Blender Git Commits

2.5

- Lasso select back (CTRL+LMB), object centers only,
and for editmesh. See template in wm_operators.c

- Circle select for editmode back. Currently it still
uses Ckey, and is only a temporary mode, not persistant.
Persistant circle select can be added later with
tweak gesture keymap? We'll see. :) The old circle
select was actually annoying that it was so sticky.

RNA
* Added skeleton code for particle/object_force/userdef.
* More Object properties wrapped.
* Added User Preferences display in outliner.

2.5 - Action Editor: Toggle settings

* Added back enable (Ctrl-Shift-W), disable (Alt-W), and toggle (Shift-W) operators to channels view. They are separate operators which use the same backend code.

* Fixed icon-drawing for IPO-curve 'protect' buttons. After doing this, I've realised that many tools will need some alterations to take this into account. That commit will come later.

cleaned up bmesh flags. current design is the first 16 bits of BMHeader->flag are reserved for the originial element flags; the next 5 are used by various flags that replace single-variable flags, with the temporary flags coming after this. further work may be warranted. also, renamed all BMESH_ macros to BM_ ones, since this inconsistency was kindof confusing.

2.5

Bugfix in OSX ghost!
The GHOST_TWindowState was set to 'invisble' always... which is a
state now handled correctly in Blender. It caused the window to not
refresh when resizing.

fixed a few things, actually the customdata functions were there, there was just some find/replace errors

made walker api compile, though I'm sure it doesn't work and needs a good read-through. also made conversion code compile, though I think it may be missing some customdata stuff. note, this may not actually link.

Added RNA for the ParticleInstance, Explode, and Bevel modifiers.

2.5
*******
small commit, after some help from ton i have fixed the clear transformation modifiers, they now redraw correctly, they was missing a depsgraph tag, also did a little tidying up

* renamed some of the RNA sound identifiers
* was missing a MEM_freeN in previous commit

Added RNA for the MeshDeform, Cast, and Smooth Modifiers. Also some changes to pointer types in other places.

python support for reporting with operators.
* errors in python called operators are raised as errors
* Python defined operators errors are reported as errors (not full traceback yet)
* added BKE_reports_string, same as BKE_reports_print but it returns a string rather then printing it.
* WM_operator_name_call optionally takes an initialized report struct

Added RNA for the UVProject modifier.

Had a couple problems with this one; the uv layer and the projectors. The projectors in particular is confusing; I figured this would be one of the cases handled by RNA automagically, but maybe I need to define custom functions for it? Anyway, marked the code with XXX.

2.5 - Silencing more MSVC warnings

Added RNA for the Displace modifier. Once again, not sure how to handle uvlayer.

Added RNA for the Array modifier.

-> Fixed BM_remove_tagged_XXX functions

BM_remove_tagged/verts/edges/faces are now API functions
and use operator flag layers to do their work. This meant
that I also removed the BM_Delete_XXX functions since they
were no longer relevant.

-> Changes to BMO_Set/Clear/TestFlag

the BMO flag functions should be taking
a void pointer as argument and then do
an explicit cast within the body of the functions.

-> Moved some stuff out of dupeops into API

-BM_Select selects an element regardless of it's type
-BM_Is_Selected tests whether or not an element is selected
regardless of it's type.
-BM_Copy_Attributes copies custom data/flags/material indices/ect.

The individual select functions should be made static and in future
any additions to structures will require changes to BM_Copy_Attributes.
Operator writers should always use this function instead of copying such
data directly.

Todo:
BM_Copy_Attributes could optionally accept a bitflag defining what
types of data should be copied. ie:

BMESH_CUSTOMDATA|BMESH_HIDE|BMESH_SHARP, ect.

Note to joe:

I think this is the best way to deal with it in the API from now on.
It strikes good balance between keeping implementation details wrapped
up in the API while not having to worry about building some complex
attribute system.

So I suppose this makes the additions you made to bmesh_marking.h redundant?

Made CMake build a little faster by not rebuilding the DNA and RNA stuff when building creator.