Crash was happening on windows platforms only and was caused
by some specifics about how CRT works.
Basically, blender and all of the .dll are compiled with /MT
flag, which means blender.exe and all .dll are using separate
environments. This makes it impossible to pass file descriptors
from blender to other dll, because it becomes invalid in the dll.
And this is exactly what was happening: OIIO was trying to open
movie file with all known plugins and one of them was zlib. And
the way OIIO was using zlib API is opening the file using Boost
and passing a file descriptor to zlib. And since zlib was a
dynamic library this lead to general issues using this descriptor
in zlib code.
Solved by linking to zlib statically. This allows to safely pass
file descriptor to zlib API. Alternative would be to compile all
the stuff with /MD flag, but that's much bigger and less robust
way to fix the issue.
Tested on windows using msvc2008, scons plus cmake both 32 and 64
bit versions. Seems to be working fine.
Further tweaks for mingw and msvc2012 could be needed tho.
New OSX "Life resize" was also being called during opening of windows - when things are
not initialized yet. Crashed on opening full-screen window (which is animated in OSX).
* Panorama Camera was not working for Blender Internal and Cycles anymore.
Code checked for Freestyle render layer flag, but as the flag is enabled per default, it broke Panorama camera in all cases. It made no sense to check this on a render layer basis anyway... check if Freestyle is enabled in general now instead.
The rationale of this revision is to provide an easy way to make text datablocks internal
from within Python (i.e., by setting these properties to True and the 'filepath' property
to empty string).
Now add asserts to make sure this layer is only added once the modifier stack has been calculated.
this saves normal layer being calculated whenver vertex normals need updating.
many modifiers were calculating normals, when those normals were ignored by the next modifier.
now flag normals as dirty and recalculate for modifiers that set use `dependsOnNormals()` callback.
Quick test on mesh with 12 modifiers (mostly build type), calculated normals 6 times, now it only runs once - so this will give some speedup too.
Now button in the toolshelf behaves this way:
- User clicks on "Add Marker"
- Then he clicks where the marker should get placed
Patch by Marcos Couto (ocf) with own modifications.
Implements an automatic keyframe selection algorithm which uses
couple of approaches to find out best keyframes candidates:
- First, slightly modifier Pollefeys's criteria is used, which
limits correspondence ration from 80% to 100%. This allows to
reject keyframe candidate early without doing heavy math in
cases there're not much common features with first keyframe.
- Second step is based on Geometric Robust Information Criteria
(aka GRIC), which checks whether features motion between
candidate keyframes is better defined by homography or
fundamental matrices.
To be a good keyframe candidate, fundamental matrix need to
define motion better than homography (in this case F-GRIC will
be smaller than H-GRIC).
This two criteria are well described in this paper:
http://www.cs.ait.ac.th/~mdailey/papers/Tahir-KeyFrame.pdf
- Final step is based on estimating reconstruction error of
a full-scene solution using candidate keyframes. This part
is based on the following paper:
ftp://ftp.tnt.uni-hannover.de/pub/papers/2004/ECCV2004-TTHBAW.pdf
This step requires reconstruction using candidate keyframes
and obtaining covariance matrix of 3D points positions.
Reconstruction was done pretty much straightforward using
other simple pipeline routines, and for covariance estimation
pseudo-inverse of Hessian is used, which is in this case
(J^T * J)+, where + denotes pseudo-inverse.
Jacobian matrix is estimating using Ceres evaluate API.
This is also crucial to get rid of possible gauge ambiguity,
which is in our case made by zero-ing 7 (by gauge freedoms
number) eigen values in pseudo-inverse.
There're still room for improving and optimizing the code,
but we need some point to start with anyway :)
Thanks to Keir Mierle and Sameer Agarwal who assisted a lot
to make this feature working.
BF_PYTHON_INC can contain more than one path (in the mono-string format), don't know how this could not be found earlier, completely broke build of GE on any recent Debian/Ubuntu distro???
VS 2010 and 2012 have stdint.h so we should use them for consistency instead of defining the inttypes ourselves.
Added if condition for for stdint.h include for >= VS 2010
P.S.: Compiled and tested on VS2008-2012 and MinGW with scons and CMake ;)
Was missing keymap for REGION_TIMER, which ended up
in missing updates happening.
Added this kind of timer to RNA, so keymaps could
bind to it.
Also made 3ds max keymap working again. Too bad it
was broken in 2.67 and 2.67a :S
