Blender Git Loki

Fix triangulating concave quads

Resulting triangles could be pointing in opposing directions.

Fix: Never write previews for undo steps.

This may eat quite q bit of mem in the end (though nothing critical),
and icons (material ones e.g.) often ended up out of sync after undo.

Fix T42927: Triangulate-beauty fails on quads

Regressions since 2.69.

Docs: comments (hash table & beauty fill)

Fix T42917 shadow maps not working on ATIs.

This is yet another issue with framebuffers. There are two issues: We
need the framebuffer fully bound to check for completeness and when we
bind a depth texture as frame buffer we need to disable read/write.

Fix typo in VBO

Fix nasty typo in CPP RNA that would break on int or bool dynarrays...

Fix T42888: Separate and Combine HSV distorts the hue value

These nodes were assuming sRGB input/output which is for sure wrong for the
shader pipeline which works in the linear space.

So now conversion to/from linear space happens in these nodes which makes them
making sence in the shader context but which might change look and feel of
existing scenes.

Remove NULL checking variable after using it.

Fix T42864: Knife-project fails with 'cut-through'

Removing edge-hit-doubles could incorrectly de-duplicate intersections between different edges.

Gave noticeable errors cutting through geometry that overlapped exactly.

Fix for clip proxies builder reloading original footage frames

This isn't needed if the source is the image sequence.

Improve tip for simple deform origin

Fix T42864 (partial): knife-project included uncut backfaces

When knife-project ran with cut-through disabled,
it would still try to select faces behind the cut.

Now check if the faces are obscured.

Fix T42864 (partial): Knife-project had too-low precision

Fix missing UI ID messages and/or I18N contexts for MovieClip/Mask/Palette/PaintCurve ID types.

Tracking: Fix wrong logic in tracks synchronization

There was some stupidness in the way how tracks are synchronized from the job
to actual DNA data leading to all sort of weird and wonderful failures again.

Tracking: Avoid possible treading issue reading the frame to sync

Tracking: Fix possible race condition accessing the tracks

Writing to the tracks was already inside the lock section, but
reading was not. This could have lead to race condition leading
to all sorts of weird and wonderful artifacts.

IME fix: Quit two assert failures

We could now also remove the win->ime_data assert in wm_window_IME_end,
but think it's better to leave it for further development.

Fix for idproperty comparison

comparing arrays had flipped logic