Blender Git Loki

allow execution mode to be given as an argument to operators from python (requested by algorith)
example. bpy.ops.tfm.rotate('INVOKE_REGION_WIN', pivot=(0,1,2), ......)

bpy_array.c - was too strict with types, 0 should be allowed as well as 0.0 in a float array.

Bugfix for usage of uninitialized variable on windows (props_ptr.data needs to be set to NULL before calling uiItemFullO() ) - please check if this also compiles on gcc

2.5: fix missing LIB_NEEDLINK check in windowmanager reading,
would cause crash with linking/appending.

2.5/Multires:

* Added back multires delete higher levels (new operator + button)

2.5 - A few bugfixes...

* Autoside renaming tools in EditMode for armatures now works again. (Wrong property name)
* Action used by NLA Strips can now be chosen/changed to another action

2.5
*****
first commit in a long time, and its great to be back!

commited Select Mirror operator for objects
eg. L.sword->R.sword

added to 3dview select menu aswel

the hotkey is shift-ctrl-m (hope its not taken)

Missing header include for non-linux OS "BLI_exist()"

sound init/exit so at least the player opens

changes needed for building the blenderplayer with cmake on linux.

made editmesh<->bmesh conversion work again, hopefully (bleh). also made BLI_edgehash use mempools for allocating the bucket entries, for speed.

Grease Pencil: Datablock bugfixes

* Grease Pencil datablocks can now be properly browsed/added/unlinked from the UI panels

* Made Grease Pencil use the brush icon for now. A proper icon for this would be nice ;)

BlenderPlayer linking again for cmake - 148 errors gone.

After talking with Ton and Campbell we agreed that it wouldn't hurt to have blenderplayer again (specially now since BGE is almost 100% working in 2.5).
However in order to make it link, I needed to bring back stubs, a lot of so-called bad calls.

I'm not sure how we should proceed from here, but it looks like people could start to take a look at source/blenderplayer/bad_level_calls_stubs/stubs.c and fix their own modules/functions

** NOTE: I removed the sound calls from BlenderPlayer. In order to fix it look at //XXX ADD SOUND in GPG_Application.cpp and GPC_Engine

*** tested in CMake+MSVC.
- Scons is not building !!! (why does the building systems have to be so different?)
And someone may like to fix make. (take a look at /trunk/source/blender/blenkernel/bad_level_call_stubs/Makefile )

**** it may work better inside /source/gameengine/GamePlayer

remove Py_CmpToRich (copy of py3.0 function), instead only support == and != for PyRNA and KX_PySequence types.
mesh1 > mesh2 # will raise an error.

Fix thread hanging problem (mostly seen with material preview, but that sneaky f*er could strike any time).

Story time: Once upon a time, in the green valley of fileselect, BLI_end_threads would get called on an empty threadbase, depending on the result of a previous call to readdir(). The function would then gladly decrement thread_level to -1 which would cause all kinds of fun havoc. THE END.

Made sure thread_level is only incremented and decremented when needed. The caller should never have to make sure of that, especially since it already lets you call with a null threadbase.

Please report any further hang (and how to reproduce, if possible).

* actually commit the sndfile dll copying (and not just claiming I did)

* KX_PythonSeq - comparisons work again. eg. act1.sensors == act2.sensors, had to copy Py_CmpToRich inline grr!, mailed python-dev about this.
* Shift-Click on states in the logic UI works again.
* New Logic Space has all the view options pressed.

compile fix

compile fix

commit of patch 19268, rotate uvs by wael oraiby

Blender 2.5
* recent files now just write content of G.recent_files, was adding untitled.blend!
* removed unused and now superfluous code reading the .Bfs file (is done in fsmenu now)