtf/powf/... functions with preprocessor and therefore always used sqrt/pow/...
Float functions are generally 20-50% faster than their equivalents for double type.
Issue was caused by FILE multilayer ImBuf sharing buffers with render
result, and SEQUENCE multilayer ImBufs duplicating buffers. Which is
nice by it's own. But, changing image source wouldn't remove any loaded
image buffers, meaning if you've got loaded FILE multilayers they'll
likely became invalid.
That behavior of handling multilayers on changing source was done as
a fix for #24976, which is now actually not needed (removing check
for multilayer doesn't change behavior at all).
Just to be sure added check to RNA, so signal wouldn't be fired if
source wasn't actually changed form a menu.
Problem is that preview render database is being used, but the new datablock
function still uses G.main, and the mesh was being added to one database but
(attempted to be) removed from another. The proper fix would get rid of the
globals here, will look into that later.
- Cycles materials now render in Blender Internal too, skipping the nodes.
Not very useful, but at least things then show up on renders and in
previews.
- Node editor: if wrong shader nodes are in a tree, they draw with thene
color RED ALERT headers now. (Switching render engine will show it).
Two issues in texture_changed:
- Missing NULL check for material (slot could be empty)
- Materials could be linked to object, added special check for this
Issue was caused by the fact that guarded allocator is not thread-safe and
generated images/movies could allocate memory when loading pixels to Cycles.
Currently solved by switching memory allocator to using mutex lock (the same
as sued for jobs) when viewport rendering is used.
Nicer solution would be to make guarded allocator thread-safe by using atomic
operations and lock-free lists, but that's more serious change.
non-threadsafe usage of guarded allocator.
Also added small chunk of code to check consistency of begin/end
threaded malloc.
All this additional checks are commented and wouldn't affect on
builds, however found them helpful to troubleshoot issues so
decided to commit it to SVN.
This will be enabled if the file was saved with dynamic topology on,
but we don't automatically re-enter dynamic-topology mode when loading
a file so remove the flag.
Fixes bug [#33956]
projects.blender.org/tracker/?func=detail&aid=33956&group_id=9&atid=498
This means the deformation on the input to the modifier can be re-applied ontop of the mesh cache.
In practice this is most useful for using corrective shape-keys with mesh-cache.
BKE_rigidbody_aftertrans_update() can be called before rigid body is
validated so check before updating physics_object.
Thanks to Antony Riakiotakis (psy-fi) for the initial patch.
For example, if you're making a chain of bones (e.g. for a spine) by extruding
the tip joint of an initial bone, the bone rolls would be: 0 (for the initial
bone), 180, -180, 180, -180, etc. This has the undesirable effect of causing
B-Bones to twist to match the roll values at the other end of the bone.
The fix here seems to improve the situation in this case: bone roll values don't
flip or change anymore (in fact, the bone axes stay perfectly aligned now, as
they should). It also doesn't seem to cause any problems in other common cases I
checked.
Added export of multiple UV coordinates and vertex colour attributes.
A debugging option to export the strands without using the cache has also been removed.
