The problem was caused by the fact that BKE_scene_copy() was simply doing "memcpy"
to duplicate render layers including lineset settings without taking care of pointers
in the linesets. For this reason, freeing the original scene and copied one resulted
in freeing allocated memory buffers twice.
Now BKE_scene_copy() properly duplicates linesets as part of render layers.
Also some code clean-up was made in the modified files.
Problem report by IRIE Shinsuke (with a patch to fix the crash). Many thanks!
* Disabling Skeleton Sketching now refreshes the view properly, so that strokes
don't linger on even after being disabled
* Added the delete operator to the panel
Although the bug report here wasn't exactly clear about what exactly was wrong,
it soon became apparent that the UI stuff here was in need of some love.
Changes:
* Ported over missing tooltips from 2.49 (i.e. most of them)
* Fixed a few incorrect tooltips (mostly the subdivision length settings)
* Made the autonaming and number/side settings slightly clearer - number/side
are used to replace placeholders in the names of template bones (&N and &S
respectively) when autonaming is disabled. When it is enabled, these values are
determined automatically.
Limit Constraints (i.e. Limit Location/Rotation/Scale/Distance) with 'For
Transform' option enabled would still be applied when transforming objects even
when they were turned off using the eye icon in the constraint panel headers.
The "off" flag was never added to the checks for muted or disabled constraints
to be skipped here.
draw_item *is* optional (it then uses default C function), even though there is not much sense to register a class without it, except for our default UI_UL_list!
Patch by Sergey, .blend by Thomas and some further tweaks by me.
Still to solve later: allow external engines to specify own preview .blend, for
now the code here is doing too much magic hacking on the preview scene still.
Issue was caused by recursive call of RE_BlenderFrame. Solved by
reshuffling image pool init/free in do_render_all_options.
Should be harmless, but doublecheck on this is welcome.
without NPOT support. Was wrong stride used for memcpy leading
to wrong memory writes in def_internal_icon.
It's a regression since matcap commit.
Should fix the following reports:
- #33993: Crash on Blender startup (Vista x32)
- #33996: Latest build crashes on win xp
Headerless panels are not supposed to be closed ever. But if user saves a blend with a stardard panel closed, then dev decides to make this panel headerless, when user open again its blend, the panel is closed and has no more header, so it becomes invisible!
This commit simply checks, at draw time, that a headerless panel is never closed (and repoen it if necessary)!
add an assert because if the mesh is in an invalid state the mask layer can exist but the mask pointer still be NULL (noticed this while looking into a different bug).
You can see a screenshot of the funcionality here:
http://wiki.blender.org/index.php/Doc:2.6/manual/Game_Engine/Performance/Display/Framerate_and_Profile
This patch creates a bar-like graph to quickly allow the game dev to see the performance changes.
Also it changes the font to monospace (too allow ' ' padding) and reduced shadow border to
match the blenderplayer one.
Patch finalized and commited at Global Game Jam Vancouver (last one, time to sleep forever now)
StrokeAttribute class methods were renamed from camel case to lower case plus underscore.
Also made changes to shaders.pyTVertexOrientationShader that uses the affected methods.
* The API syntax of StrokeVertex and StrokeAttribute was updated by means of getter/setter
properties instead of class methods. Python style modules (including the Parameter Editor
implementation) were updated accordingly.
* Code clean-up was done for a few Python style modules, mostly by removing duplicated
definitions of stroke shaders and fixing indentation.
(and marking rna as deprecated)
I talked with Benoit Bolsee and Mitchell Stokes and they both agreed that
the feature should be removed.
In case someone was actually using it the rna is still available. But next
release we remove both the rna, the DNA and the flag in the code.
I did a simple benchmark with tons of cubes, and the DBVT culling (use_occlusion_culling=True)
always perform better than when it's off. Even when no occluder objects are in the scene.
Also generate rigid body constraint types in py bullet code from RNA enum values (simpler than having to sync the code when something is changed here!).
Side note: RNA API about icons still needs to expose icons for enum values, and conversion funcs between icon_name and icon_value!
Oops! So there was an Unpack button, hidden in the Nkey properties of Image window.
This was drawn next to greyed-out buttons, didn't notice it well.
I also now found the unpack() menu in editors/util, and there's an Image unpack op.
Also the RNA api has an unpack!
Will remove the generic unpack op, and add a sound and vfont unpack instead.
Change Scene.frame_set so that it ensures subframe in range [0,1[ as Blender
expects, otherwise some things like physics point cache lookups don't get
evaluated properly.
- The "ID" buttons (for browse images, for example) now show a Pack icon, for packed
Images. Using this button allows unpack.
- Pack and unpack operations now give a Info report on what happened.
- Not restored yet: option to set "AutoPack".
The operator names all show up in the Search button. As such is nicer if they
can all have the main words capitalized.
e.g. "Snap strips" should be "Snap Strips"
"Copy to clipboard" should be "Copy to Clipboard"
This was done with a mix of bash tools, regex, and manual work because I'm too rushed into regex :)
+ fix bge stereo eye separation tooltip
The object color option in the materials always worked, however the initial color of the object
was never passed to the game.
We are now passing it only, only when the object's mesh has a material that has ((shade_flags & OB_COLOR)).
There reason to not always set the object color is probably due to performance (the m_bUseObjectColor
flag in KX_GameObject). This patch still respect that.
Bug report from Mike Pan, as part of our book nitty-gritties review work.
Bugfix during Vancouver Global Game Jam :)
