Age | Commit message (Collapse) | Author |
|
cleanup
|
|
|
|
investigation lead to finding 3 bugs here...
- transform key input handling didnt ignore minus key on an unsigned value as it should.
- not being able to set numcuts to 0 made typing in numbers not very useful.
- backspace would set the cuts to an unsigned value.
|
|
wire paint drawing into draw_mesh_paint().
|
|
surfaces
Changed one of the intersection tests to use < rather than <=
The sharp and smooth modes look fine now for the special case that was
reported broken; blocks mode looks correct but "jumps" slightly from
one octree resolution to another, so may need additional corrections.
|
|
Removed a lot of unused code, added comments and some clearer
naming. Minor code shuffles and style cleanup too.
|
|
Submitted by Daniel Balster
|
|
clarified with Daniel Salazar & Campbell Barton)
|
|
|
|
|
|
- avoid calculating face center when the face isnt used at all.
- avoid copying vectors when its not needed.
|
|
It was a regression since texture buffer merge from tomato, now ensure
all display happens in sRGB space as it used to be before.
|
|
like the user preferences.
Also renamed "View Docs" menu entry to "Python Documentation".
|
|
before the version patch which it ideally shouldn't, but moving it to a different
location is tricky.
|
|
|
|
|
|
from David M (erwin94)
|
|
to pass options, instead of having one parameter per boolean flag (i.e. replaces current useRenderParams and isFinalCalc by a single ModifierApplyFlag flag. ModifierApplyFlag is an enum defined in BKE_modifier.h). This way we won't anymore have to edit all modifier files when e.g. adding a new control flag!
Should have no effect over modifier behavior.
|
|
from Sebastian Nell (codemanx), with some edits
- Changed key shortcut to Shift+Enter.
- made into its own operator since it works differently to delete.
|
|
|
|
not work and making other stereo modes not work when the graphics card did not
support quadbuffer stereo.
|
|
change modifier to use the average of the RGB since perceptual conversion isn't really needed for modifiers.
|
|
smooth seems to be a keyword in glsl and using it as a name for function property caused issues.
|
|
functions
|
|
|
|
Made smarter check for composite output node which now checks if there are
composite output nodes in group nodes.
|
|
- Fixed suddenly disappearing preview image from track widget
- Also suddenly stopping tracking should be fixed now
Both issues were caused by mixing working with floats and bytes and
in some cases it was wrong or missed checks done for this.
|
|
|
|
|
|
the old file path was split correctly into base path + filename.
|
|
Patch by Gaia Clary
This patch removes the index numbering from materials, which is not necessary.
|
|
|
|
** I did not update rst docs. I will post on the quick-hack list. If no one volunteers to do it, I have it done before next release. **
this patch add the barebone functions to do glsl shaders in an addon.
For simplicity sakes it supports old GLSL, but not all the new modern methods (glBindFragDataLocation, ...).
It should be enough to make an addon to help with bge 2d filters.
I'm using this with a 3dview callback and it's working nicely:
self._handle = context.region.callback_add(draw_callback_px, (self, context), 'POST_VIEW')
I was considering not to commit this patch due to it not covering all OpenGL new methods.
But frankly it's not the first time bgl is expanded upon individual needs (guilty of that myself).
If a python developer knows enough to use OpenGL and need new functions, s/he should also be able to patch Blender for that,
so I see no problems here.
Another solution is to move to PyOpenGL or similar any soon.
glShaderSource implemented differently than opengl standards. It's taking one source at a time (instead of an array of sources).
This is similar with PyOpenGL solution.
glActiveTexture
glAttachShader
glCompileShader
glCreateProgram
glCreateShader
glDeleteProgram
glDeleteShader
glDetachShader
glGetAttachedShaders
glGetProgramInfoLog
glGetProgramiv
glGetShaderInfoLog
glGetShaderiv
glGetShaderSource
glGetUniformLocation
glIsProgram
glIsShader
glLinkProgram
glShaderSource
glUniform1f
glUniform2f
glUniform3f
glUniform4f
glUniform1fv
glUniform2fv
glUniform3fv
glUniform4fv
glUniform1i
glUniform2i
glUniform3i
glUniform4i
glUniform1iv
glUniform2iv
glUniform3iv
glUniform4iv
glUniformMatrix2fv
glUniformMatrix3fv
glUniformMatrix4fv
glUniformMatrix2x3fv
glUniformMatrix3x2fv
glUniformMatrix2x4fv
glUniformMatrix4x2fv
glUniformMatrix3x4fv
glUniformMatrix4x3fv
glUseProgram
glValidateProgram
|
|
|
|
space size and location were outdated often, and already computed on demand
by blender internal, now do that through RNA as well.
|
|
Cleanups in brute_region_tracker.cc.
|
|
|
|
|
|
undo.
The way this got updated from the context is a bit unreliable, and for handlers
the update couldn't happen because there is no context passed in. Now it's
updated from setup_app_data, which is where the change actually happens. I left
in the other updates to be sure but they should not be needed anymore.
|
|
|
|
Fix part of thread safety issue, there's still something else wrong.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This happens because of how output node index is initializing in assign_index function: itterator goes to the beginning of the nodes list using node->prev and then reviews the whole node list to find first unused index. The problem is that node's initialization now is getting called before node was added to node tree, so all output nodes have got equal index.
|
|
function has to check that before making the update call.
|
|
vector subtract and normalize for every vertex). now store the previous normalized vector for re-use.
also add BKE_mesh_poly_calc_angles() which is mostly a reference for now.
|