<blush>Missing "2*" caused AO tables to be only initialized half</blush>
Result was bad AO quality in render, and unpredictable brightness. This bug
happened in previous commit, when fixing random table issues.
never work correctly with it, the way it is currently implemented.
The selection won't draw correctly either with TextOnCurve, and this is
really asking a bit much..
What to do...
rendered with MBlur or Fields.
This is a fix for now, but I've already noticed several pending issues for
Blender's internal time control (time ipos, global time control, startframe
offsets, etc). That's for another time! (pun not intended :)
- Toolbox item for shaded mode had wrong hotkey
- Added countall() for select menu stuff (in 3d header, groupmenu)
That way the stats in infoheader are updated.
- Tooltip was wrong for X-axis mirror mode Armatures
- Move UVTEXTTOOL variables into global Gip struct (like Gvp for
vertex paint). This will probably be moved into SpaceImage later,
so it is saved with the .blend file.
- Disable tool drawing. a better solution needs to be found.
- Panel button layout is still the same, this will change.
- Removed the NAN_TPT define, it has no use anymore.
Note: The "Insert Text" button pretty much replaces this, together with
the fact that editing 3d text is now much more powerful than the
usual text editor ;-)
IK: needed to change abs() to TNT::abs() (thanks brecht!)
Freetype: FT_ENCODING_UNICODE typecast to FT_CharMap (thanks intrr!)
header_image: missing include for liballoc stuff
Ketsji: needed to declare m_hitObject as public in KX_MouseFocusSensor.h.
This is probably the wrong thing to do, but KX_MouseFocusSensor.cpp
compilation fails on line 279 otherwise.
Further information is available here:
http://wiki.blender.org/bin/view.pl/Blenderdev/UnicodeFont3D
Shortlist of features:
- Unicode character support for Font3D
- UI to select characters from Unicode character list
- UI to select Unicode table areas
- Optimized character loading (Load only those characters which are used
in font object)
Please test extensively if it breaks anything, try also loading/saving
files, packing fonts, etc.
The official text regression file in the regression suite should be a
good start.
Thanks to mikasaari for this very useful addition!
Still not good, i.e. getting these when quitting: Error Totblock: 4
new bpytriple len: 60 0x8889bdc ... 'cause nothing frees them..
Changed the loop that parsed input args to PyArg_ParseTuple to have
support for passing ints from Python too as the floats that are the
coordinates. Didn't find PyInt_AsFloat and figured that this is an ok
way anyhow.
Changed the default handle mode from AUTO to ALIGN, which is the same
as in UI and more useful at least for me.
Little sanifying in CurNurb (this was done with Ton).
- "make track" in posemode only worked for objects, so disabled it for now
- in weightpaint, "Clear rotation/location" worked on armature-pose, as well
as on object. Was confusing... disabled clearing object in weightpaint.
conversion of data still. Remains a painful issue to get things converted
and in same time prevent things from calculated twice. :)
Anyhoo, issue was that old files with armatures in hidden layers went wrong
Contributed by Toni Alatalo (antont).
Support for Taper Objects for Curves.
Code for curnurb.setFlagU() method not changed as per discussion on
bf-python mail list.
much... made it 10 times larger, to match the max zoom level.
Nevertheless; on our ancient todo is that we once should give number
buttons 2 limits. One for the actual min/max values, and one for the best
sliding experience (matching zoom level or distance for example).
Using "make parent armature with creating vertex groups" option gave
corrupted memory when one or more bones had the "Deform" option not set.
Was caused by old boneclass SKINNABLE variable. To end this confusement,
removed this define from code.
Note for future Armature tool coders; don't use this bonelooper
functionality. Operations in Object mode for armatures should use the
Pose Channels, which is a simple ListBase to go over. I have to recode
the 'create vertexgroups' still...
- Pontus Lidman contributed a new module: Blender.Key + access to key objects from NMesh, Lattice and Curve + docs (thanks and sorry for taking so long to check/commit the patch!)
- Allowing EVENT spacehandlers to call the file selector (scriptlinks in general are not allowed, but this special case should be able to). Requested by Paolo Colombo (thanks!)
- tiny doc update (Ken Hughes pointed an error in the space handlers example)
I didn't have time to update the Key module to follow the current bpython design, will do that later and also test it better than I did.
