That was not really a bug (code working as expected), but the way tex context was handled was a bit raw, now it is much smarter:
* Default fallback context (when current one is no more valid) will now choose "most specific" ones first (i.e. material/lamp/particules before world and "others").
* When using that default fallback context, previous one is stored and we try to revive it later, if possible. Thus e.g. object[mat tex ctxt] -> empty[default world ctxt] -> object[mat tex ctxt] is now working as expected.
* However, when user explicitely or implicitely (through e.g. going to Material context...) sets a tex context, previous one is not stored, so that only default fallback context switch may later automatically revive a previous (presumably user-set) context.
calloc. Since we copy the first 1/1.3 part of the new array from the
existing nodes, only the rest 0.3/1.3 should be initialized to zero.
This should in theory cut down the times of occasional hangs with
dyntopo, since my guess is that it is caused by dynamic reallocations.
Maybe a linked list structure would help here? This is a bigger change
though, leaving as is for now.
Also, minor cleanup, delete duplicate ghash deletion and remove unneeded
commented code.
Topic says it all :)
Jut implemented operator to duplicate mask points
and segments between them (exactly the same behavior
as Curve object duplication in edit mode).
Does not copy animation, but that's tricky and likely
not needed anyway.
That ended up in tricky code trying to mimic depsgraph
branch behavior API-wise preserving texspace and bound
box calculation compatible with previous releases.
So for now bring cu->disp back to the trunk but keep
texpsace and boundbox APIs the same as in the branch.
This keeps texpsapce and boundbox behavior fully compatible
with previous releases and still makes API the same as
for meshes.
This mimics the behaviour we have in the Clip Editor.
I personally would prefer if we had no border once in fullscreen
(current border is 5 pixels).
I will consult Sergey Sharybin first to see if we can change that in the clip editor as well (though there I
believe the border is useful - the bottom of the editor is used to indicate 'tracked' frames.
before that if you pressed 'F', fit_view would be True every time (e.g.,
even if you pressed 'HOME' which is supposed to call the operator
without the fit_view).
I'm not sure why we need a fullscreen without 'fit_view' but anyways,
that's a different issue.
Simply recalc sequence len for audio (and meta!) strips when modifying fps value. Note start, startofs and endofs are also updated, to try to keep final pos and length as consistent as possible.
The reason of this is because PATH_MAX is not guaranteed
to be defined on all platforms and Hurd doesn't define it.
So either we need to support arbitrary long file path or
we need to define own maximum path length.
The rule here would be:
- If it's not big trouble to support arbitrary long paths
(i.e. in ghost by using std::string instead of char*)
then arbitrary long path shall be implemented.
- For other cases to use PATH_MAX please include BLI_fileops.h
which takes care of making sure PATH_MAX is defined.
Additional change: get rid of own changes made yesterday
which were supposed to make storage.c work fine in cases
PATH_MAX is not define, but on the second though it lead
to unneeded complication of the code.
Thanks Campbell for review!
gpencil_data_get_active and gpencil_data_get_active_v3d did not have consistent behavior when we had an active object, but not on any visible layer (the first would return the default scene gpd in this case, while the first was still returning active object's one). Now they both return scene's one.
initialization. In sculpt mode, when using shift-click to switch to the
smooth brush it was possible to ommit initialization of curvemapping for
that brush.
* Actually check if normal recalculation flags exists and set it when
splitting nodes in dyntopo. Right now, nodes that need GPU buffer update
will always get in the list to get their normals generated, but to avoid
a possible future breakage better do it right now.
* Avoid keeping deleted/removed vertices in vertex-to-node and unique
vertex hashes, since some checks rely on those and may go awry if these
still exist. Also they pollute the hashes, and may hurt performance
somewhat.
It's a bit dumb to store render-time bevel list in
object's curve_cache, but that's how blender already
used to work for ages.
Proper fix is suspended for tomorrow :)
Turn off pbvh normal update flag after recalculation, saves
recalculating normals every frame when not stroking the mesh.
For this to work reliably with undo we need to support original normals
in the bm_log (was marked as a TODO already in the code), so that
undoing avoids having invalid normals in the mesh (since we don't update
every frame anymore). This was added in this commit as well.
Also added some (disabled) quite paranoid checks in the bmesh valication
code for dyntopo hoping to catch the real normal update issue. No luck
there yet.
These are not animatable! Note this is the case of most (all?) render settings, maybe we should go over both Cycles and internal ones, there are still quite a bunch of them that are marked as animatable... :/
