Blender Git Commits

Big Transform Manipulator Merge

*NOTE*: Some UI decision done in this commit will most likely be revised, all flame shall go in /dev/null. Constructive discussions of course welcomed.

This commit merges manipulator orientation selection back in "traditional" transform.
That's how it works:
- The dropdown in the 3D view header is always visible
- The orientation chosen will be used when choosing an axis with MMB and for the *second* key press of X,Y,Z
However, Local orientation doesn't use the one calculated by the manipulator. This is to ensure that multiple object local and armatures in pose still works as before.
- Alt-Space (to change the orientation) works during transform


New Transform orientation: View, using the view axis.

Fixes for the following bugs:
- Constraint projection code "jammed" if input vector was <0,0,0>, reported a couple of times on IRC. Thanks to Basse for the example file.
- Transform on texspace crashed on objects without texspace data (camera, lamp, ...). This was reported in tracker.
- Numinput with lock constraints didn't work correctly. Reported on elysiun

Probably some others that I'm forgetting

I also moved a couple of functions around in an attempt to make things clearer.

- Armature editmode now supports lasso. Please be aware that in editmode,
armatures actually only consists of points, flushing bone selection
based on points that are selected.

Renamed the "Lines" drawmode for bones to "Sticks". Thanks Basse! :)

- Fix; crash in selecting in editmode armature (commit 1 hour ago)
- Fix; in editmode, bones didn't transform (same commit)
- Fix; in drawmode "Sticks", names could go weird in editmode (commit
two days ago)

- added ModifierData flag, is modifier enabled in editmode
- added modifier type flag: should modifier be enabled by default for
active in editmode
- added subsurf "debug incremental" option instead of G.rt==52 (it becomes
a slightly useful feature now for debugging how well a modifier works
with incremental subsurf... maybe important for future python modifier
developers)
- shuffled modifier button layout just to keep people guessing
- switched back to drawing editmesh face centers not through derivedmesh,
I didn't think this one through, forgot that centers were also used for
selection. have to think about what to do about this, should be either
(a) don't draw centers with a cage active (optimal mode) or (b) come up
with api to draw centers through derivedmesh and also handle selection.
- changed recalc_editnormals to also follow the len(no)==0.0 use vertex
co convention

- change mesh_calc_normals to set vertices with len(no)==0
to normalised coordinate (convention in blender, helps with
halo)
- removed vertexnormals(), vertexnormals_mesh()
- removed CTX_NO_NOR_RECALC (always assume already calculated)
- change NMesh.c to call mesh_calc_normals
- chance load_editMesh to call mesh_calc_normals after done
converting instead of using editmesh normals
- update recalc_editnormals to also calc vertex normals (whats
4 more adds and a sqrt among friends)

Its hard to believe, but it just might be the case that there
are only two places mesh normals are calculated now (renderer
and kernel)

Killed silly modal PoseMode mode! :)

- PoseMode now is a state Armature Objects can be in. So, while in PoseMode
for an Armature, you can just select another Object or Armature.
- The old PoseMode options (transform, insert keys etc) are accessible
with making the Armature Object 'active' (and have object in PoseMode).
- At this moment no multiple Poses can be transformed/edited at the same
time.
- The old hotkey CTRL+TAB, and view3d header menu, still work to set an
Object's PoseMode

It was quite a lot recode, so tests & reports are welcome.

Oh, as a bonus I added Lasso Select for Bones in PoseMode! It selects using
only the line between root and tip of the Bone.

- cleaned up calc_vertexnormals, did unnecessary temporary array alloc
and passed over faces twice
- changed init_render_mesh to *always* call calc_vertexnormals. Vertex normal
calculation is insignificant time wise in comparison to modern renders so
it wasn't a big savings anyway, and vertex normals were sometimes incorrect.
Still issue remaining regarding what to do with wire normals.

- removed mface->puno flags, was only used to flip normals in display
and this is better left to user (whee this was a fun commit! so
much deleting!)
- removed mesh_calculate_vertex_normals (replaced by mesh_calc_normals)

added bandoler's SConscript file for Bullet

Added code in loopcut to undo cut if edgeslide was cancelled. (For you Chris!)


Also Added some code for UV's in subdivide. This is close, but still is not perfect :(

Added point.c to the SConscript file

_update to msvc7 project_
- added point to python
ps
- the point commit remove vector division as it is undefined

_new point class and update_
- adds a new point class
* point/ vector math (p + v = p, p - p = v, etc.)
* points can be transformed by matrices/quats
* wraps 'place vector' type vectors that have no magnitude
- wrapped toXXX() methods work correctly
* toXXX() will NOT wrap data (this is due to the fact that wrapped data cannot be converted)
* added a 'wrapped' attribute to mathutils classes to determine wether the object is accessing python or blender data
- added the ability to negate vectors/points with "-vec"
* deprecated vector.negate()
- added the ability to shorhand inverse matrices with "~mat" (tilde)
- conversion between vector/point with toXXX() methods

- added DerivedMesh.drawMapped{VertNormals,FaceNormals,FaceCenters}EM
functions and implementation for EditmeshDerivedMesh
- switch drawobject to drawing normals/centers through the DerivedMesh
- added G_DRAW_VNORMALS flag and button, implementation is not yet complete
because editmesh normals are not updated regularly
- switch editmesh draw buttons to use uiDefButBit (can't we get some
monkey to convert all of the uiDefBut calls with TOG|BIT type? It
makes grepping the source much nicer)

- fix a bug with softbody, verts were copied out of order (meant
history was lost... I think, dunno this code well.)
- commented out code to do merging of quad with 3 verts shared in
mirror mode... didnt seem worth the effort and mesh still wasnt
perfect afterwards
- bug fix, indices for triangles were not swapped correctly in
mirror, could lead to crash with subsurf in editmode

- agh!!!! horrible code alert!!! pushdata/popdata in build_particle_system
actually copy Object data and then store it back later. Of course this
means changes to object during the call are lost (and very hard to track
down). Eekadoodle!

Oh yeah, actual commit was to move call to fetch derived mesh above
pushdata, because this call can alloc cache's in the Object (which won't
get freed later if it is after pushdata)

- move build_particle_system to makeDispList, it was getting called
recursively in certain situations. needs review.

Events added for:

- clear parent (didn't do a recalc data for deformed objects)
- add/remove object name in modifier menu (redraw event)

Error in drawing of new armature options;
bglPolygonOffset(0) was accidentally called without it being set before,
causing window matrix to go weird. :)

Motion blur didn't work for Object Ipos (rest worked OK), so simple
fix. :)

Also removed the ugly old 'update for newframe' from renderwin.c, the
new call scene_update_for_newframe() does it much friendlier.

- moved mesh_getVertexCos to mesh.c and prototyped
- make mesh_modifier build vertex locations on demand