15 years agoDid put softbody time scaling stuff to functions,
Jens Ole Wund [Tue, 12 Apr 2005 21:36:21 +0000 (21:36 +0000)]
Did put softbody time scaling stuff to functions,
so we can decide later what proper timing is meant to be.
i've prepared something in static float sb_time_scale(Object *ob) [softbody.c] (ton :) )

'hacked in' particle collision system to match softbodies needs
naa .. i don't realy like the hack to int pdDoDeflection(..) in kernel .. effect.c  (but it works :) )
so .. until we'll have a 'nice' collision detector this is what 'softbodies simulator can live with'
did not remove intentionally test function static int sb_deflect_test(..) for further discussions  sbcol6.blend should work for a while

15 years agoBug fix #2438
Ton Roosendaal [Tue, 12 Apr 2005 11:53:46 +0000 (11:53 +0000)]
Bug fix #2438

Snapping code checked for Hooks in object, but forgot to check first if
the Hook->parent existed. Crashed on 'empty' hooks.

15 years agoBug fix #2323
Ton Roosendaal [Tue, 12 Apr 2005 11:45:06 +0000 (11:45 +0000)]
Bug fix #2323

The addition in 2.36 to correct bumpmapping for View and Object rotation,
made many nice bumpmap setups in older files less nice. Especially apparent
when you apply "flipped" mappings in "Map Input" Panel, like map the X to Z
and the Z to X. A correction then gives wrong results.

Since the "correct bumpmapping" is actually a new option, causing extra
rendertime too, I've made it an option now. With the "Map To" Panel being
fully used, I've put the new button option (temporally I hope) in the first
"Texture" Panel in material context.

Will be updated in release log page!

15 years agoBug fix #2394
Ton Roosendaal [Tue, 12 Apr 2005 10:08:04 +0000 (10:08 +0000)]
Bug fix #2394

While in camera view, going to local axis rotation (R, X, X) the view
became instable. Was still using the view3d view matrix, instead of
initialized TransInfo->viewmat.

Martin; you could check on the mouse motion in this case, for some reason
it only uses horizontal motion?

ALso; renamed button in InfoWindow "Disable sound" to "Disable Game Sound".

15 years agoQuick fix for bug #2393 (same as #2218), deformed meshes were not transformed
Alfredo de Greef [Mon, 11 Apr 2005 23:23:25 +0000 (23:23 +0000)]
Quick fix for bug #2393 (same as #2218), deformed meshes were not transformed
correctly anymore because of changes in other blender code.

15 years agoTransform constraints with Ykey were mixed up, setting text to global
Martin Poirier [Mon, 11 Apr 2005 22:29:20 +0000 (22:29 +0000)]
Transform constraints with Ykey were mixed up, setting text to global
instead of local and vice versa.

Thanks to GSR's email for helping me notice that.

15 years agoSmall transform fixes.
Martin Poirier [Mon, 11 Apr 2005 21:19:38 +0000 (21:19 +0000)]
Small transform fixes.

Rotation acted weird in perspective mode. Works much better using viewinv
instead of persinv to get the rotation axis. What would be even better is
using real view vector depending on transformation center (with
getViewVector), but that can wait a bit.

Meta Element resize needed some code to reset radius changes or axis size
changes when switching from one to the other (my fault, I didn't think of this
when telling Jiri how to implement it).

15 years ago - It is possible to scale radius of MetaElem again.
Jiri Hnidek [Mon, 11 Apr 2005 19:31:05 +0000 (19:31 +0000)]
 - It is possible to scale radius of MetaElem again.
 - It is possible to change dx, dy and dz params in 3Dview with manipulators
    or with some shortcuts (S-X, S-Y, S-Z).
 - User can scale stiffness when green circle is selected.
 - MetaElem is selected with RMB click at green or red circle, then user can
    grab or rotate with MetaElem as usual.

 - Screenshot:

 - Captured video:

15 years agoBug fix #2411
Ton Roosendaal [Mon, 11 Apr 2005 17:55:20 +0000 (17:55 +0000)]
Bug fix #2411

(Looks like big commit, but is mostly just change of API call!)

Particle emittors now can be parented to an armature Bone, and give the
correct path for each particle. Note that this doesn't work for deform!
And, for each particle the entire armature is evaluated, all actions and
NLA strips.

It used to work a little while BTW, but the code just called ALL armatures
and made ALL deforms again. Was quite slow... thats why the API call
change: do_all_actions() now accepts Object * to only do that one. With
a NULL it now does all still. Will disapppear in recode of course...

15 years agoSilly pointer set on wrong place in code, caused Bones transform in
Ton Roosendaal [Mon, 11 Apr 2005 15:14:07 +0000 (15:14 +0000)]
Silly pointer set on wrong place in code, caused Bones transform in
PoseMode not to work on branched armatures with multiple selected.

Thanks Richie for the test file!

15 years agoBug fix #2413
Ton Roosendaal [Mon, 11 Apr 2005 14:01:34 +0000 (14:01 +0000)]
Bug fix #2413

Usage of uninitialezed pointer in 'extrude edges only' could crash.

15 years agoBug fix #2415
Ton Roosendaal [Mon, 11 Apr 2005 13:49:05 +0000 (13:49 +0000)]
Bug fix #2415

Envmap rendering was using Gauss filtering, causing dark outlines in edges
of cube pictures. Now it temporally disables gauss in envmap render, which
isn't really noticable since the pics are filtered + gaussed anyway in the
end result.

15 years agoBug fix #2428
Ton Roosendaal [Mon, 11 Apr 2005 10:45:20 +0000 (10:45 +0000)]
Bug fix #2428

The "Separate" and "Split" commands in editmode Mesh didn't take loose
edges into account correctly (leaving them in). These edges now get
removed correctly.

Note; this changes the tool XKEY "Delete->Faces" slightly. This already
removed entire faces and loose selected vertices. Now also removes loose
selected edges. More consistant, but maybe a bit confusing...

15 years agoBug fix #2431
Ton Roosendaal [Mon, 11 Apr 2005 10:20:17 +0000 (10:20 +0000)]
Bug fix #2431

Underflow case in Blinn spec... resulting in large negative values to be
returned in certain cases, causing (again!) black dots.
Thanks randall for clear report with nice demo file. :)

15 years agoBug fix #2424
Ton Roosendaal [Mon, 11 Apr 2005 08:49:42 +0000 (08:49 +0000)]
Bug fix #2424

With a sun lamp shining exact perpendicular to a face, the result was
undefined (black or full lit). Was caused by using acos() without checking
the input range. Float calculus can result in small under or overflows...
therefore: always use the arithb.c function saacos() !!!

Found more acos() usage in rendercore (in blinn and arealight too), removed
all of them. Might solve the white or black dots people were complaining

15 years agoResize didn't work with bezier handles correctly. That is because bezier has prelemin...
Martin Poirier [Sun, 10 Apr 2005 20:09:31 +0000 (20:09 +0000)]
Resize didn't work with bezier handles correctly. That is because bezier has preleminary work for Individual Center in edit mode.
While I was there, I cleaned up individual center a bit, it was messing with helpline.

For fun, set the center mode to Individual Center, create a bezier circle, change handle to free (H) and Resize or Rotate. Fun isn't it! :)

15 years agoMerged all the internal transform includes into transform.h and moves it to /src...
Martin Poirier [Sun, 10 Apr 2005 18:33:19 +0000 (18:33 +0000)]
Merged all the internal transform includes into transform.h and moves it to /src to /include.
Split the conversion fonctions and sorting functions from transform.c into transform_conversions.c
Update MSVC 6.0 projectiles and SConscript accordingly.
Editview still included transform.h, replaced that for BIF_transform.h, the external include.

15 years ago - don't apply creasing to edges with no adjacent faces... things
Daniel Dunbar [Sun, 10 Apr 2005 15:04:52 +0000 (15:04 +0000)]
 - don't apply creasing to edges with no adjacent faces... things
   go wacky.

15 years ago - bug fix #2433, mesh not drawn correctly when only edges
Daniel Dunbar [Sun, 10 Apr 2005 14:31:04 +0000 (14:31 +0000)]
 - bug fix #2433, mesh not drawn correctly when only edges
   (and Mesh.medge).

15 years agoBugfix for disappearing hilight bug and code re-org for text editor
Stephen Swaney [Sun, 10 Apr 2005 14:01:41 +0000 (14:01 +0000)]
Bugfix for disappearing hilight bug  and code re-org for text editor
from Ricki Myers (themyers).

Comes with nice juicy commit msg, too!

- source/blender/blenkernel/BKE_text.h

- Removed indent_paste, uncommen, unindent_lines, comment_paste,
uncomment_paste, uncomment, set_tabs.

All these functions cut and re-added text (I felt this was
unsafe). whicch is was caused the highlight loss.

- Now the only functions are Indent, Unindent, comment, uncomment,
  setcurr_tab.  All these functions only take one @parm (struct Text)

-indent(struct Text *text)
copy's the selected text in a MEM_mallocN line by line added a
tab at the begginning

- Unindent(struct Text *text)
Tests if current line starts with a tab.
if TAB remove it
- comment(struct Text *text)
copy's the selected text in a MEM_mallocN and adding a # at the begginning

- Uncomment(struct Text *text)
Tests if current line starts with a #.
if # remove it

- setcurr_tab (Text *text)
Checks for Tabs pri. to any text
if : is found and not in a comment then Tabs is increased by one
if "return", "break", "pass" is found then Tabs is decreased

- blender/source/blender/src/header_text.c

Changed: txt_cut_sel(text);
* no more cutting of the text

- source/blender/src/drawtext.c

set_tabs(Text *text) just calls setcurr_tab(text);

15 years agoUncommitted my last commit.
Rob Haarsma [Sun, 10 Apr 2005 08:46:47 +0000 (08:46 +0000)]
Uncommitted my last commit.

15 years ago - made shadedisplist not free all object displist data...
Daniel Dunbar [Sun, 10 Apr 2005 00:00:34 +0000 (00:00 +0000)]
 - made shadedisplist not free all object displist data...
   this messed with data caches which messed with assumptions made
   by drawing code (and was just silly not to mention). here
   be dragons...

15 years agoFix for bug #1576.
Chris Want [Sat, 9 Apr 2005 23:38:35 +0000 (23:38 +0000)]
Fix for bug #1576.

Constraints on bones that point to objects/bones outside of
an armature weren't working. The figure_pose_updating()
does a bunch of flagging based on the constraints on the
scene to figure out what armatures/bones need updating
(even when it is not the armature that is being manipulated).

15 years agoClean up for Makefiles: some of the directories were being printed
Chris Want [Sat, 9 Apr 2005 22:40:34 +0000 (22:40 +0000)]
Clean up for Makefiles: some of the directories were being printed
wrong (e.g., "/Fuzzics" and "/Joystick", and some of  the solid stuff).
This was particularly noticable during "make clean".

15 years agoAdded console printout when rendering in parts, to keep track of progress.
Rob Haarsma [Sat, 9 Apr 2005 09:18:32 +0000 (09:18 +0000)]
Added console printout when rendering in parts, to keep track of progress.
Requested by Relic.

15 years ago - bug fix #2400 (I forgot to add some parameters in uiDefButF() function)
Jiri Hnidek [Fri, 8 Apr 2005 10:30:46 +0000 (10:30 +0000)]
 - bug fix #2400 (I forgot to add some parameters in uiDefButF() function)

15 years agoToggle between Connected and Normal PET.
Martin Poirier [Thu, 7 Apr 2005 18:28:41 +0000 (18:28 +0000)]
Toggle between Connected and Normal PET.

Alt-C during transform (hotkey subject to change)

No visual clue right now other than what is transformed and it gets resetted everytime (not kept as a global flag).

So, still a lot to do regarding that.

15 years ago - fix for crash in separate_mesh... thats some pretty nasty code
Daniel Dunbar [Wed, 6 Apr 2005 17:21:49 +0000 (17:21 +0000)]
 - fix for crash in separate_mesh... thats some pretty nasty code

15 years agoFixed uninitialized causing almost infinite loop.
Jens Ole Wund [Wed, 6 Apr 2005 10:29:04 +0000 (10:29 +0000)]
Fixed uninitialized causing almost infinite loop.
Wind is media friction now, to have consitent relaxing when the squall is gone, fixed direrction.

15 years agofixed small glitch multiplying nodemass twice (thanks ton)
Jens Ole Wund [Tue, 5 Apr 2005 20:28:32 +0000 (20:28 +0000)]
fixed small glitch multiplying nodemass twice (thanks ton)
claened up sbObjectStep(...) to follow 'time step rules'
added really care for framerate in scene
renamed arguments in softbody_calc_forces(...); and softbody_apply_forces(...); for better reading
fixed particle integration to be ODE solver compatible

15 years agoAdded OPENGL includes to this one as well.
Kent Mein [Tue, 5 Apr 2005 15:06:41 +0000 (15:06 +0000)]
Added OPENGL includes to this one as well.


15 years agoAdds OpenGL includes for details see:
Kent Mein [Tue, 5 Apr 2005 15:02:21 +0000 (15:02 +0000)]
Adds OpenGL includes for details see:


15 years agoSupport for SGI Prism:
Chris Want [Tue, 5 Apr 2005 14:33:43 +0000 (14:33 +0000)]
Support for SGI Prism:

* set python version to 2.2;
* set link flags for ia64 linux to the same as i386, alpha.

Thanks to Wayne Vieira for this!

15 years agoModified the warnings used (was using c++ warnings for c files)
Kent Mein [Tue, 5 Apr 2005 14:32:35 +0000 (14:32 +0000)]
Modified the warnings used (was using c++ warnings for c files)


15 years agoCrease transformation
Martin Poirier [Mon, 4 Apr 2005 23:37:15 +0000 (23:37 +0000)]
Crease transformation

Behavior is a bit different. Instead of showing the weird two value thing like the old transform did, it always show the weight variation in +/- that is applied to every edge. Works a bit like scale. Move away from center to increase, move closer to decrease (between -1.0 and 1.0).
Works with PET (not using connection yet).

Also removed some missed global Trans variable abuse.

15 years agoMSVC 6.0 projectfiles
Martin Poirier [Mon, 4 Apr 2005 23:30:14 +0000 (23:30 +0000)]
MSVC 6.0 projectfiles

Removing subsurf.c from the BKE_blenkernel project.

15 years ago- Made SoftBody work with Particle Force Fields.
Ton Roosendaal [Mon, 4 Apr 2005 18:09:47 +0000 (18:09 +0000)]
- Made SoftBody work with Particle Force Fields.

- Added new (Particle) Deflector; type Wind.
  Wind gives constant directional force. It is animatable (Ipos) and reacts
  to Object scaling. Also uses FallOff. Works for particles and SoftBody
  quick movie check;
  test file is in

- Added MaxDist option for forcefields, to control its influence better.
  Is drawn as circle in 3d window.

Forcefields are a bit weak still... should react to scaling, or not; in
that case drawing should indicate it (done for spherical field now).

15 years agoMoved the start of an ifdef so it includes vars specific to the ifdef
Kent Mein [Mon, 4 Apr 2005 15:57:50 +0000 (15:57 +0000)]
Moved the start of an ifdef so it includes vars specific to the ifdef
to get rid of warnings about unused vars (on non __APPLE__ system)


15 years agoYuck, forgot to uncommit the ugly version of Mat3ToSize(). Caused scaling
Ton Roosendaal [Mon, 4 Apr 2005 14:21:09 +0000 (14:21 +0000)]
Yuck, forgot to uncommit the ugly version of Mat3ToSize(). Caused scaling
to do very weird things on rotated objects.

15 years ago - change make_orco_displist_mesh to use current editmesh as source
Daniel Dunbar [Mon, 4 Apr 2005 12:22:33 +0000 (12:22 +0000)]
 - change make_orco_displist_mesh to use current editmesh as source
   if needed.
 - don't return cached subsurf during editmode, order is messed up (fixes
   orco rendering).

15 years ago - dropped user specified vertex data functions for ccgsubsurf,
Daniel Dunbar [Mon, 4 Apr 2005 05:21:20 +0000 (05:21 +0000)]
 - dropped user specified vertex data functions for ccgsubsurf,
   didn't really add value. hardcoded to use float[3] now.

15 years ago - removed DispListMesh.{editedge,editface}
Daniel Dunbar [Mon, 4 Apr 2005 04:50:41 +0000 (04:50 +0000)]
 - removed DispListMesh.{editedge,editface}
 - removed displistmesh_from_{mesh,editmesh}
 - removed EditVert.ssco
 - removed unused functions for DispListMesh DerivedMesh

 Still need lots more testing for this stuff.

15 years ago - removed subsurf.c from SConscript
Daniel Dunbar [Mon, 4 Apr 2005 04:21:01 +0000 (04:21 +0000)]
 - removed subsurf.c from SConscript

15 years ago - moved subsurf_calculate_limit_positions to using ccgsubsurf
Daniel Dunbar [Mon, 4 Apr 2005 04:20:32 +0000 (04:20 +0000)]
 - moved subsurf_calculate_limit_positions to using ccgsubsurf
 - removed ME_CCG_SUBSURF define
 - dropped CCGSubSurf from editing menu... subsurf is always
   CCG now.

 NOTE: If you saved a file with CCGSubSurf set on the button
 will show up blank (will still work as a subsurf). Just change
 type back to Catmull-Clark.

 Wave bye-bye to HyperMesh...

15 years ago - work-in-progress: subsurf level 0 now just acts like a regular mesh,
Daniel Dunbar [Mon, 4 Apr 2005 03:38:21 +0000 (03:38 +0000)]
 - work-in-progress: subsurf level 0 now just acts like a regular mesh,
   before it would follow subsurf code path by creating fake displistmesh,
   etc... kinda fun for testing but just annoying now.
 - change to creasing behavior, to make sure that with full creasing effect
   is just simple subdivision.

15 years ago - kinda crazy, but replaced old ME_SIMPLE_SUBDIV method with just
Daniel Dunbar [Mon, 4 Apr 2005 02:55:56 +0000 (02:55 +0000)]
 - kinda crazy, but replaced old ME_SIMPLE_SUBDIV method with just
   using regular ccgsubsurf but with crease set to full.

15 years agoAdded a NO_NOR_RECALC context flag. With Individual extrude, the normals are already...
Martin Poirier [Mon, 4 Apr 2005 02:16:56 +0000 (02:16 +0000)]
Added a NO_NOR_RECALC context flag. With Individual extrude, the normals are already set by the extrude code, recalculating breaks them.

Cosmetic changes, renamed CTX_NOPET to CTX_NO_PET

CKey during transform now clears the constraint.

15 years ago - sorry, forgot to save before commit, part of vertex normal fix
Daniel Dunbar [Sun, 3 Apr 2005 22:02:03 +0000 (22:02 +0000)]
 - sorry, forgot to save before commit, part of vertex normal fix

15 years agoRemoves global variable access from TransData conversion fonctions.
Martin Poirier [Sun, 3 Apr 2005 21:52:11 +0000 (21:52 +0000)]
Removes global variable access from TransData conversion fonctions.

Added a NULL check on TransInfo->data before sorting. At that point, the conversion could have been cancel if there's no selected elements (caused segfaults with PET).
Another NULL check on TransInfo->data in postTrans fonction, since that is called on cancel on no selection too.

Was missing a matrix multiplication when setting TransData->dist on meshes which caused PET on scaled meshes to be bad. (bug report 2395: )

15 years ago - added vertex normal calculation functionality to CCG subsurf (CCGSS
Daniel Dunbar [Sun, 3 Apr 2005 21:52:10 +0000 (21:52 +0000)]
 - added vertex normal calculation functionality to CCG subsurf (CCGSS
   is fast, but can be rather tedious to work with...)
 - vertex normals for smooth faces draw correctly now... this code also
   switched to drawing with GL_QUAD_STRIP which can be quite a bit faster
   (depends how fast graphics card calcs lighting, I get 50% faster here)

15 years ago - BLI_STRUCT_OFFSET was not defined correctly, fixed
Daniel Dunbar [Sun, 3 Apr 2005 21:45:43 +0000 (21:45 +0000)]
 - BLI_STRUCT_OFFSET was not defined correctly, fixed

15 years agoError in Softbody goal-min and goal-max correction. It was giving
Ton Roosendaal [Sun, 3 Apr 2005 20:57:31 +0000 (20:57 +0000)]
Error in Softbody goal-min and goal-max correction. It was giving
not as result the range as indicated!

15 years agoSoftBody:
Ton Roosendaal [Sun, 3 Apr 2005 20:13:10 +0000 (20:13 +0000)]

- on add new SoftBody, it creates automatically edges in Mesh now, gives
  too confusing results otherwise
- if no edges exist in mesh, it also doesnt add diagonals for faces in


- fixed stupid 'used unitialized' gcc warning (sorry theeth!)
- made dualAxisConstraint() accept string too
- little cleanup of prints in using Manipulators

15 years agoAdded more dependencies to softbody, for correct recalc
Ton Roosendaal [Sun, 3 Apr 2005 16:57:16 +0000 (16:57 +0000)]
Added more dependencies to softbody, for correct recalc
For example; subsurfed softbody after play anim (ALT+A) didn't get
cleared correctly. Darn where_is_object stuff!

15 years ago - hidden edges were being drawn when DRAWEDGES was off
Daniel Dunbar [Sun, 3 Apr 2005 15:13:38 +0000 (15:13 +0000)]
 - hidden edges were being drawn when DRAWEDGES was off

15 years agoNew minimal & 2D drawing type for Manipulators; try F10 "rt" on 5 (or 6 for
Ton Roosendaal [Sun, 3 Apr 2005 13:12:08 +0000 (13:12 +0000)]
New minimal & 2D drawing type for Manipulators; try F10 "rt" on 5 (or 6 for
same style in trackball arcs)

Full collection can be found temporally here;

15 years agoHidden verts using mesh connected PET caused segfaults.
Martin Poirier [Sun, 3 Apr 2005 06:09:31 +0000 (06:09 +0000)]
Hidden verts using mesh connected PET caused segfaults.
Was caused by the scratch array offsets counting hidden verts when, obviously, it shouldn't.

Also simplified the check for non connected when setting dist. This *might* have caused some errors.

15 years agoupdate to add subsurf csg stuff
Joseph Gilbert [Sat, 2 Apr 2005 20:26:12 +0000 (20:26 +0000)]
update to add subsurf csg stuff

15 years agoMore softbody goodness;
Ton Roosendaal [Sat, 2 Apr 2005 19:52:32 +0000 (19:52 +0000)]
More softbody goodness;

- Added Softbody effect for Lattices (not too useful yet without
  vertexgroups though)
- Added default vertex "goal" value + button, to be assigned when no
  vertexgroup exists (or vertex isn't in the group)
- Made softmin and softmax work as documented (defining min and max range)
- made changes in buttons send 'update' signal to softbody


- added Nkey Properties to show lattice coordinates

15 years ago*** empty log message ***
Joseph Gilbert [Sat, 2 Apr 2005 18:42:30 +0000 (18:42 +0000)]
*** empty log message ***

15 years agoJacked up max resolu and resolv for Surface and Curve to 1024
Ton Roosendaal [Sat, 2 Apr 2005 18:42:16 +0000 (18:42 +0000)]
Jacked up max resolu and resolv for Surface and Curve to 1024

15 years ago - fix for drawing of hidden edges in editmode (w/ optimal off)
Daniel Dunbar [Sat, 2 Apr 2005 18:37:18 +0000 (18:37 +0000)]
 - fix for drawing of hidden edges in editmode (w/ optimal off)
   kinda hacky because it is in derivedmesh implementation not drawing
   logic where I would prefer it to be.

15 years ago - mesh was drawing edges underneath selected edges in optimal mode,
Daniel Dunbar [Sat, 2 Apr 2005 18:21:50 +0000 (18:21 +0000)]
 - mesh was drawing edges underneath selected edges in optimal mode,
   waste of time!
 - made solid face drawing respect hidden faces
 - made selection work for hidden face drawing when select visible is on

15 years ago - fix unused variable warning
Daniel Dunbar [Sat, 2 Apr 2005 18:20:03 +0000 (18:20 +0000)]
 - fix unused variable warning

15 years agoA little housekeeping:
Stephen Swaney [Sat, 2 Apr 2005 17:01:56 +0000 (17:01 +0000)]
A little housekeeping:
Object.c - expand evil 1 character name.  nicer error msg.
doc/ - clarify Object.getData() doc

15 years ago - add include for allqueue prototype
Daniel Dunbar [Sat, 2 Apr 2005 15:56:16 +0000 (15:56 +0000)]
 - add include for allqueue prototype

15 years agoAdded signal OB_SB_REDO for ob->softflag to force a recalc.
Ton Roosendaal [Sat, 2 Apr 2005 15:55:15 +0000 (15:55 +0000)]
Added signal OB_SB_REDO for ob->softflag to force a recalc.
Now used after weight painting, so it updates immediately.

15 years ago - got tired of str[n]casecmp not declared warnings
Daniel Dunbar [Sat, 2 Apr 2005 15:36:57 +0000 (15:36 +0000)]
 - got tired of str[n]casecmp not declared warnings
 - added BLI_str[n]casecmp, use instead of regular versions
 - rewrote BLI_str[n]casecmp to not be stupid

15 years agoIntegration stage of Softbody project
Ton Roosendaal [Sat, 2 Apr 2005 13:57:23 +0000 (13:57 +0000)]
Integration stage of Softbody project

User level notes are in Wiki here;
And will be added in CMS later.

Tech level notes are still pending, but here's the most relevant ones;

- made ob->soft struct SoftBody to hold all settings, and read/save in
- added (temporal!) conversion for the old settings. So: read old files
  with softbody experiments now, and save over!
- cleaned API calls for softbody, which are only 5 of them now:
  sbObjectStep()          (animation steps)
  sbObjectToSoftbody()    (full re-initialize data)
  sbObjectReset()         (only reset motion)
- API calls accepts time in frames now, within softbody.c it converts

Further, internally code was cleaned some (missing tabs etc). Also tried
to keep a well defined structure with hints how to add support for more
objects. Can write notes about that...

15 years agoMedian center with PET was wrong (reported by JMS).
Martin Poirier [Fri, 1 Apr 2005 20:04:24 +0000 (20:04 +0000)]
Median center with PET was wrong (reported by JMS).

Was dividing running sum of vectors by the total number of elements, not by the number of selected elements, as it should have.

15 years ago - bug fix for old subsurf, crash when using meshes with free vertices
Daniel Dunbar [Fri, 1 Apr 2005 19:51:28 +0000 (19:51 +0000)]
 - bug fix for old subsurf, crash when using meshes with free vertices

15 years ago - pickle transform didn't work well, backed out...
Daniel Dunbar [Fri, 1 Apr 2005 19:15:45 +0000 (19:15 +0000)]
 - pickle transform didn't work well, backed out...

15 years agoRemoved casting warnings from transform_manipulator.
Martin Poirier [Fri, 1 Apr 2005 18:39:39 +0000 (18:39 +0000)]
Removed casting warnings from transform_manipulator.
PET is now checked as a transform flag during transform. Much cleaner that way for NOPET context (extrude, duplicate)
Added Sphere as a new PET falloff mode (icon and all) and reordered them around in a more logical fashion (IMHO).

Also brought back Subsurf toggle with Shift-O (was supposed to be for PET object mode but that was canned).

15 years agoTeeth's msvc6 projectfiles commit solved almost all my bf compile probs, only one...
Rob Haarsma [Fri, 1 Apr 2005 11:26:03 +0000 (11:26 +0000)]
Teeth's msvc6 projectfiles commit solved almost all my bf compile probs, only one problem remained;
the including of <MT_assert.h> in the NG_network project. So here's a fix.

15 years ago - some code for testing the pickle matrix
Daniel Dunbar [Fri, 1 Apr 2005 06:13:30 +0000 (06:13 +0000)]
 - some code for testing the pickle matrix

15 years ago - part of creasing changes
Daniel Dunbar [Fri, 1 Apr 2005 03:02:13 +0000 (03:02 +0000)]
 - part of creasing changes

15 years ago - apply spike crease bug fix to ccgsubsurf
Daniel Dunbar [Fri, 1 Apr 2005 02:52:55 +0000 (02:52 +0000)]
 - apply spike crease bug fix to ccgsubsurf

15 years agoSince new transform and subsurf are default now.. no more need for the
Chris Burt [Fri, 1 Apr 2005 02:28:39 +0000 (02:28 +0000)]
Since new transform and subsurf are default now.. no more need for the
defines in the makefiles. Removed to prevent gunky buildup.

15 years ago - take out compile time conditional behavior of creasing for ccg
Daniel Dunbar [Fri, 1 Apr 2005 02:07:34 +0000 (02:07 +0000)]
 - take out compile time conditional behavior of creasing for ccg
   subsurf (always on)
 - bug fix, creasing wasn't evaluated properly after incremental change

15 years agoSacre bleh another bugfix!
Martin Poirier [Fri, 1 Apr 2005 01:25:00 +0000 (01:25 +0000)]
Sacre bleh another bugfix!

Mesh Connected PET was acting weird on geometry that curved on itself (reported by Desoto). Turned out the part of the floodfill working on verts already connected was just doing one case out of the two useful ones (little drawings helped :P).

15 years agoPush/Pull (aka: non proportionnal resize, original idea by Robert Tiess) slipped...
Martin Poirier [Thu, 31 Mar 2005 22:18:08 +0000 (22:18 +0000)]
Push/Pull (aka: non proportionnal resize, original idea by Robert Tiess) slipped in my last commit a bit, so might as well finish it.

Accesible with Shift-P (was free), the behavior is much like Shrink/Fatten. Pull mouse toward you to pull toward center, push away to push away. Constraint and PET works with it too, of course.

15 years agoTransform fixes
Martin Poirier [Thu, 31 Mar 2005 21:02:35 +0000 (21:02 +0000)]
Transform fixes

Connected PET for mesh had an infinite loop in some case. Fixed an made a bit faster.
Renamed the numinput flag.
Corrected a conflict in TransInfo flags. This was a crash waiting to happen.
NO_CONSTRAINT flag now correctly used to disable MMB and constraint hotkeys when needed.

Added check for the manipulator flag when drawing them before applying transformation matrix. Was really weird when doing MMB with rotation.

15 years agoChanging hard codec max frame release for a define (in blendef.h with the other maxim...
Martin Poirier [Thu, 31 Mar 2005 18:49:52 +0000 (18:49 +0000)]
Changing hard codec max frame release for a define (in blendef.h with the other maximums).

Tried to change it everywhere, hopefully I didn't miss any.

Also raised max frame to 30000 (maintainer of that code can safely modify it back since it is in one spot now ;) )
Needed that limit for a marketing video contract, might as well commit it.

15 years agoSome MSVC 6.0 projectfiles updates:
Martin Poirier [Thu, 31 Mar 2005 18:39:38 +0000 (18:39 +0000)]
Some MSVC 6.0 projectfiles updates:

BIF_transform.h added in BL_src
Adding moto include paths to EXP_expressions
SDL include paths and correct python version for all targets in SCA_gamelogic
KX_RayCast.cpp/h in KX_ketsji
SumoPHYCallbackBridge.cpp/.h in PHY_Sumo

15 years agoAdding CCG stuff to MSVC 6.0 projectfiles
Martin Poirier [Thu, 31 Mar 2005 18:22:43 +0000 (18:22 +0000)]
Adding CCG stuff to MSVC 6.0 projectfiles

15 years ago - potential fix for crash on amd64
Daniel Dunbar [Thu, 31 Mar 2005 17:15:44 +0000 (17:15 +0000)]
 - potential fix for crash on amd64

15 years ago - switch to drawing aging in and out of optimal mode
Daniel Dunbar [Thu, 31 Mar 2005 16:54:47 +0000 (16:54 +0000)]
 - switch to drawing aging in and out of optimal mode
 - switch to not using arena allocator in editmode for ccgsubsurf (memory
   doesn't get freed during incremental)

15 years agoRemove USE_CCGSUBSURFLIB stuff since it is always on now.
Stephen Swaney [Thu, 31 Mar 2005 16:21:08 +0000 (16:21 +0000)]
Remove USE_CCGSUBSURFLIB stuff since it is always on now.

15 years ago - disable overriding color (to show aging) during selection drawing
Daniel Dunbar [Thu, 31 Mar 2005 16:16:22 +0000 (16:16 +0000)]
 - disable overriding color (to show aging) during selection drawing

15 years ago - fix warnings for gcc
Daniel Dunbar [Thu, 31 Mar 2005 16:08:29 +0000 (16:08 +0000)]
 - fix warnings for gcc
 - some subsurf settings not correctly changed in editmode

15 years ago - adapt to reprototyping in BKE_DerivedMesh.h
Daniel Dunbar [Thu, 31 Mar 2005 15:57:18 +0000 (15:57 +0000)]
 - adapt to reprototyping in BKE_DerivedMesh.h

15 years ago - adapt caching check for G.editMesh->derived
Daniel Dunbar [Thu, 31 Mar 2005 15:55:31 +0000 (15:55 +0000)]
 - adapt caching check for G.editMesh->derived

15 years ago - removed USE_CCGSUBSURF define (always on now)
Daniel Dunbar [Thu, 31 Mar 2005 15:44:05 +0000 (15:44 +0000)]
 - removed USE_CCGSUBSURF define (always on now)
 - added G.editMesh->derived pointer... idea is to use this
   for mesh derived from editmesh instead of Mesh->derived (as the
   derived mesh tends to vary depending on what it came from).

   This part could be cleaner, also there may problems with it not
   being invalidated correctly.

 - And most importantly: In case people were beginning to worry all
   these edits were just crazy zr stuff, the big point comes about:
   Incremental subsurf calculation is now enabled. This gives massive
   speed improvements when editing a large mesh.

   For the eye-candy happy: try setting G.rt==52 before entering editmode
   and the edges and vertices (in optimal mode) will switch to displaying
   visually the age since a region has last been calculated. Lots of

15 years ago - fix some signed/unsigned comparison warnings
Daniel Dunbar [Thu, 31 Mar 2005 15:15:12 +0000 (15:15 +0000)]
 - fix some signed/unsigned comparison warnings

15 years ago - perhaps I should commit a version that compiles
Daniel Dunbar [Thu, 31 Mar 2005 13:56:50 +0000 (13:56 +0000)]
 - perhaps I should commit a version that compiles

15 years ago - committed VecLength function for theeth
Daniel Dunbar [Thu, 31 Mar 2005 13:44:23 +0000 (13:44 +0000)]
 - committed VecLength function for theeth

15 years agoDifferent floodfill method for PET connected with mesh. More accurate but needs two...
Martin Poirier [Wed, 30 Mar 2005 23:19:42 +0000 (23:19 +0000)]
Different floodfill method for PET connected with mesh. More accurate but needs two temporary scratch arrays to keep vectors and nearest element pointers.

Will comment the code more later tonight, I'm a bit pissed at the moment (stupid errors right in your face. grrrr)

15 years ago - zero derived and decimated pointers on copy
Daniel Dunbar [Wed, 30 Mar 2005 18:06:42 +0000 (18:06 +0000)]
 - zero derived and decimated pointers on copy

15 years ago - added dlm->medge for displistmesh_from_mesh
Daniel Dunbar [Wed, 30 Mar 2005 17:40:51 +0000 (17:40 +0000)]
 - added dlm->medge for displistmesh_from_mesh

15 years ago - Claims of speed improvement last night were slightly
Daniel Dunbar [Wed, 30 Mar 2005 17:32:23 +0000 (17:32 +0000)]
 - Claims of speed improvement last night were slightly
   overrated. Turns out there was a bug in recalc code where
   in editmode mesh was getting reevaluated all the time (because
   ob->disp doesn't get built in editmode). This should fix.
 - Fix some normal calculation/backwards face drawing