while. This may not fix all cases but should at least solve the issue when
rendering with cycles.
The cause was a race condition on C->data.recursion, with multiple threads
accessing context at the same time. Cycles itself does not access context
from the render thread, but the bpy api would do a context update for any
callback in case e.g. a new file got loaded. Disabled that now in non-main
threads.
The ideal solution would be to not allow any context access at all from threads
but that's not so simple to implement, especially not this close to release.
- memset(..., -1) is used incorrectly even though it worked: MOD_solidify.c - thanks Halley from IRC for noticing this. use fill_vn_i() instead.
- quiet warnings in editmesh_slide.c
- cleanup comments in bmesh and some other minor comment additions.
--debug
--debug-ffmpeg
--debug-python
--debug-events
--debug-wm
This makes debug output easier to read - event debug prints would flood output too much before.
For convenience:
--debug-all turns all debug flags on (works as --debug did before).
also removed some redundant whitespace in debug prints and prefix some prints with __func__ to give some context.
- filename comparison was invalid
- was stopping on the first traceback, which would reference the caller but not the error line (when the error was in a function).
this report exposed multiple bugs in blender when using a non utf8 compatible home directory.
- bpy.utils.script_paths() would crash when homedir wasn't utf8 (reported bug)
- PyC_DefaultNameSpace() - would raise an error when running when __file__ was non utf8.
- preset filepath property was not set to accept non utf8.
- bpy.paths.display_name would raise an error on non utf8 paths, (used for preset draw)
fix for collection functions not showing up in __dir__,
eg, console autocomplete didnt show up bpy.data.libraries.load
also fix refcounting leak with returning attributes from collections.
Not all file formats/calls are supported yet. It will be expended.
Please from now on use BLI_fopen, BLI_* for file manipulations.
For non-windows systems BLI_fopen just calls fopen.
For Windows, the utf-8 string is translated to utf-16 string in order to call UTF version of the function.
Wrap customdata, so far you can access the data layers in a pythonic way but not manipulate the customdata yet.
provides dictionary like access to customdata layers, eg:
texpoly = bm.faces.tex["UVMap"]
print(bm.verts.shape.keys()) # un-intended pun, keys() works on all layers.
print("MyInt" in bm.edges.int) # __contains__
layer = bm.faces.get("CheckForLayer")
we cant ensure buildinfo to have utf8 encoding so access it as bytes via python - a different workaround then this patch made.
also use C style string formatting for sys_info.py
added some missing functions too - which are not used yep but should be there for api completeness.
* CDDM_set_mloop
* CDDM_set_mpoly
* BLI_mempool_count
corrently allows to create and loop over verts/edges/faces, access selection and selection modes.
this is still WIP, access to face, edge verts is still missing, no access to UV's, no access to editing operations yet.
When the api is ready it will be documented by sphinx like mathutils, blf, aud.
made some small edits
- removed changes to AVI reading since the data types are apart of the format spec.
- absf -> abs for a double value in render code.
Python operator subclasses and operator types each get their own SRNA, causing double ups for bpy.types.__dir__()
From the operator type - these share names.
* ot->ext.srna
* ot->srna
Note that this conflict is still there, this only disables 'ot->ext.srna' from being included in dir(bpy.types).
This script was defining an operator within the panels draw function, while its possible to support this its really asking for trouble.
the fix is to raise an error when this happens.
also fix crash passing non classes to register_class/unregister_class
