- python 3.2 does 'import site' on startup which now tries to parse pyconfig.h which isn't copied. so for now just run without importing 'site', alternative would be to copy the header file for posix systems.
- cache PYTHON_VERSION variable so it can be set to 3.2, needed for copying python installation's other then 3.1.
mode with drivers
ChildOf constraints added using the PoseBone.constraints.new() method
via Python scripts instead of using the operator (this latter method
is still the preferred/recommended method) were not getting some
critical flags set, causing errors arising from space conversions
being performed more than once.
- negative ray casts would invert the offset direction.
this meant if positive and negative were enabled at once and the mesh was slightly inside & outside the object it wrapped,
the offset would be applied in opposite directions.
This way the offset is always along the vertex normal.
- allow negative offset from RNA, could be useful and no benefit to disable.
While testing Apply Visual transforms last night, I noticed that
setting a constraint to use its owner as its target was allowed and
didn't trigger any warnings. This clearly doesn't do any good and is
different from the old behaviour.
After talking with Ton Roosendaal we agreed on making those changes.
I wonder if we should go for "Start Game" instead of "From File". But Space is not a constraint here so be it.
- Armature Actuator now only shows the "Secondary Target" option when the Bone Constraint supports it (IK only now).
-- that may be overkill I don't know. It shouldn't slow down the UI considerably, so it should be fine. Easy to revert if needed though.
- renaming things such as "Start frame" to "Start Frame"
When using masks or other simple 3D elements in composites, doing
a layer re-rendering on a node is a bit clumsy all the time.
This commit does two things to help:
- new hotkey "Z" in node editor automatically finds render layer
that changed and re-renders it + composites
- option "Auto Render" does same, but then after every transform
edit in 3D window
The latter is experimental; real & proper system for this requires
full threaded render support (like previews). But it works!
Demo file:
http://download.blender.org/demo/test/auto_composite.blend
Important fix:
After any render, all the render layers were tagged "changed", which
caused any edit to first totally recomposte everthing. Now it only
composites changes.
Implementation notes
- DAG scene flush now sets 'changed' flags in render layer nodes
- Added notifier for 'transform finished' to trigger the update,
this is temporarily.
Drivers for Scene, World, and Compositing Nodes now "work" (well,
sort-of)! Previously they were strictly restricted to object-
accessible data only; now they can function across the board (give or
take some weak spots).
Although there is still no depsgraph support so that these properties
update properly when their source controls are changed (this will
probably require a lot more work), they can still update under other
circumstances (i.e. frame change and/or manual refresh flushing via
mouse movement, etc.)
As the depsgraph tagging support is lacking, these just get always
executed for now, which might potentially be quite sluggish, though it
is hoped that there are so few of these top-level datablocks with
drivers hooked up that this is barely an issue in practice. At least I
haven't noticed any substantial slowdowns for animation playback, so
it should probably be fine.
* There were a lot of settings in the particle panels that made no sense for simple hair and only cluttered up the ui.
* Now these settings are hidden by default unless "advanced" hair options are shown.
* Without advanced options the particle velocity controls are replaced by a simple "hair length" value, which actually corresponds to the grown hair length in blender units.
* Some hair effector options that are actually very useful were not shown in ui. These are now found in the "field weights" panel.
Moving this code to the same method as is used for Apply Visual
Transform (for bones case), which uses standardised decomposition
methods instead of the improvised ones used here before. This was in
an attempt to get IK chains Visual Keyed correctly, though this
doesn't seem to have improved the situation much.
Also, the decompositon for objects here now takes care of euler
rotation modes.
TODO: other rotation modes should get included here too...
Simplified the code for Apply Visual Transform as Pose.
This makes it work for IK chains now, since there aren't any
recalculations of the pose involved now, which were contributing to
the pose getting incorrect as the applying process went on, since the
IK chain was getting calculated differently with each successive
change to the chain.
Probably wouldn't cause a problem but manually editing these types through python could easily crash blender.
also changed cmake, stub-makefile default build dir to be lower case and leave out architecture string, easier for documentation.
Use ../build/linux/ rather then ../build/Linux_i686/
When using Pose Lib with a search string, if the search fails after
typing a character and then you try to backspace to get back to the
previous search string, the pose does not refresh until you type in
another valid letter such that the search works again.
Did a logic reshuffled in the get next preview pose function so that
this is dealt with properly
- [#25951] Hardcoded Keys in PoseLib modal operator register twice per
use
Now only the KM_PRESS events are handled. KM_RELEASE events for these
are ignored...
- While checking the above, I also found that the view-manipulation
while previewing functionality was broken. This was one of the
benefits of this UI approach over a fixed screenshot-only listview.
Fixed now :)
Fix in crashing displacement last friday, assumed wrongly that
the old 'new bump' failed for displacement too. It still did
work (somewhat). This restores the situation again.
- Bugfix #25937
Child-of constraint now behaves like regular parent-child
relationship when all options are set. This prevents the
errors that can happen when decomposing non-uniform matrices.
- Todo item
The area corner hotspots for splitting/merging were far too
narrow. Now it uses a circular distance to detect whether
the hotspot is active. Also cleaned up drawing code for it.
* After countless different bugs particles should now render correctly inside dupligroups.
* Only particles with metaball visualization are still problematic, this is mostly due to the ancient metaball code.
* I'll also add a test file for some of the situations, so that hopefully these cases stay fixed :)
- 'axis' arg was not coerced from a tuple like other args now do.
- 'axis' arg was being modified in-place (VERY BAD).
- also made new function matrix_3x3_as_4x4().
3D View:
In quad-view mode, the options to set back/bottom/left views were
blocked. Now they work again.
Note: the oparator polls don't have a check for properties... so it
cannot be finegrained here. Checks are inside code.
These should not have any effect on render results, except in some cases with
you have overlapping faces, where the noise seems to be slightly reduced.
There are some performance improvements, for simple scenes I wouldn't expect
more than 5-10% to be cut off the render time, for sintel scenes we got about
50% on average, that's with millions of polygons on intel quad cores. This
because memory access / cache misses were the main bottleneck for those scenes,
and the optimizations improve that.
Interal changes:
* Remove RE_raytrace.h, raytracer is now only used by render engine again.
* Split non-public parts rayobject.h into rayobject_internal.h, hopefully
makes it clearer how the API is used.
* Added rayintersection.h to contain some of the stuff from RE_raytrace.h
* Change Isect.vec/labda to Isect.dir/dist, previously vec was sometimes
normalized and sometimes not, confusing... now dir is always normalized
and dist contains the distance.
* Change VECCOPY and similar to BLI_math functions.
* Force inlining of auxiliary functions for ray-triangle/quad intersection,
helps a few percentages.
* Reorganize svbvh code so all the traversal functions are in one file
* Don't do test for root so that push_childs can be inlined
* Make shadow a template parameter so it doesn't need to be runtime checked
* Optimization in raytree building, was computing bounding boxes more often
than necessary.
* Leave out logf() factor in SAH, makes tree build quicker with no
noticeable influence on raytracing on performance?
* Set max childs to 4, simplifies traversal code a bit, but also seems
to help slightly in general.
* Store child pointers and child bb just as fixed arrays of size 4 in nodes,
nearly all nodes have this many children, so overall it actually reduces
memory usage a bit and avoids a pointer indirection.
Patch #25901 by Tobias Oelgarte.
Bone transformations would be converted back and forth between different
representations when changing modes, which due to numerical errors could
lead to bone transformations slowly changing as you edit the armature.
Now the editmode head, tail and roll values are stored in bones and used
directly when entering edit mode. Head and tail were already there but
now we ensure they are the exact same value, roll was not yet there, so
we have a version patch for it.
The sub version was incremented to 1 for the version patch.
Using displacement with new bump caused NaN values or crashes.
Since displace code doesn't provide derivatives, it now falls
back to "old bump" by default.
