2005-04-10 18:33:19 +00:00
|
|
|
/**
|
|
|
|
* $Id$
|
|
|
|
*
|
|
|
|
* ***** BEGIN GPL/BL DUAL LICENSE BLOCK *****
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License
|
|
|
|
* as published by the Free Software Foundation; either version 2
|
|
|
|
* of the License, or (at your option) any later version. The Blender
|
|
|
|
* Foundation also sells licenses for use in proprietary software under
|
|
|
|
* the Blender License. See http://www.blender.org/BL/ for information
|
|
|
|
* about this.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software Foundation,
|
|
|
|
* Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
|
|
|
*
|
|
|
|
* The Original Code is Copyright (C) 2001-2002 by NaN Holding BV.
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* The Original Code is: all of this file.
|
|
|
|
*
|
|
|
|
* Contributor(s): none yet.
|
|
|
|
*
|
|
|
|
* ***** END GPL/BL DUAL LICENSE BLOCK *****
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifdef HAVE_CONFIG_H
|
|
|
|
#include <config.h>
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef WIN32
|
|
|
|
#include <unistd.h>
|
|
|
|
#else
|
|
|
|
#include <io.h>
|
|
|
|
#endif
|
2005-08-23 18:13:30 +00:00
|
|
|
#include <string.h>
|
|
|
|
#include <math.h>
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
#include "MEM_guardedalloc.h"
|
|
|
|
|
|
|
|
#include "DNA_action_types.h"
|
|
|
|
#include "DNA_armature_types.h"
|
|
|
|
#include "DNA_camera_types.h"
|
|
|
|
#include "DNA_curve_types.h"
|
|
|
|
#include "DNA_effect_types.h"
|
|
|
|
#include "DNA_image_types.h"
|
|
|
|
#include "DNA_ipo_types.h"
|
|
|
|
#include "DNA_key_types.h"
|
|
|
|
#include "DNA_lamp_types.h"
|
|
|
|
#include "DNA_lattice_types.h"
|
|
|
|
#include "DNA_mesh_types.h"
|
|
|
|
#include "DNA_meshdata_types.h"
|
|
|
|
#include "DNA_meta_types.h"
|
2005-08-11 02:23:52 +00:00
|
|
|
#include "DNA_modifier_types.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "DNA_object_types.h"
|
2005-05-04 12:53:35 +00:00
|
|
|
#include "DNA_object_force.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "DNA_scene_types.h"
|
|
|
|
#include "DNA_screen_types.h"
|
2005-08-23 18:13:30 +00:00
|
|
|
#include "DNA_space_types.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "DNA_texture_types.h"
|
|
|
|
#include "DNA_view3d_types.h"
|
|
|
|
#include "DNA_world_types.h"
|
|
|
|
#include "DNA_userdef_types.h"
|
|
|
|
#include "DNA_property_types.h"
|
|
|
|
#include "DNA_vfont_types.h"
|
|
|
|
#include "DNA_constraint_types.h"
|
|
|
|
|
|
|
|
#include "BKE_action.h"
|
|
|
|
#include "BKE_armature.h"
|
|
|
|
#include "BKE_blender.h"
|
|
|
|
#include "BKE_curve.h"
|
2005-05-04 12:53:35 +00:00
|
|
|
#include "BKE_constraint.h"
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
#include "BKE_depsgraph.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BKE_displist.h"
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
#include "BKE_DerivedMesh.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BKE_effect.h"
|
2005-05-04 12:53:35 +00:00
|
|
|
#include "BKE_font.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BKE_global.h"
|
|
|
|
#include "BKE_ipo.h"
|
|
|
|
#include "BKE_lattice.h"
|
|
|
|
#include "BKE_mball.h"
|
2005-08-23 18:13:30 +00:00
|
|
|
#include "BKE_mesh.h"
|
2005-08-11 02:23:52 +00:00
|
|
|
#include "BKE_modifier.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BKE_object.h"
|
2005-05-04 12:53:35 +00:00
|
|
|
#include "BKE_softbody.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BKE_utildefines.h"
|
|
|
|
|
2005-07-27 10:37:20 +00:00
|
|
|
#include "BIF_editaction.h"
|
|
|
|
#include "BIF_editview.h"
|
|
|
|
#include "BIF_editlattice.h"
|
|
|
|
#include "BIF_editconstraint.h"
|
|
|
|
#include "BIF_editarmature.h"
|
|
|
|
#include "BIF_editmesh.h"
|
2005-08-23 18:13:30 +00:00
|
|
|
#include "BIF_editsima.h"
|
2005-07-27 10:37:20 +00:00
|
|
|
#include "BIF_gl.h"
|
|
|
|
#include "BIF_poseobject.h"
|
2005-12-01 19:04:57 +00:00
|
|
|
#include "BIF_meshtools.h"
|
2005-07-27 10:37:20 +00:00
|
|
|
#include "BIF_mywindow.h"
|
|
|
|
#include "BIF_resources.h"
|
|
|
|
#include "BIF_screen.h"
|
|
|
|
#include "BIF_space.h"
|
|
|
|
#include "BIF_toolbox.h"
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BSE_view.h"
|
|
|
|
#include "BSE_edit.h"
|
|
|
|
#include "BSE_editipo.h"
|
|
|
|
#include "BSE_editipo_types.h"
|
2005-05-04 12:53:35 +00:00
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BDR_editobject.h" // reset_slowparents()
|
2005-12-07 12:36:26 +00:00
|
|
|
#include "BDR_unwrapper.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
#include "BLI_arithb.h"
|
2005-10-30 13:50:42 +00:00
|
|
|
#include "BLI_blenlib.h"
|
2005-04-10 18:33:19 +00:00
|
|
|
#include "BLI_editVert.h"
|
|
|
|
|
|
|
|
#include "blendef.h"
|
|
|
|
|
|
|
|
#include "mydevice.h"
|
|
|
|
|
|
|
|
extern ListBase editNurb;
|
|
|
|
extern ListBase editelems;
|
|
|
|
|
|
|
|
#include "transform.h"
|
|
|
|
|
2005-05-04 12:53:35 +00:00
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
/* ************************** Functions *************************** */
|
|
|
|
|
|
|
|
static void qsort_trans_data(TransInfo *t, TransData *head, TransData *tail) {
|
|
|
|
TransData pivot = *head;
|
|
|
|
TransData *ihead = head;
|
|
|
|
TransData *itail = tail;
|
|
|
|
short connected = t->flag & T_PROP_CONNECTED;
|
|
|
|
|
|
|
|
while (head < tail)
|
|
|
|
{
|
|
|
|
if (connected) {
|
|
|
|
while ((tail->dist >= pivot.dist) && (head < tail))
|
|
|
|
tail--;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
while ((tail->rdist >= pivot.rdist) && (head < tail))
|
|
|
|
tail--;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (head != tail)
|
|
|
|
{
|
|
|
|
*head = *tail;
|
|
|
|
head++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (connected) {
|
|
|
|
while ((head->dist <= pivot.dist) && (head < tail))
|
|
|
|
head++;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
while ((head->rdist <= pivot.rdist) && (head < tail))
|
|
|
|
head++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (head != tail)
|
|
|
|
{
|
|
|
|
*tail = *head;
|
|
|
|
tail--;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
*head = pivot;
|
|
|
|
if (ihead < head) {
|
|
|
|
qsort_trans_data(t, ihead, head-1);
|
|
|
|
}
|
|
|
|
if (itail > head) {
|
|
|
|
qsort_trans_data(t, head+1, itail);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void sort_trans_data_dist(TransInfo *t) {
|
|
|
|
TransData *start = t->data;
|
|
|
|
int i = 1;
|
|
|
|
|
|
|
|
while(i < t->total && start->flag & TD_SELECTED) {
|
|
|
|
start++;
|
|
|
|
i++;
|
|
|
|
}
|
|
|
|
qsort_trans_data(t, start, t->data + t->total - 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void sort_trans_data(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *sel, *unsel;
|
|
|
|
TransData temp;
|
|
|
|
unsel = t->data;
|
|
|
|
sel = t->data;
|
|
|
|
sel += t->total - 1;
|
|
|
|
while (sel > unsel) {
|
|
|
|
while (unsel->flag & TD_SELECTED) {
|
|
|
|
unsel++;
|
|
|
|
if (unsel == sel) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
while (!(sel->flag & TD_SELECTED)) {
|
|
|
|
sel--;
|
|
|
|
if (unsel == sel) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
temp = *unsel;
|
|
|
|
*unsel = *sel;
|
|
|
|
*sel = temp;
|
|
|
|
sel--;
|
|
|
|
unsel++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* distance calculated from not-selected vertex to nearest selected vertex
|
|
|
|
warning; this is loops inside loop, has minor N^2 issues, but by sorting list it is OK */
|
|
|
|
static void set_prop_dist(TransInfo *t, short with_dist)
|
|
|
|
{
|
|
|
|
TransData *tob;
|
|
|
|
int a;
|
2005-08-23 18:13:30 +00:00
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
for(a=0, tob= t->data; a<t->total; a++, tob++) {
|
|
|
|
|
|
|
|
tob->rdist= 0.0f; // init, it was mallocced
|
|
|
|
|
|
|
|
if((tob->flag & TD_SELECTED)==0) {
|
|
|
|
TransData *td;
|
|
|
|
int i;
|
|
|
|
float dist, vec[3];
|
|
|
|
|
|
|
|
tob->rdist = -1.0f; // signal for next loop
|
|
|
|
|
|
|
|
for (i = 0, td= t->data; i < t->total; i++, td++) {
|
|
|
|
if(td->flag & TD_SELECTED) {
|
|
|
|
VecSubf(vec, tob->center, td->center);
|
|
|
|
Mat3MulVecfl(tob->mtx, vec);
|
|
|
|
dist = Normalise(vec);
|
|
|
|
if (tob->rdist == -1.0f) {
|
|
|
|
tob->rdist = dist;
|
|
|
|
}
|
|
|
|
else if (dist < tob->rdist) {
|
|
|
|
tob->rdist = dist;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else break; // by definition transdata has selected items in beginning
|
|
|
|
}
|
|
|
|
if (with_dist) {
|
|
|
|
tob->dist = tob->rdist;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ************************** CONVERSIONS ************************* */
|
|
|
|
|
|
|
|
/* ********************* texture space ********* */
|
|
|
|
|
|
|
|
static void createTransTexspace(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *td;
|
|
|
|
Object *ob;
|
|
|
|
ID *id;
|
|
|
|
|
|
|
|
ob= OBACT;
|
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.
2005-07-24 06:25:00 +00:00
|
|
|
|
|
|
|
id= ob->data;
|
|
|
|
if(id==NULL || !ELEM3( GS(id->name), ID_ME, ID_CU, ID_MB )) {
|
|
|
|
t->total = 0;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
t->total = 1;
|
|
|
|
td= t->data= MEM_callocN(sizeof(TransData), "TransTexspace");
|
|
|
|
td->ext= t->ext= MEM_callocN(sizeof(TransDataExtension), "TransTexspace");
|
|
|
|
|
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
VECCOPY(td->center, ob->obmat[3]);
|
|
|
|
td->ob = ob;
|
|
|
|
|
|
|
|
Mat3CpyMat4(td->mtx, ob->obmat);
|
|
|
|
Mat3Inv(td->smtx, td->mtx);
|
|
|
|
|
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.
2005-07-24 06:25:00 +00:00
|
|
|
if( GS(id->name)==ID_ME) {
|
2005-04-10 18:33:19 +00:00
|
|
|
Mesh *me= ob->data;
|
|
|
|
me->texflag &= ~AUTOSPACE;
|
|
|
|
td->loc= me->loc;
|
|
|
|
td->ext->rot= me->rot;
|
|
|
|
td->ext->size= me->size;
|
|
|
|
}
|
|
|
|
else if( GS(id->name)==ID_CU) {
|
|
|
|
Curve *cu= ob->data;
|
|
|
|
cu->texflag &= ~CU_AUTOSPACE;
|
|
|
|
td->loc= cu->loc;
|
|
|
|
td->ext->rot= cu->rot;
|
|
|
|
td->ext->size= cu->size;
|
|
|
|
}
|
|
|
|
else if( GS(id->name)==ID_MB) {
|
|
|
|
MetaBall *mb= ob->data;
|
|
|
|
mb->texflag &= ~MB_AUTOSPACE;
|
|
|
|
td->loc= mb->loc;
|
|
|
|
td->ext->rot= mb->rot;
|
|
|
|
td->ext->size= mb->size;
|
|
|
|
}
|
|
|
|
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
VECCOPY(td->ext->irot, td->ext->rot);
|
|
|
|
VECCOPY(td->ext->isize, td->ext->size);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* edge (for crease) ***** */
|
|
|
|
|
|
|
|
static void createTransEdge(TransInfo *t) {
|
|
|
|
TransData *td = NULL;
|
|
|
|
EditMesh *em = G.editMesh;
|
|
|
|
EditEdge *eed;
|
|
|
|
float mtx[3][3], smtx[3][3];
|
|
|
|
int count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
|
|
|
|
|
|
|
for(eed= em->edges.first; eed; eed= eed->next) {
|
|
|
|
if(eed->h==0) {
|
|
|
|
if (eed->f & SELECT) countsel++;
|
|
|
|
if (propmode) count++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (countsel == 0)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if(propmode) {
|
|
|
|
t->total = count;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
t->total = countsel;
|
|
|
|
}
|
|
|
|
|
|
|
|
td= t->data= MEM_callocN(t->total * sizeof(TransData), "TransCrease");
|
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
|
|
|
for(eed= em->edges.first; eed; eed= eed->next) {
|
|
|
|
if(eed->h==0 && (eed->f & SELECT || propmode)) {
|
|
|
|
/* need to set center for center calculations */
|
|
|
|
VecAddf(td->center, eed->v1->co, eed->v2->co);
|
|
|
|
VecMulf(td->center, 0.5f);
|
|
|
|
|
|
|
|
td->loc= NULL;
|
|
|
|
if (eed->f & SELECT)
|
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
else
|
|
|
|
td->flag= 0;
|
|
|
|
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = &(eed->crease);
|
|
|
|
td->ival = eed->crease;
|
|
|
|
|
|
|
|
td++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* pose mode ************* */
|
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
static bKinematicConstraint *has_targetless_ik(bPoseChannel *pchan)
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
{
|
2005-10-30 13:50:42 +00:00
|
|
|
bConstraint *con= pchan->constraints.first;
|
|
|
|
|
|
|
|
for(;con; con= con->next) {
|
|
|
|
if(con->type==CONSTRAINT_TYPE_KINEMATIC) {
|
|
|
|
bKinematicConstraint *data= con->data;
|
|
|
|
|
|
|
|
if(data->tar==NULL)
|
|
|
|
return data;
|
|
|
|
if(data->tar->type==OB_ARMATURE && data->subtarget[0]==0)
|
|
|
|
return data;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
static void apply_targetless_ik(Object *ob)
|
|
|
|
{
|
|
|
|
bPoseChannel *pchan, *parchan, *chanlist[256];
|
|
|
|
bKinematicConstraint *data;
|
|
|
|
int segcount;
|
|
|
|
|
|
|
|
/* now we got a difficult situation... we have to find the
|
|
|
|
target-less IK pchans, and apply transformation to the all
|
|
|
|
pchans that were in the chain */
|
|
|
|
|
|
|
|
for (pchan=ob->pose->chanbase.first; pchan; pchan=pchan->next) {
|
|
|
|
data= has_targetless_ik(pchan);
|
|
|
|
if(data && (data->flag & CONSTRAINT_IK_AUTO)) {
|
|
|
|
|
|
|
|
/* fill the array with the bones of the chain (armature.c does same, keep it synced) */
|
|
|
|
segcount= 0;
|
|
|
|
|
|
|
|
/* exclude tip from chain? */
|
|
|
|
if(!(data->flag & CONSTRAINT_IK_TIP))
|
|
|
|
parchan= pchan->parent;
|
|
|
|
else
|
|
|
|
parchan= pchan;
|
|
|
|
|
|
|
|
/* Find the chain's root & count the segments needed */
|
|
|
|
for (; parchan; parchan=parchan->parent){
|
|
|
|
chanlist[segcount]= parchan;
|
|
|
|
segcount++;
|
|
|
|
|
|
|
|
if(segcount==data->rootbone || segcount>255) break; // 255 is weak
|
|
|
|
}
|
|
|
|
for(;segcount;segcount--) {
|
|
|
|
Bone *bone;
|
|
|
|
float rmat[4][4], tmat[4][4], imat[4][4];
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* pose_mat(b) = pose_mat(b-1) * offs_bone * channel * constraint * IK */
|
|
|
|
/* we put in channel the entire result of rmat= (channel * constraint * IK) */
|
|
|
|
/* pose_mat(b) = pose_mat(b-1) * offs_bone * rmat */
|
|
|
|
/* rmat = pose_mat(b) * inv( pose_mat(b-1) * offs_bone ) */
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
parchan= chanlist[segcount-1];
|
|
|
|
bone= parchan->bone;
|
2005-11-17 22:07:06 +00:00
|
|
|
bone->flag |= BONE_TRANSFORM; /* ensures it gets an auto key inserted */
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
if(parchan->parent) {
|
2005-11-05 20:02:04 +00:00
|
|
|
Bone *parbone= parchan->parent->bone;
|
2005-10-30 13:50:42 +00:00
|
|
|
float offs_bone[4][4];
|
|
|
|
|
|
|
|
/* offs_bone = yoffs(b-1) + root(b) + bonemat(b) */
|
|
|
|
Mat4CpyMat3(offs_bone, bone->bone_mat);
|
|
|
|
|
|
|
|
/* The bone's root offset (is in the parent's coordinate system) */
|
|
|
|
VECCOPY(offs_bone[3], bone->head);
|
|
|
|
|
|
|
|
/* Get the length translation of parent (length along y axis) */
|
2005-11-05 20:02:04 +00:00
|
|
|
offs_bone[3][1]+= parbone->length;
|
2005-10-30 13:50:42 +00:00
|
|
|
|
|
|
|
/* pose_mat(b-1) * offs_bone */
|
2005-11-05 20:02:04 +00:00
|
|
|
if(parchan->bone->flag & BONE_HINGE) {
|
|
|
|
/* the rotation of the parent restposition */
|
|
|
|
Mat4CpyMat4(rmat, parbone->arm_mat); /* rmat used as temp */
|
|
|
|
|
|
|
|
/* the location of actual parent transform */
|
|
|
|
VECCOPY(rmat[3], offs_bone[3]);
|
|
|
|
offs_bone[3][0]= offs_bone[3][1]= offs_bone[3][2]= 0.0f;
|
|
|
|
Mat4MulVecfl(parchan->parent->pose_mat, rmat[3]);
|
|
|
|
|
|
|
|
Mat4MulMat4(tmat, offs_bone, rmat);
|
|
|
|
}
|
2005-10-31 10:10:01 +00:00
|
|
|
else
|
|
|
|
Mat4MulMat4(tmat, offs_bone, parchan->parent->pose_mat);
|
Three new features:
1) Stride Bone
For walkcycles, you could already set an NLA strip to cycle over a path
based on a preset distance value. This cycling happens based on a linear
interpolation, with constant speed.
Not all cycles have a constant speed however, like hopping or jumping.
To ensure a perfect slipping-less foot contact, you now can set a Bone
in an Armature to define the stride. This "Stride Bone" then becomes a
sort-of ruler, a conveyor belt, on which the character walks. When using
the NLA "Use Path" option, it then tries to keep the Stride Bone entirely
motionless on the path, by cancelling out its motion (for the entire
Armature). This means that the animation keys for a Stride Bone have to be
exactly negative of the desired path. Only, at choice, the X,Y or Z Ipo
curve is used for this stride.
Examples:
http://www.blender.org/bf/0001_0040.avi
The top armature shows the actual Action, the bottom armature has been
parented to a Path, using the Stride Bone feature.
http://www.blender.org/bf/0001_0080.avi
Here the Stride Bone has a number of children, creating a ruler to be
used as reference while animating.
Test .blend:
http://www.blender.org/bf/motionblender1.blend
Notes:
- Note that action keys for Bones work local, based on the Bone's
orientation as set in EditMode. Therefore, an Y translation always
goes in the Bone's direction.
- To be able to get a "solvable" stride, the animation curve has
to be inverse evaluated, using a Newton Raphson root solver. That
means you can only create stride curves that keep moving forward, and
cannot return halfway.
- Set the Stride Bone in the Editing Buttons, Bone Panel. You can set
change the name or set the axis in the NLA Window, Strip Properties Panel.
- Files in this commit will move to the blender.org release section.
2) Armature Ghosting
In EditButtons, Armature Panel, you can set an armature to draw ghosts.
The number value denotes the amount of frames that have to be drawn extra
(for the active action!) around the current frame.
Ghosts only evaluate its own Pose, executing it's Actions, Constraints and
IK. No external dependencies are re-evaluated for it.
3) NLA/Action time control
If you click in the NLA window on the action (linked to Object), it makes
sure the Timing as drawn in the Action editor is not corrected for NLA.
If you also set the Object to "Action", this timing will be executed on the
Object as well (not NLA time).
(It's a bit confusing... will make a good doc & maybe review UI!)
2005-11-01 12:44:30 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
Mat4Invert(imat, tmat);
|
2005-07-04 05:16:54 +00:00
|
|
|
}
|
|
|
|
else {
|
2005-10-30 13:50:42 +00:00
|
|
|
Mat4CpyMat3(tmat, bone->bone_mat);
|
|
|
|
|
|
|
|
VECCOPY(tmat[3], bone->head);
|
|
|
|
Mat4Invert(imat, tmat);
|
2005-07-04 05:16:54 +00:00
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
/* result matrix */
|
|
|
|
Mat4MulMat4(rmat, parchan->pose_mat, imat);
|
2005-07-19 16:12:56 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* apply and decompose, doesn't work for constraints or non-uniform scale well */
|
|
|
|
{
|
|
|
|
float rmat3[3][3], qmat[3][3], imat[3][3], smat[3][3];
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
Mat3CpyMat4(rmat3, rmat);
|
|
|
|
|
|
|
|
/* quaternion */
|
|
|
|
Mat3ToQuat(rmat3, parchan->quat);
|
|
|
|
|
|
|
|
/* for size, remove rotation */
|
|
|
|
QuatToMat3(parchan->quat, qmat);
|
|
|
|
Mat3Inv(imat, qmat);
|
|
|
|
Mat3MulMat3(smat, rmat3, imat);
|
|
|
|
Mat3ToSize(smat, parchan->size);
|
|
|
|
|
|
|
|
VECCOPY(parchan->loc, rmat[3]);
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
|
|
|
|
data->flag &= ~CONSTRAINT_IK_AUTO;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
static void add_pose_transdata(TransInfo *t, bPoseChannel *pchan, Object *ob, TransData *td)
|
|
|
|
{
|
|
|
|
Bone *bone= pchan->bone;
|
|
|
|
float pmat[3][3], omat[3][3];
|
|
|
|
float vec[3];
|
|
|
|
|
|
|
|
VECCOPY(vec, pchan->pose_mat[3]);
|
|
|
|
VECCOPY(td->center, vec);
|
|
|
|
|
|
|
|
td->ob = ob;
|
|
|
|
td->flag= TD_SELECTED|TD_USEQUAT;
|
|
|
|
td->protectflag= pchan->protectflag;
|
|
|
|
|
|
|
|
td->loc = pchan->loc;
|
|
|
|
VECCOPY(td->iloc, pchan->loc);
|
|
|
|
|
|
|
|
td->ext->rot= NULL;
|
|
|
|
td->ext->quat= pchan->quat;
|
|
|
|
td->ext->size= pchan->size;
|
|
|
|
|
|
|
|
QUATCOPY(td->ext->iquat, pchan->quat);
|
|
|
|
VECCOPY(td->ext->isize, pchan->size);
|
|
|
|
|
|
|
|
/* proper way to get the parent transform + own transform */
|
|
|
|
Mat3CpyMat4(omat, ob->obmat);
|
|
|
|
if(pchan->parent) {
|
|
|
|
if(pchan->bone->flag & BONE_HINGE)
|
|
|
|
Mat3CpyMat4(pmat, pchan->parent->bone->arm_mat);
|
|
|
|
else
|
|
|
|
Mat3CpyMat4(pmat, pchan->parent->pose_mat);
|
|
|
|
|
|
|
|
Mat3MulSerie(td->mtx, pchan->bone->bone_mat, pmat, omat, 0,0,0,0,0); // dang mulserie swaps args
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
Mat3MulMat3(td->mtx, omat, pchan->bone->bone_mat); // huh, transposed?
|
|
|
|
}
|
|
|
|
|
|
|
|
Mat3Inv (td->smtx, td->mtx);
|
|
|
|
|
|
|
|
/* for axismat we use bone's own transform */
|
|
|
|
Mat3CpyMat4(pmat, pchan->pose_mat);
|
|
|
|
Mat3MulMat3(td->axismtx, omat, pmat);
|
|
|
|
Mat3Ortho(td->axismtx);
|
|
|
|
|
|
|
|
if(t->mode==TFM_BONESIZE) {
|
|
|
|
bArmature *arm= t->poseobj->data;
|
|
|
|
|
|
|
|
if(arm->drawtype==ARM_ENVELOPE) {
|
|
|
|
td->loc= NULL;
|
|
|
|
td->val= &bone->dist;
|
|
|
|
td->ival= bone->dist;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
// abusive storage of scale in the loc pointer :)
|
|
|
|
td->loc= &bone->xwidth;
|
|
|
|
VECCOPY (td->iloc, td->loc);
|
|
|
|
td->val= NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* in this case we can do target-less IK grabbing */
|
|
|
|
if(t->mode==TFM_TRANSLATION) {
|
|
|
|
bKinematicConstraint *data= has_targetless_ik(pchan);
|
|
|
|
if(data) {
|
|
|
|
if(data->flag & CONSTRAINT_IK_TIP) {
|
|
|
|
VECCOPY(data->grabtarget, pchan->pose_tail);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
VECCOPY(data->grabtarget, pchan->pose_head);
|
|
|
|
}
|
|
|
|
td->loc = data->grabtarget;
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
data->flag |= CONSTRAINT_IK_AUTO;
|
|
|
|
|
|
|
|
/* only object matrix correction */
|
|
|
|
Mat3CpyMat3 (td->mtx, omat);
|
|
|
|
Mat3Inv (td->smtx, td->mtx);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void bone_children_clear_transflag(ListBase *lb)
|
|
|
|
{
|
|
|
|
Bone *bone= lb->first;
|
|
|
|
|
|
|
|
for(;bone;bone= bone->next) {
|
|
|
|
bone->flag &= ~BONE_TRANSFORM;
|
|
|
|
bone_children_clear_transflag(&bone->childbase);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* sets transform flags in the bones, returns total */
|
|
|
|
static void set_pose_transflags(TransInfo *t, Object *ob)
|
|
|
|
{
|
2005-12-07 12:36:26 +00:00
|
|
|
bArmature *arm= ob->data;
|
2005-10-30 13:50:42 +00:00
|
|
|
bPoseChannel *pchan;
|
|
|
|
Bone *bone;
|
|
|
|
|
|
|
|
t->total= 0;
|
|
|
|
|
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
|
|
|
bone= pchan->bone;
|
2005-12-07 12:36:26 +00:00
|
|
|
if(bone->layer & arm->layer) {
|
|
|
|
if(bone->flag & BONE_SELECTED)
|
|
|
|
bone->flag |= BONE_TRANSFORM;
|
|
|
|
else
|
|
|
|
bone->flag &= ~BONE_TRANSFORM;
|
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* make sure no bone can be transformed when a parent is transformed */
|
|
|
|
/* since pchans are depsgraph sorted, the parents are in beginning of list */
|
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
|
|
|
bone= pchan->bone;
|
|
|
|
if(bone->flag & BONE_TRANSFORM)
|
|
|
|
bone_children_clear_transflag(&bone->childbase);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* now count, and check if we have autoIK or have to switch from translate to rotate */
|
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
|
|
|
bone= pchan->bone;
|
|
|
|
if(bone->flag & BONE_TRANSFORM) {
|
|
|
|
t->total++;
|
|
|
|
|
2005-11-01 19:32:54 +00:00
|
|
|
if(t->mode==TFM_TRANSLATION) {
|
2005-11-17 18:01:07 +00:00
|
|
|
if( has_targetless_ik(pchan)==NULL ) {
|
|
|
|
if(pchan->parent && (pchan->bone->flag & BONE_CONNECTED))
|
|
|
|
t->mode= TFM_ROTATION;
|
|
|
|
else if((pchan->protectflag & OB_LOCK_LOC)==OB_LOCK_LOC)
|
2005-11-01 19:32:54 +00:00
|
|
|
t->mode= TFM_ROTATION;
|
2005-10-30 13:50:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* frees temporal IKs */
|
|
|
|
static void pose_grab_with_ik_clear(Object *ob)
|
|
|
|
{
|
|
|
|
bKinematicConstraint *data;
|
|
|
|
bPoseChannel *pchan;
|
|
|
|
bConstraint *con;
|
|
|
|
|
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
|
|
|
for(con= pchan->constraints.first; con; con= con->next) {
|
|
|
|
if(con->type==CONSTRAINT_TYPE_KINEMATIC) {
|
|
|
|
data= con->data;
|
|
|
|
if(data->flag & CONSTRAINT_IK_TEMP) {
|
|
|
|
BLI_remlink(&pchan->constraints, con);
|
|
|
|
MEM_freeN(con->data);
|
|
|
|
MEM_freeN(con);
|
|
|
|
pchan->constflag &= ~(PCHAN_HAS_IK|PCHAN_HAS_TARGET);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* adds the IK to pchan */
|
|
|
|
static void pose_grab_with_ik_add(bPoseChannel *pchan)
|
|
|
|
{
|
|
|
|
bKinematicConstraint *data;
|
|
|
|
bConstraint *con;
|
|
|
|
|
|
|
|
/* rule: not if there's already an IK on this channel */
|
|
|
|
for(con= pchan->constraints.first; con; con= con->next)
|
|
|
|
if(con->type==CONSTRAINT_TYPE_KINEMATIC)
|
|
|
|
break;
|
|
|
|
|
|
|
|
if(con) {
|
|
|
|
/* but, if this is a targetless IK, we make it auto anyway (for the children loop) */
|
|
|
|
data= has_targetless_ik(pchan);
|
|
|
|
if(data)
|
|
|
|
data->flag |= CONSTRAINT_IK_AUTO;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
con = add_new_constraint(CONSTRAINT_TYPE_KINEMATIC);
|
|
|
|
BLI_addtail(&pchan->constraints, con);
|
|
|
|
pchan->constflag |= (PCHAN_HAS_IK|PCHAN_HAS_TARGET); /* for draw, but also for detecting while pose solving */
|
|
|
|
data= con->data;
|
|
|
|
data->flag= CONSTRAINT_IK_TIP|CONSTRAINT_IK_TEMP|CONSTRAINT_IK_AUTO;
|
|
|
|
VECCOPY(data->grabtarget, pchan->pose_tail);
|
|
|
|
data->rootbone= 1;
|
|
|
|
|
|
|
|
/* we include only a connected chain */
|
|
|
|
while(pchan && (pchan->bone->flag & BONE_CONNECTED)) {
|
|
|
|
data->rootbone++;
|
|
|
|
pchan= pchan->parent;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* bone is a canditate to get IK, but we don't do it if it has children connected */
|
|
|
|
static void pose_grab_with_ik_children(bPose *pose, Bone *bone)
|
|
|
|
{
|
|
|
|
Bone *bonec;
|
|
|
|
int wentdeeper= 0;
|
|
|
|
|
|
|
|
/* go deeper if children & children are connected */
|
|
|
|
for(bonec= bone->childbase.first; bonec; bonec= bonec->next) {
|
|
|
|
if(bonec->flag & BONE_CONNECTED) {
|
|
|
|
wentdeeper= 1;
|
|
|
|
pose_grab_with_ik_children(pose, bonec);
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
if(wentdeeper==0) {
|
|
|
|
bPoseChannel *pchan= get_pose_channel(pose, bone->name);
|
|
|
|
if(pchan)
|
|
|
|
pose_grab_with_ik_add(pchan);
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* main call which adds temporal IK chains */
|
|
|
|
static void pose_grab_with_ik(Object *ob)
|
|
|
|
{
|
2005-12-07 12:36:26 +00:00
|
|
|
bArmature *arm= ob->data;
|
2005-10-30 13:50:42 +00:00
|
|
|
bPoseChannel *pchan, *pchansel= NULL;
|
|
|
|
|
|
|
|
if(ob==NULL || ob->pose==NULL || (ob->flag & OB_POSEMODE)==0)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/* rule: only one Bone */
|
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
2005-12-07 12:36:26 +00:00
|
|
|
if(pchan->bone->layer & arm->layer) {
|
|
|
|
if(pchan->bone->flag & BONE_SELECTED) {
|
|
|
|
if(pchansel)
|
|
|
|
break;
|
|
|
|
pchansel= pchan;
|
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
if(pchan || pchansel==NULL) return;
|
|
|
|
|
|
|
|
/* rule: if selected Bone is not a root bone, it gets a temporal IK */
|
|
|
|
if(pchansel->parent) {
|
|
|
|
/* only adds if there's no IK yet */
|
|
|
|
pose_grab_with_ik_add(pchansel);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
/* rule: go over the children and add IK to the tips */
|
|
|
|
pose_grab_with_ik_children(ob->pose, pchansel->bone);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
2005-07-23 18:52:31 +00:00
|
|
|
/* only called with pose mode active object now */
|
2005-10-30 13:50:42 +00:00
|
|
|
static void createTransPose(TransInfo *t, Object *ob)
|
2005-04-10 18:33:19 +00:00
|
|
|
{
|
|
|
|
bArmature *arm;
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
bPoseChannel *pchan;
|
2005-04-10 18:33:19 +00:00
|
|
|
TransData *td;
|
|
|
|
TransDataExtension *tdx;
|
|
|
|
int i;
|
|
|
|
|
2005-08-29 18:03:03 +00:00
|
|
|
t->total= 0;
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
/* check validity of state */
|
2005-07-23 18:52:31 +00:00
|
|
|
arm=get_armature (ob);
|
|
|
|
if (arm==NULL || ob->pose==NULL) return;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
if (arm->flag & ARM_RESTPOS) {
|
|
|
|
if(t->mode!=TFM_BONESIZE) {
|
|
|
|
notice ("Pose edit not possible while Rest Position is enabled");
|
|
|
|
return;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
2005-07-23 18:52:31 +00:00
|
|
|
if (!(ob->lay & G.vd->lay)) return;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* do we need to add temporal IK chains? */
|
|
|
|
if((arm->flag & ARM_AUTO_IK) && t->mode==TFM_TRANSLATION)
|
|
|
|
pose_grab_with_ik(ob);
|
|
|
|
|
|
|
|
/* set flags and count total (warning, can change transform to rotate) */
|
|
|
|
set_pose_transflags(t, ob);
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
if(t->total==0) return;
|
2005-08-23 18:13:30 +00:00
|
|
|
|
2005-08-16 10:17:02 +00:00
|
|
|
t->flag |= T_POSE;
|
2005-11-01 20:58:37 +00:00
|
|
|
t->poseobj= ob; /* we also allow non-active objects to be transformed, in weightpaint */
|
2005-08-16 10:17:02 +00:00
|
|
|
|
2005-11-01 20:58:37 +00:00
|
|
|
/* make sure the lock is set OK, unlock can be accidentally saved? */
|
|
|
|
ob->pose->flag |= POSE_LOCKED;
|
|
|
|
ob->pose->flag &= ~POSE_DO_UNLOCK;
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
/* init trans data */
|
2005-04-11 15:14:07 +00:00
|
|
|
td = t->data = MEM_callocN(t->total*sizeof(TransData), "TransPoseBone");
|
|
|
|
tdx = t->ext = MEM_callocN(t->total*sizeof(TransDataExtension), "TransPoseBoneExt");
|
2005-04-10 18:33:19 +00:00
|
|
|
for(i=0; i<t->total; i++, td++, tdx++) {
|
|
|
|
td->ext= tdx;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
|
|
|
}
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
|
|
|
|
/* use pose channels to fill trans data */
|
2005-04-10 18:33:19 +00:00
|
|
|
td= t->data;
|
2005-07-23 18:52:31 +00:00
|
|
|
for(pchan= ob->pose->chanbase.first; pchan; pchan= pchan->next) {
|
2005-10-30 13:50:42 +00:00
|
|
|
if(pchan->bone->flag & BONE_TRANSFORM) {
|
|
|
|
add_pose_transdata(t, pchan, ob, td);
|
|
|
|
td++;
|
|
|
|
}
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
if(td != (t->data+t->total)) printf("Bone selection count error\n");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* armature ************** */
|
|
|
|
|
|
|
|
static void createTransArmatureVerts(TransInfo *t)
|
|
|
|
{
|
|
|
|
EditBone *ebo;
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
bArmature *arm= G.obedit->data;
|
2005-04-10 18:33:19 +00:00
|
|
|
TransData *td;
|
2005-07-16 19:07:02 +00:00
|
|
|
float mtx[3][3], smtx[3][3], delta[3], bonemat[3][3];
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
t->total = 0;
|
2005-07-16 19:07:02 +00:00
|
|
|
for (ebo=G.edbo.first;ebo;ebo=ebo->next) {
|
2005-12-07 12:36:26 +00:00
|
|
|
if(ebo->layer & arm->layer) {
|
|
|
|
if (t->mode==TFM_BONESIZE) {
|
|
|
|
if (ebo->flag & BONE_SELECTED)
|
|
|
|
t->total++;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
if (ebo->flag & BONE_TIPSEL)
|
|
|
|
t->total++;
|
|
|
|
if (ebo->flag & BONE_ROOTSEL)
|
|
|
|
t->total++;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!t->total) return;
|
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
2005-10-29 20:08:25 +00:00
|
|
|
td = t->data = MEM_callocN(t->total*sizeof(TransData), "TransEditBone");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
for (ebo=G.edbo.first;ebo;ebo=ebo->next){
|
2005-08-20 09:55:11 +00:00
|
|
|
ebo->oldlength= ebo->length; // length==0.0 on extrude, used for scaling radius of bone points
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
if(ebo->layer & arm->layer) {
|
|
|
|
if (t->mode==TFM_BONE_ENVELOPE) {
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
if (ebo->flag & BONE_ROOTSEL){
|
|
|
|
td->val= &ebo->rad_head;
|
|
|
|
td->ival= *td->val;
|
|
|
|
|
|
|
|
VECCOPY (td->center, ebo->head);
|
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td->loc = NULL;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
|
|
|
|
td++;
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
}
|
2005-12-07 12:36:26 +00:00
|
|
|
if (ebo->flag & BONE_TIPSEL){
|
|
|
|
td->val= &ebo->rad_tail;
|
|
|
|
td->ival= *td->val;
|
|
|
|
VECCOPY (td->center, ebo->tail);
|
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td->loc = NULL;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
|
|
|
|
td++;
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
}
|
2005-07-16 19:07:02 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
}
|
|
|
|
else if (t->mode==TFM_BONESIZE) {
|
|
|
|
if (ebo->flag & BONE_SELECTED) {
|
|
|
|
if(arm->drawtype==ARM_ENVELOPE) {
|
|
|
|
td->loc= NULL;
|
|
|
|
td->val= &ebo->dist;
|
|
|
|
td->ival= ebo->dist;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
// abusive storage of scale in the loc pointer :)
|
|
|
|
td->loc= &ebo->xwidth;
|
|
|
|
VECCOPY (td->iloc, td->loc);
|
|
|
|
td->val= NULL;
|
|
|
|
}
|
|
|
|
VECCOPY (td->center, ebo->head);
|
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
|
|
|
|
/* use local bone matrix */
|
|
|
|
VecSubf(delta, ebo->tail, ebo->head);
|
|
|
|
vec_roll_to_mat3(delta, ebo->roll, bonemat);
|
|
|
|
Mat3MulMat3(td->mtx, mtx, bonemat);
|
|
|
|
Mat3Inv(td->smtx, td->mtx);
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->axismtx, td->mtx);
|
|
|
|
Mat3Ortho(td->axismtx);
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
|
|
|
|
td++;
|
|
|
|
}
|
2005-07-16 19:07:02 +00:00
|
|
|
}
|
2005-12-07 12:36:26 +00:00
|
|
|
else {
|
|
|
|
if (ebo->flag & BONE_TIPSEL){
|
|
|
|
VECCOPY (td->iloc, ebo->tail);
|
|
|
|
VECCOPY (td->center, td->iloc);
|
|
|
|
td->loc= ebo->tail;
|
|
|
|
td->flag= TD_SELECTED;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
td++;
|
|
|
|
}
|
|
|
|
if (ebo->flag & BONE_ROOTSEL){
|
|
|
|
VECCOPY (td->iloc, ebo->head);
|
|
|
|
VECCOPY (td->center, td->iloc);
|
|
|
|
td->loc= ebo->head;
|
|
|
|
td->flag= TD_SELECTED;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
2005-07-16 19:07:02 +00:00
|
|
|
|
2005-12-07 12:36:26 +00:00
|
|
|
td++;
|
|
|
|
}
|
2005-07-16 19:07:02 +00:00
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* meta elements ********* */
|
|
|
|
|
|
|
|
static void createTransMBallVerts(TransInfo *t)
|
|
|
|
{
|
|
|
|
MetaElem *ml;
|
|
|
|
TransData *td;
|
|
|
|
TransDataExtension *tx;
|
|
|
|
float mtx[3][3], smtx[3][3];
|
|
|
|
int count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
|
|
|
|
|
|
|
/* count totals */
|
|
|
|
for(ml= editelems.first; ml; ml= ml->next) {
|
|
|
|
if(ml->flag & SELECT) countsel++;
|
|
|
|
if(propmode) count++;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* note: in prop mode we need at least 1 selected */
|
|
|
|
if (countsel==0) return;
|
|
|
|
|
|
|
|
if(propmode) t->total = count;
|
|
|
|
else t->total = countsel;
|
|
|
|
|
2005-10-29 20:08:25 +00:00
|
|
|
td = t->data= MEM_callocN(t->total*sizeof(TransData), "TransObData(MBall EditMode)");
|
|
|
|
tx = t->ext = MEM_callocN(t->total*sizeof(TransDataExtension), "MetaElement_TransExtension");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
|
|
|
for(ml= editelems.first; ml; ml= ml->next) {
|
|
|
|
if(propmode || (ml->flag & SELECT)) {
|
|
|
|
td->loc= &ml->x;
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
|
2005-04-11 19:31:05 +00:00
|
|
|
if(ml->flag & SELECT) td->flag= TD_SELECTED | TD_USEQUAT | TD_SINGLESIZE;
|
2005-04-10 18:33:19 +00:00
|
|
|
else td->flag= TD_USEQUAT;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td->ext = tx;
|
|
|
|
td->tdi = NULL;
|
|
|
|
|
2005-04-11 19:31:05 +00:00
|
|
|
/* Radius of MetaElem (mass of MetaElem influence) */
|
|
|
|
if(ml->flag & MB_SCALE_RAD){
|
|
|
|
td->val = &ml->rad;
|
|
|
|
td->ival = ml->rad;
|
|
|
|
}
|
|
|
|
else{
|
|
|
|
td->val = &ml->s;
|
|
|
|
td->ival = ml->s;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* expx/expy/expz determine "shape" of some MetaElem types */
|
2005-04-10 18:33:19 +00:00
|
|
|
tx->size = &ml->expx;
|
|
|
|
tx->isize[0] = ml->expx;
|
|
|
|
tx->isize[1] = ml->expy;
|
|
|
|
tx->isize[2] = ml->expz;
|
|
|
|
|
2005-04-11 19:31:05 +00:00
|
|
|
/* quat is used for rotation of MetaElem */
|
2005-04-10 18:33:19 +00:00
|
|
|
tx->quat = ml->quat;
|
|
|
|
QUATCOPY(tx->iquat, ml->quat);
|
|
|
|
|
|
|
|
tx->rot = NULL;
|
|
|
|
|
|
|
|
td++;
|
|
|
|
tx++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* curve/surface ********* */
|
|
|
|
|
|
|
|
static void calc_distanceCurveVerts(TransData *head, TransData *tail) {
|
|
|
|
TransData *td, *td_near = NULL;
|
|
|
|
for (td = head; td<=tail; td++) {
|
|
|
|
if (td->flag & TD_SELECTED) {
|
|
|
|
td_near = td;
|
|
|
|
td->dist = 0.0f;
|
|
|
|
}
|
|
|
|
else if(td_near) {
|
|
|
|
float dist;
|
|
|
|
dist = VecLenf(td_near->center, td->center);
|
|
|
|
if (dist < (td-1)->dist) {
|
|
|
|
td->dist = (td-1)->dist;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
td->dist = dist;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
td->dist = MAXFLOAT;
|
|
|
|
td->flag |= TD_NOTCONNECTED;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
td_near = NULL;
|
|
|
|
for (td = tail; td>=head; td--) {
|
|
|
|
if (td->flag & TD_SELECTED) {
|
|
|
|
td_near = td;
|
|
|
|
td->dist = 0.0f;
|
|
|
|
}
|
|
|
|
else if(td_near) {
|
|
|
|
float dist;
|
|
|
|
dist = VecLenf(td_near->center, td->center);
|
|
|
|
if (td->flag & TD_NOTCONNECTED || dist < td->dist || (td+1)->dist < td->dist) {
|
|
|
|
td->flag &= ~TD_NOTCONNECTED;
|
|
|
|
if (dist < (td+1)->dist) {
|
|
|
|
td->dist = (td+1)->dist;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
td->dist = dist;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void createTransCurveVerts(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *td = NULL;
|
|
|
|
Nurb *nu;
|
|
|
|
BezTriple *bezt;
|
|
|
|
BPoint *bp;
|
|
|
|
float mtx[3][3], smtx[3][3];
|
|
|
|
int a;
|
|
|
|
int count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
|
|
|
|
|
|
|
/* count total of vertices, check identical as in 2nd loop for making transdata! */
|
|
|
|
for(nu= editNurb.first; nu; nu= nu->next) {
|
|
|
|
if((nu->type & 7)==CU_BEZIER) {
|
|
|
|
for(a=0, bezt= nu->bezt; a<nu->pntsu; a++, bezt++) {
|
|
|
|
if(bezt->hide==0) {
|
|
|
|
if(bezt->f1 & 1) countsel++;
|
|
|
|
if(bezt->f2 & 1) countsel++;
|
|
|
|
if(bezt->f3 & 1) countsel++;
|
|
|
|
if(propmode) count+= 3;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
for(a= nu->pntsu*nu->pntsv, bp= nu->bp; a>0; a--, bp++) {
|
|
|
|
if(bp->hide==0) {
|
|
|
|
if(propmode) count++;
|
|
|
|
if(bp->f1 & 1) countsel++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/* note: in prop mode we need at least 1 selected */
|
|
|
|
if (countsel==0) return;
|
|
|
|
|
|
|
|
if(propmode) t->total = count;
|
|
|
|
else t->total = countsel;
|
2005-10-29 20:08:25 +00:00
|
|
|
t->data= MEM_callocN(t->total*sizeof(TransData), "TransObData(Curve EditMode)");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
|
|
|
td = t->data;
|
|
|
|
for(nu= editNurb.first; nu; nu= nu->next) {
|
|
|
|
if((nu->type & 7)==CU_BEZIER) {
|
2005-06-11 02:01:26 +00:00
|
|
|
TransData *head, *tail;
|
|
|
|
head = tail = td;
|
2005-04-10 18:33:19 +00:00
|
|
|
for(a=0, bezt= nu->bezt; a<nu->pntsu; a++, bezt++) {
|
|
|
|
if(bezt->hide==0) {
|
|
|
|
if(propmode || (bezt->f1 & 1)) {
|
|
|
|
VECCOPY(td->iloc, bezt->vec[0]);
|
|
|
|
td->loc= bezt->vec[0];
|
|
|
|
VECCOPY(td->center, bezt->vec[1]);
|
|
|
|
if(bezt->f1 & 1) td->flag= TD_SELECTED;
|
|
|
|
else td->flag= 0;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td++;
|
|
|
|
count++;
|
2005-06-11 02:01:26 +00:00
|
|
|
tail++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
/* THIS IS THE CV, the other two are handles */
|
|
|
|
if(propmode || (bezt->f2 & 1)) {
|
|
|
|
VECCOPY(td->iloc, bezt->vec[1]);
|
|
|
|
td->loc= bezt->vec[1];
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
if(bezt->f2 & 1) td->flag= TD_SELECTED;
|
|
|
|
else td->flag= 0;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = &(bezt->alfa);
|
|
|
|
td->ival = bezt->alfa;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td++;
|
|
|
|
count++;
|
2005-06-11 02:01:26 +00:00
|
|
|
tail++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
if(propmode || (bezt->f3 & 1)) {
|
|
|
|
VECCOPY(td->iloc, bezt->vec[2]);
|
|
|
|
td->loc= bezt->vec[2];
|
|
|
|
VECCOPY(td->center, bezt->vec[1]);
|
|
|
|
if(bezt->f3 & 1) td->flag= TD_SELECTED;
|
|
|
|
else td->flag= 0;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td++;
|
|
|
|
count++;
|
2005-06-11 02:01:26 +00:00
|
|
|
tail++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
2005-06-11 02:01:26 +00:00
|
|
|
else if (propmode && head != tail) {
|
|
|
|
calc_distanceCurveVerts(head, tail-1);
|
|
|
|
head = tail;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
2005-06-11 02:01:26 +00:00
|
|
|
if (propmode && head != tail)
|
2005-07-08 06:55:10 +00:00
|
|
|
calc_distanceCurveVerts(head, tail-1);
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
else {
|
2005-06-11 02:01:26 +00:00
|
|
|
TransData *head, *tail;
|
|
|
|
head = tail = td;
|
2005-04-10 18:33:19 +00:00
|
|
|
for(a= nu->pntsu*nu->pntsv, bp= nu->bp; a>0; a--, bp++) {
|
|
|
|
if(bp->hide==0) {
|
|
|
|
if(propmode || (bp->f1 & 1)) {
|
|
|
|
VECCOPY(td->iloc, bp->vec);
|
|
|
|
td->loc= bp->vec;
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
if(bp->f1 & 1) td->flag= TD_SELECTED;
|
|
|
|
else td->flag= 0;
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = &(bp->alfa);
|
|
|
|
td->ival = bp->alfa;
|
|
|
|
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td++;
|
|
|
|
count++;
|
2005-06-11 02:01:26 +00:00
|
|
|
tail++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
2005-06-11 02:01:26 +00:00
|
|
|
else if (propmode && head != tail) {
|
|
|
|
calc_distanceCurveVerts(head, tail-1);
|
|
|
|
head = tail;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
2005-06-11 02:01:26 +00:00
|
|
|
if (propmode && head != tail)
|
|
|
|
calc_distanceCurveVerts(head, tail-1);
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* lattice *************** */
|
|
|
|
|
|
|
|
static void createTransLatticeVerts(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *td = NULL;
|
|
|
|
BPoint *bp;
|
|
|
|
float mtx[3][3], smtx[3][3];
|
|
|
|
int a;
|
|
|
|
int count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
|
|
|
|
|
|
|
bp= editLatt->def;
|
|
|
|
a= editLatt->pntsu*editLatt->pntsv*editLatt->pntsw;
|
|
|
|
while(a--) {
|
2005-11-21 11:15:46 +00:00
|
|
|
if(bp->hide==0) {
|
|
|
|
if(bp->f1 & 1) countsel++;
|
|
|
|
if(propmode) count++;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
bp++;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* note: in prop mode we need at least 1 selected */
|
|
|
|
if (countsel==0) return;
|
|
|
|
|
|
|
|
if(propmode) t->total = count;
|
|
|
|
else t->total = countsel;
|
2005-10-29 20:08:25 +00:00
|
|
|
t->data= MEM_callocN(t->total*sizeof(TransData), "TransObData(Lattice EditMode)");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
|
|
|
td = t->data;
|
|
|
|
bp= editLatt->def;
|
|
|
|
a= editLatt->pntsu*editLatt->pntsv*editLatt->pntsw;
|
|
|
|
while(a--) {
|
|
|
|
if(propmode || (bp->f1 & 1)) {
|
|
|
|
if(bp->hide==0) {
|
|
|
|
VECCOPY(td->iloc, bp->vec);
|
|
|
|
td->loc= bp->vec;
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
if(bp->f1 & 1) td->flag= TD_SELECTED;
|
|
|
|
else td->flag= 0;
|
|
|
|
Mat3CpyMat3(td->smtx, smtx);
|
|
|
|
Mat3CpyMat3(td->mtx, mtx);
|
|
|
|
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
|
|
|
|
|
|
|
td++;
|
|
|
|
count++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
bp++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ********************* mesh ****************** */
|
|
|
|
|
|
|
|
/* proportional distance based on connectivity */
|
2006-01-03 17:22:42 +00:00
|
|
|
#define E_VEC(a) (vectors + (3 * (a)->tmp.l))
|
|
|
|
#define E_NEAR(a) (nears[((a)->tmp.l)])
|
2005-04-10 18:33:19 +00:00
|
|
|
static void editmesh_set_connectivity_distance(int total, float *vectors, EditVert **nears)
|
|
|
|
{
|
|
|
|
EditMesh *em = G.editMesh;
|
|
|
|
EditVert *eve;
|
|
|
|
EditEdge *eed;
|
|
|
|
int i= 0, done= 1;
|
|
|
|
|
|
|
|
/* f2 flag is used for 'selection' */
|
2006-01-03 17:22:42 +00:00
|
|
|
/* tmp.l is offset on scratch array */
|
2005-04-10 18:33:19 +00:00
|
|
|
for(eve= em->verts.first; eve; eve= eve->next) {
|
|
|
|
if(eve->h==0) {
|
2006-01-03 17:22:42 +00:00
|
|
|
eve->tmp.l = i++;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
if(eve->f & SELECT) {
|
|
|
|
eve->f2= 2;
|
|
|
|
E_NEAR(eve) = eve;
|
|
|
|
E_VEC(eve)[0] = 0.0f;
|
|
|
|
E_VEC(eve)[1] = 0.0f;
|
|
|
|
E_VEC(eve)[2] = 0.0f;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
eve->f2 = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Floodfill routine */
|
|
|
|
/*
|
|
|
|
At worst this is n*n of complexity where n is number of edges
|
|
|
|
Best case would be n if the list is ordered perfectly.
|
|
|
|
Estimate is n log n in average (so not too bad)
|
|
|
|
*/
|
|
|
|
while(done) {
|
|
|
|
done= 0;
|
|
|
|
|
|
|
|
for(eed= em->edges.first; eed; eed= eed->next) {
|
|
|
|
if(eed->h==0) {
|
|
|
|
EditVert *v1= eed->v1, *v2= eed->v2;
|
|
|
|
float *vec2 = E_VEC(v2);
|
|
|
|
float *vec1 = E_VEC(v1);
|
|
|
|
|
|
|
|
if (v1->f2 + v2->f2 == 4)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (v1->f2) {
|
|
|
|
if (v2->f2) {
|
|
|
|
float nvec[3];
|
|
|
|
float len1 = VecLength(vec1);
|
|
|
|
float len2 = VecLength(vec2);
|
|
|
|
float lenn;
|
|
|
|
/* for v2 if not selected */
|
|
|
|
if (v2->f2 != 2) {
|
|
|
|
VecSubf(nvec, v2->co, E_NEAR(v1)->co);
|
|
|
|
lenn = VecLength(nvec);
|
|
|
|
if (lenn - len1 > 0.00001f && len2 - lenn > 0.00001f) {
|
|
|
|
VECCOPY(vec2, nvec);
|
|
|
|
E_NEAR(v2) = E_NEAR(v1);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
else if (len2 - len1 > 0.00001f && len1 - lenn > 0.00001f) {
|
|
|
|
VECCOPY(vec2, vec1);
|
|
|
|
E_NEAR(v2) = E_NEAR(v1);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/* for v1 if not selected */
|
|
|
|
if (v1->f2 != 2) {
|
|
|
|
VecSubf(nvec, v1->co, E_NEAR(v2)->co);
|
|
|
|
lenn = VecLength(nvec);
|
|
|
|
if (lenn - len2 > 0.00001f && len1 - lenn > 0.00001f) {
|
|
|
|
VECCOPY(vec1, nvec);
|
|
|
|
E_NEAR(v1) = E_NEAR(v2);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
else if (len1 - len2 > 0.00001f && len2 - lenn > 0.00001f) {
|
|
|
|
VECCOPY(vec1, vec2);
|
|
|
|
E_NEAR(v1) = E_NEAR(v2);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
v2->f2 = 1;
|
|
|
|
VecSubf(vec2, v2->co, E_NEAR(v1)->co);
|
|
|
|
if (VecLength(vec1) - VecLength(vec2) > 0.00001f) {
|
|
|
|
VECCOPY(vec2, vec1);
|
|
|
|
}
|
|
|
|
E_NEAR(v2) = E_NEAR(v1);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (v2->f2) {
|
|
|
|
v1->f2 = 1;
|
|
|
|
VecSubf(vec1, v1->co, E_NEAR(v2)->co);
|
|
|
|
if (VecLength(vec2) - VecLength(vec1) > 0.00001f) {
|
|
|
|
VECCOPY(vec1, vec2);
|
|
|
|
}
|
|
|
|
E_NEAR(v1) = E_NEAR(v2);
|
|
|
|
done = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void VertsToTransData(TransData *td, EditVert *eve)
|
|
|
|
{
|
|
|
|
td->flag = 0;
|
|
|
|
td->loc = eve->co;
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
|
|
|
|
// Setting normals
|
|
|
|
VECCOPY(td->axismtx[2], eve->no);
|
|
|
|
td->axismtx[0][0] =
|
|
|
|
td->axismtx[0][1] =
|
|
|
|
td->axismtx[0][2] =
|
|
|
|
td->axismtx[1][0] =
|
|
|
|
td->axismtx[1][1] =
|
|
|
|
td->axismtx[1][2] = 0.0f;
|
|
|
|
|
|
|
|
td->ext = NULL;
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
2005-12-01 19:04:57 +00:00
|
|
|
td->tdmir= NULL;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* *********************** CrazySpace correction. Now without doing subsurf optimal ****************** */
|
|
|
|
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
static void make_vertexcos__mapFunc(void *userData, int index, float *co, float *no_f, short *no_s)
|
|
|
|
{
|
|
|
|
float *vec = userData;
|
|
|
|
|
|
|
|
vec+= 3*index;
|
|
|
|
VECCOPY(vec, co);
|
|
|
|
}
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* hurmf, copy from buttons_editing.c, i have to sort this out what it means... */
|
|
|
|
static void modifiers_setOnCage(void *ob_v, void *md_v)
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
{
|
2006-01-04 15:37:48 +00:00
|
|
|
Object *ob = ob_v;
|
|
|
|
ModifierData *md;
|
|
|
|
|
|
|
|
int i, cageIndex = modifiers_getCageIndex(ob, NULL );
|
|
|
|
|
|
|
|
for( i = 0, md=ob->modifiers.first; md; ++i, md=md->next )
|
|
|
|
if( md == md_v ) {
|
|
|
|
if( i >= cageIndex )
|
|
|
|
md->mode ^= eModifierMode_OnCage;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* disable subsurf temporal, get mapped cos, and enable it */
|
|
|
|
static float *get_crazy_mapped_editverts(void)
|
|
|
|
{
|
|
|
|
DerivedMesh *dm;
|
|
|
|
ModifierData *md;
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
float *vertexcos;
|
2006-01-04 15:37:48 +00:00
|
|
|
int needsFree;
|
|
|
|
int i;
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
for( i = 0, md=G.obedit->modifiers.first; md; ++i, md=md->next ) {
|
|
|
|
if(md->type==eModifierType_Subsurf)
|
|
|
|
if(md->mode & eModifierMode_OnCage)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if(md) {
|
|
|
|
/* this call disables subsurf and enables the underlying modifier to deform, apparently */
|
|
|
|
modifiers_setOnCage(G.obedit, md);
|
|
|
|
/* make it all over */
|
|
|
|
makeDispListMesh(G.obedit);
|
|
|
|
}
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* now get the cage */
|
|
|
|
dm= editmesh_get_derived_cage(&needsFree);
|
|
|
|
|
|
|
|
vertexcos= MEM_mallocN(3*sizeof(float)*G.totvert, "vertexcos map");
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
dm->foreachMappedVert(dm, make_vertexcos__mapFunc, vertexcos);
|
|
|
|
|
|
|
|
if (needsFree) dm->release(dm);
|
2006-01-04 15:37:48 +00:00
|
|
|
|
|
|
|
if(md) {
|
|
|
|
/* set back the flag, no new cage needs to be built, transform does it */
|
|
|
|
modifiers_setOnCage(G.obedit, md);
|
|
|
|
}
|
|
|
|
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
return vertexcos;
|
|
|
|
}
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
#define TAN_MAKE_VEC(a, b, c) a[0]= b[0] + 0.2f*(b[0]-c[0]); a[1]= b[1] + 0.2f*(b[1]-c[1]); a[2]= b[2] + 0.2f*(b[2]-c[2])
|
|
|
|
static void set_crazy_vertex_quat(float *quat, float *v1, float *v2, float *v3, float *def1, float *def2, float *def3)
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
{
|
2006-01-04 15:37:48 +00:00
|
|
|
float vecu[3], vecv[3];
|
|
|
|
float q1[4], q2[4];
|
|
|
|
|
|
|
|
TAN_MAKE_VEC(vecu, v1, v2);
|
|
|
|
TAN_MAKE_VEC(vecv, v1, v3);
|
|
|
|
triatoquat(v1, vecu, vecv, q1);
|
|
|
|
|
|
|
|
TAN_MAKE_VEC(vecu, def1, def2);
|
|
|
|
TAN_MAKE_VEC(vecv, def1, def3);
|
|
|
|
triatoquat(def1, vecu, vecv, q2);
|
|
|
|
|
|
|
|
QuatSub(quat, q2, q1);
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
}
|
2006-01-04 15:37:48 +00:00
|
|
|
#undef TAN_MAKE_VEC
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
|
|
|
|
static void set_crazyspace_quats(float *mappedcos, float *quats)
|
|
|
|
{
|
|
|
|
EditMesh *em = G.editMesh;
|
|
|
|
EditVert *eve, *prev;
|
|
|
|
EditFace *efa;
|
2006-01-04 15:37:48 +00:00
|
|
|
float *v1, *v2, *v3, *v4;
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
int index= 0;
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* two abused locations in vertices */
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
for(eve= em->verts.first; eve; eve= eve->next, index++) {
|
2006-01-03 17:22:42 +00:00
|
|
|
eve->tmp.fp = NULL;
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
eve->prev= (EditVert *)index;
|
|
|
|
}
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* first store two sets of tangent vectors in vertices, we derive it just from the face-edges */
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
for(efa= em->faces.first; efa; efa= efa->next) {
|
|
|
|
|
2006-01-04 15:37:48 +00:00
|
|
|
/* retrieve mapped coordinates */
|
|
|
|
v1= mappedcos + 3*( (int)(efa->v1->prev) );
|
|
|
|
v2= mappedcos + 3*( (int)(efa->v2->prev) );
|
|
|
|
v3= mappedcos + 3*( (int)(efa->v3->prev) );
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v2->tmp.fp==NULL && efa->v2->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v2->co, efa->v3->co, efa->v1->co, v2, v3, v1);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v2->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
|
|
|
|
|
|
|
if(efa->v4) {
|
|
|
|
v4= mappedcos + 3*( (int)(efa->v4->prev) );
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v1->tmp.fp==NULL && efa->v1->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v1->co, efa->v2->co, efa->v4->co, v1, v2, v4);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v1->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v3->tmp.fp==NULL && efa->v3->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v3->co, efa->v4->co, efa->v2->co, v3, v4, v2);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v3->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v4->tmp.fp==NULL && efa->v4->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v4->co, efa->v1->co, efa->v3->co, v4, v1, v3);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v4->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v1->tmp.fp==NULL && efa->v1->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v1->co, efa->v2->co, efa->v3->co, v1, v2, v3);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v1->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
2006-01-04 15:47:16 +00:00
|
|
|
if(efa->v3->tmp.fp==NULL && efa->v3->f1) {
|
2006-01-04 15:37:48 +00:00
|
|
|
set_crazy_vertex_quat(quats, efa->v3->co, efa->v1->co, efa->v2->co, v3, v1, v2);
|
2006-01-04 15:47:16 +00:00
|
|
|
efa->v3->tmp.fp= quats;
|
2006-01-04 15:37:48 +00:00
|
|
|
quats+= 4;
|
|
|
|
}
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* restore abused prev pointer */
|
|
|
|
for(prev= NULL, eve= em->verts.first; eve; prev= eve, eve= eve->next)
|
|
|
|
eve->prev= prev;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
static void createTransEditVerts(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *tob = NULL;
|
|
|
|
EditMesh *em = G.editMesh;
|
|
|
|
EditVert *eve;
|
|
|
|
EditVert **nears = NULL;
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
float *vectors = NULL, *mappedcos = NULL, *quats= NULL;
|
2005-04-10 18:33:19 +00:00
|
|
|
float mtx[3][3], smtx[3][3];
|
|
|
|
int count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
2005-12-01 19:04:57 +00:00
|
|
|
int mirror= (G.scene->toolsettings->editbutflag & B_MESH_X_MIRROR);
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
// transform now requires awareness for select mode, so we tag the f1 flags in verts
|
|
|
|
if(G.scene->selectmode & SCE_SELECT_VERTEX) {
|
|
|
|
for(eve= em->verts.first; eve; eve= eve->next) {
|
|
|
|
if(eve->h==0 && (eve->f & SELECT))
|
|
|
|
eve->f1= SELECT;
|
|
|
|
else
|
|
|
|
eve->f1= 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if(G.scene->selectmode & SCE_SELECT_EDGE) {
|
|
|
|
EditEdge *eed;
|
|
|
|
for(eve= em->verts.first; eve; eve= eve->next) eve->f1= 0;
|
|
|
|
for(eed= em->edges.first; eed; eed= eed->next) {
|
|
|
|
if(eed->h==0 && (eed->f & SELECT))
|
|
|
|
eed->v1->f1= eed->v2->f1= SELECT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
EditFace *efa;
|
|
|
|
for(eve= em->verts.first; eve; eve= eve->next) eve->f1= 0;
|
|
|
|
for(efa= em->faces.first; efa; efa= efa->next) {
|
|
|
|
if(efa->h==0 && (efa->f & SELECT)) {
|
|
|
|
efa->v1->f1= efa->v2->f1= efa->v3->f1= SELECT;
|
|
|
|
if(efa->v4) efa->v4->f1= SELECT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* now we can count */
|
|
|
|
for(eve= em->verts.first; eve; eve= eve->next) {
|
|
|
|
if(eve->h==0) {
|
|
|
|
if(eve->f1) countsel++;
|
|
|
|
if(propmode) count++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* note: in prop mode we need at least 1 selected */
|
|
|
|
if (countsel==0) return;
|
|
|
|
|
|
|
|
if(propmode) {
|
|
|
|
t->total = count;
|
|
|
|
|
|
|
|
/* allocating scratch arrays */
|
2005-04-17 17:43:07 +00:00
|
|
|
vectors = (float *)MEM_mallocN(t->total * 3 * sizeof(float), "scratch vectors");
|
|
|
|
nears = (EditVert**)MEM_mallocN(t->total * sizeof(EditVert*), "scratch nears");
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
else t->total = countsel;
|
2005-10-29 20:08:25 +00:00
|
|
|
tob= t->data= MEM_callocN(t->total*sizeof(TransData), "TransObData(Mesh EditMode)");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
Mat3CpyMat4(mtx, G.obedit->obmat);
|
|
|
|
Mat3Inv(smtx, mtx);
|
|
|
|
|
|
|
|
if(propmode) editmesh_set_connectivity_distance(t->total, vectors, nears);
|
|
|
|
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
/* detect CrazySpace [tm] */
|
2005-10-26 10:49:19 +00:00
|
|
|
if(propmode==0) {
|
|
|
|
if(modifiers_getCageIndex(G.obedit, NULL)>=0) {
|
|
|
|
if(modifiers_isDeformed(G.obedit)) {
|
2006-01-04 15:37:48 +00:00
|
|
|
/* disable subsurf temporal, get mapped cos, and enable it */
|
|
|
|
mappedcos= get_crazy_mapped_editverts();
|
|
|
|
quats= MEM_mallocN( (t->total)*sizeof(float)*4, "crazy quats");
|
2005-10-26 10:49:19 +00:00
|
|
|
set_crazyspace_quats(mappedcos, quats);
|
|
|
|
}
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-12-01 19:04:57 +00:00
|
|
|
/* find out which half we do */
|
|
|
|
if(mirror) {
|
|
|
|
for (eve=em->verts.first; eve; eve=eve->next) {
|
|
|
|
if(eve->h==0 && eve->f1 && eve->co[0]!=0.0f) {
|
|
|
|
if(eve->co[0]<0.0f)
|
|
|
|
mirror = -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
for (eve=em->verts.first; eve; eve=eve->next) {
|
|
|
|
if(eve->h==0) {
|
|
|
|
if(propmode || eve->f1) {
|
|
|
|
VertsToTransData(tob, eve);
|
|
|
|
|
|
|
|
if(eve->f1) tob->flag |= TD_SELECTED;
|
|
|
|
if(propmode) {
|
|
|
|
if (eve->f2) {
|
|
|
|
float vec[3];
|
|
|
|
VECCOPY(vec, E_VEC(eve));
|
|
|
|
Mat3MulVecfl(mtx, vec);
|
|
|
|
tob->dist= VecLength(vec);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
tob->flag |= TD_NOTCONNECTED;
|
|
|
|
tob->dist = MAXFLOAT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
/* CrazySpace */
|
2006-01-03 17:22:42 +00:00
|
|
|
if(quats && eve->tmp.fp) {
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
float mat[3][3], imat[3][3], qmat[3][3];
|
|
|
|
|
2006-01-03 17:22:42 +00:00
|
|
|
QuatToMat3(eve->tmp.fp, qmat);
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
Mat3MulMat3(mat, mtx, qmat);
|
|
|
|
Mat3Inv(imat, mat);
|
|
|
|
|
|
|
|
Mat3CpyMat3(tob->smtx, imat);
|
|
|
|
Mat3CpyMat3(tob->mtx, mat);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
Mat3CpyMat3(tob->smtx, smtx);
|
|
|
|
Mat3CpyMat3(tob->mtx, mtx);
|
|
|
|
}
|
2005-12-01 19:04:57 +00:00
|
|
|
|
|
|
|
/* Mirror? */
|
|
|
|
if( (mirror>0 && tob->iloc[0]>0.0f) || (mirror<0 && tob->iloc[0]<0.0f)) {
|
|
|
|
EditVert *vmir= editmesh_get_x_mirror_vert(G.obedit, tob->iloc); /* initializes octree on first call */
|
|
|
|
if(vmir!=eve) tob->tdmir= vmir;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
tob++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (propmode) {
|
2005-04-17 17:43:07 +00:00
|
|
|
MEM_freeN(vectors);
|
|
|
|
MEM_freeN(nears);
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
New: CrazySpace [tm] correction
When Modifiers are used in Edit Mode to show the deformed result for
editing, all actual coordinates Blender works with are still the ones from
the original Cage. You can notice that with the Transform Widget or
helper lines while transforming.
Even worse, the actual transformations still happened on the original Cage
as well, making it very hard to edit. That caused the feature to be named
"CrazySpace" (baptized by Andy, afaik?).
This commit calculates the deformation transformation per vertex, and
inverse corrects it, so it's more intuitive editing this way.
Unfortunately all the deformation features of Blender don't use matrices
for defining deform, so the existing code cannot be re-used to retrieve
the correct deformation matrix per vertex. The solution I found is based
on calculating per face the transformation based on its first 3 vertices,
and store this transformation averaged in the face's vertices.
The solution can also only work on entire faces, because the full deform
can only be retrieved using 3 vertices. (using 2 vertices will miss edge-
aligned rotation, using 1 vertex can only retrieve translation).
By deriving the deformations per face, small errors will still happen,
especially on very low-poly Meshes with extreme deformations.
The only alternative I know now, is providing each vertex in
a mesh with 2 extreme small tangent vectors, which get deformed using the
existing code as well. That will mess up the existing deformation code too
much though, this solution has the benefit it works with each deform we can
up with later too.
Last note about CrazySpace: it can only be used to tweak Meshes. Do not
even try to add vertices, extrude, or duplicate. Probably we should disable
this... but preventing user errors isn't always power-user-friendly, eh. :)
2005-10-26 09:56:52 +00:00
|
|
|
/* crazy space free */
|
|
|
|
if(mappedcos)
|
|
|
|
MEM_freeN(mappedcos);
|
|
|
|
if(quats)
|
|
|
|
MEM_freeN(quats);
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|
2005-08-23 18:13:30 +00:00
|
|
|
/* ********************* UV ****************** */
|
|
|
|
|
|
|
|
static void UVsToTransData(TransData *td, TransData2D *td2d, float *uv, int selected)
|
|
|
|
{
|
|
|
|
float aspx, aspy;
|
|
|
|
|
|
|
|
transform_aspect_ratio_tface_uv(&aspx, &aspy);
|
|
|
|
|
|
|
|
/* uv coords are scaled by aspects. this is needed for rotations and
|
|
|
|
proportional editing to be consistent with the stretchted uv coords
|
|
|
|
that are displayed. this also means that for display and numinput,
|
|
|
|
and when the the uv coords are flushed, these are converted each time */
|
|
|
|
td2d->loc[0] = uv[0]*aspx;
|
|
|
|
td2d->loc[1] = uv[1]*aspy;
|
|
|
|
td2d->loc[2] = 0.0f;
|
|
|
|
td2d->loc2d = uv;
|
|
|
|
|
|
|
|
td->flag = 0;
|
|
|
|
td->loc = td2d->loc;
|
|
|
|
VECCOPY(td->center, td->loc);
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
|
|
|
|
memset(td->axismtx, 0, sizeof(td->axismtx));
|
|
|
|
td->axismtx[2][2] = 1.0f;
|
|
|
|
|
|
|
|
td->ext= NULL; td->tdi= NULL; td->val= NULL;
|
|
|
|
|
|
|
|
if(selected) {
|
|
|
|
td->flag |= TD_SELECTED;
|
|
|
|
td->dist= 0.0;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
td->dist= MAXFLOAT;
|
|
|
|
|
|
|
|
Mat3One(td->mtx);
|
|
|
|
Mat3One(td->smtx);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void createTransUVs(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *td = NULL;
|
|
|
|
TransData2D *td2d = NULL;
|
|
|
|
Mesh *me;
|
|
|
|
TFace *tf;
|
|
|
|
MFace *mf;
|
|
|
|
int a, count=0, countsel=0;
|
|
|
|
int propmode = t->flag & T_PROP_EDIT;
|
|
|
|
|
|
|
|
if(is_uv_tface_editing_allowed()==0) return;
|
|
|
|
me= get_mesh(OBACT);
|
|
|
|
|
|
|
|
/* count */
|
|
|
|
tf= me->tface;
|
|
|
|
mf= me->mface;
|
|
|
|
for(a=me->totface; a>0; a--, tf++, mf++) {
|
|
|
|
if(mf->v3 && tf->flag & TF_SELECT) {
|
|
|
|
if(tf->flag & TF_SEL1) countsel++;
|
|
|
|
if(tf->flag & TF_SEL2) countsel++;
|
|
|
|
if(tf->flag & TF_SEL3) countsel++;
|
|
|
|
if(mf->v4 && (tf->flag & TF_SEL4)) countsel++;
|
|
|
|
if(propmode)
|
|
|
|
count += (mf->v4)? 4: 3;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* note: in prop mode we need at least 1 selected */
|
|
|
|
if (countsel==0) return;
|
|
|
|
|
|
|
|
t->total= (propmode)? count: countsel;
|
2005-10-29 20:08:25 +00:00
|
|
|
t->data= MEM_callocN(t->total*sizeof(TransData), "TransObData(UV Editing)");
|
2005-08-23 18:13:30 +00:00
|
|
|
/* for each 2d uv coord a 3d vector is allocated, so that they can be
|
|
|
|
treated just as if they were 3d verts */
|
|
|
|
t->data2d= MEM_mallocN(t->total*sizeof(TransData2D), "TransObData2D(UV Editing)");
|
|
|
|
|
|
|
|
if(G.sima->flag & SI_CLIP_UV)
|
|
|
|
t->flag |= T_CLIP_UV;
|
|
|
|
|
|
|
|
td= t->data;
|
|
|
|
td2d= t->data2d;
|
|
|
|
tf= me->tface;
|
|
|
|
mf= me->mface;
|
|
|
|
for(a=me->totface; a>0; a--, tf++, mf++) {
|
|
|
|
if(mf->v3 && tf->flag & TF_SELECT) {
|
|
|
|
if(tf->flag & TF_SEL1 || propmode)
|
|
|
|
UVsToTransData(td++, td2d++, tf->uv[0], (tf->flag & TF_SEL1));
|
|
|
|
if(tf->flag & TF_SEL2 || propmode)
|
|
|
|
UVsToTransData(td++, td2d++, tf->uv[1], (tf->flag & TF_SEL2));
|
|
|
|
if(tf->flag & TF_SEL3 || propmode)
|
|
|
|
UVsToTransData(td++, td2d++, tf->uv[2], (tf->flag & TF_SEL3));
|
|
|
|
|
|
|
|
if(mf->v4 && (tf->flag & TF_SEL4 || propmode))
|
|
|
|
UVsToTransData(td++, td2d++, tf->uv[3], (tf->flag & TF_SEL4));
|
|
|
|
}
|
|
|
|
}
|
2005-12-03 23:22:31 +00:00
|
|
|
|
|
|
|
if (G.sima->flag & SI_LSCM_LIVE)
|
|
|
|
unwrap_lscm_live_begin();
|
2005-08-23 18:13:30 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void flushTransUVs(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData2D *td;
|
|
|
|
int a, width, height;
|
|
|
|
Object *ob= OBACT;
|
|
|
|
Mesh *me= get_mesh(ob);
|
|
|
|
float aspx, aspy, invx, invy;
|
|
|
|
|
|
|
|
transform_aspect_ratio_tface_uv(&aspx, &aspy);
|
|
|
|
transform_width_height_tface_uv(&width, &height);
|
|
|
|
invx= 1.0f/aspx;
|
|
|
|
invy= 1.0f/aspy;
|
|
|
|
|
|
|
|
/* flush to 2d vector from internally used 3d vector */
|
|
|
|
for(a=0, td= t->data2d; a<t->total; a++, td++) {
|
|
|
|
td->loc2d[0]= td->loc[0]*invx;
|
|
|
|
td->loc2d[1]= td->loc[1]*invy;
|
|
|
|
|
|
|
|
if((G.sima->flag & SI_PIXELSNAP) && (t->state != TRANS_CANCEL)) {
|
|
|
|
td->loc2d[0]= floor(width*td->loc2d[0])/width;
|
|
|
|
td->loc2d[1]= floor(height*td->loc2d[1])/height;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if((G.sima->flag & SI_BE_SQUARE) && (t->state != TRANS_CANCEL))
|
|
|
|
be_square_tface_uv(me);
|
|
|
|
|
|
|
|
/* this is overkill if G.sima->lock is not set, but still needed */
|
|
|
|
object_uvs_changed(ob);
|
|
|
|
}
|
|
|
|
|
|
|
|
int clipUVTransform(TransInfo *t, float *vec, int resize)
|
|
|
|
{
|
|
|
|
TransData *td;
|
|
|
|
int a, clipx=1, clipy=1;
|
|
|
|
float aspx, aspy, min[2], max[2];
|
|
|
|
|
|
|
|
transform_aspect_ratio_tface_uv(&aspx, &aspy);
|
|
|
|
min[0]= min[1]= 0.0f;
|
|
|
|
max[0]= aspx; max[1]= aspy;
|
|
|
|
|
|
|
|
for(a=0, td= t->data; a<t->total; a++, td++) {
|
|
|
|
DO_MINMAX2(td->loc, min, max);
|
|
|
|
}
|
|
|
|
|
|
|
|
if(resize) {
|
|
|
|
if(min[0] < 0.0f && t->center[0] > 0.0f && t->center[0] < aspx*0.5f)
|
|
|
|
vec[0] *= t->center[0]/(t->center[0] - min[0]);
|
|
|
|
else if(max[0] > aspx && t->center[0] < aspx)
|
|
|
|
vec[0] *= (t->center[0] - aspx)/(t->center[0] - max[0]);
|
|
|
|
else
|
|
|
|
clipx= 0;
|
|
|
|
|
|
|
|
if(min[1] < 0.0f && t->center[1] > 0.0f && t->center[1] < aspy*0.5f)
|
|
|
|
vec[1] *= t->center[1]/(t->center[1] - min[1]);
|
|
|
|
else if(max[1] > aspy && t->center[1] < aspy)
|
|
|
|
vec[1] *= (t->center[1] - aspy)/(t->center[1] - max[1]);
|
|
|
|
else
|
|
|
|
clipy= 0;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
if(min[0] < 0.0f)
|
|
|
|
vec[0] -= min[0];
|
|
|
|
else if(max[0] > aspx)
|
|
|
|
vec[0] -= max[0]-aspx;
|
|
|
|
else
|
|
|
|
clipx= 0;
|
|
|
|
|
|
|
|
if(min[1] < 0.0f)
|
|
|
|
vec[1] -= min[1];
|
|
|
|
else if(max[1] > aspy)
|
|
|
|
vec[1] -= max[1]-aspy;
|
|
|
|
else
|
|
|
|
clipy= 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (clipx || clipy);
|
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
/* **************** IpoKey stuff, for Object TransData ********** */
|
|
|
|
|
|
|
|
/* storage of bezier triple. thats why -3 and +3! */
|
|
|
|
static void set_tdi_old(float *old, float *poin)
|
|
|
|
{
|
|
|
|
old[0]= *(poin);
|
|
|
|
old[3]= *(poin-3);
|
|
|
|
old[6]= *(poin+3);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* while transforming */
|
|
|
|
void add_tdi_poin(float *poin, float *old, float delta)
|
|
|
|
{
|
|
|
|
if(poin) {
|
|
|
|
poin[0]= old[0]+delta;
|
|
|
|
poin[-3]= old[3]+delta;
|
|
|
|
poin[3]= old[6]+delta;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* fill ipokey transdata with old vals and pointers */
|
|
|
|
static void ipokey_to_transdata(IpoKey *ik, TransData *td)
|
|
|
|
{
|
|
|
|
extern int ob_ar[]; // blenkernel ipo.c
|
|
|
|
TransDataIpokey *tdi= td->tdi;
|
|
|
|
BezTriple *bezt;
|
|
|
|
int a, delta= 0;
|
|
|
|
|
2005-05-08 20:44:59 +00:00
|
|
|
td->val= NULL; // is read on ESC
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
for(a=0; a<OB_TOTIPO; a++) {
|
|
|
|
if(ik->data[a]) {
|
|
|
|
bezt= ik->data[a];
|
|
|
|
|
|
|
|
switch( ob_ar[a] ) {
|
|
|
|
case OB_LOC_X:
|
|
|
|
case OB_DLOC_X:
|
|
|
|
tdi->locx= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_LOC_Y:
|
|
|
|
case OB_DLOC_Y:
|
|
|
|
tdi->locy= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_LOC_Z:
|
|
|
|
case OB_DLOC_Z:
|
|
|
|
tdi->locz= &(bezt->vec[1][1]); break;
|
|
|
|
|
|
|
|
case OB_DROT_X:
|
|
|
|
delta= 1;
|
|
|
|
case OB_ROT_X:
|
|
|
|
tdi->rotx= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_DROT_Y:
|
|
|
|
delta= 1;
|
|
|
|
case OB_ROT_Y:
|
|
|
|
tdi->roty= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_DROT_Z:
|
|
|
|
delta= 1;
|
|
|
|
case OB_ROT_Z:
|
|
|
|
tdi->rotz= &(bezt->vec[1][1]); break;
|
|
|
|
|
|
|
|
case OB_SIZE_X:
|
|
|
|
case OB_DSIZE_X:
|
|
|
|
tdi->sizex= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_SIZE_Y:
|
|
|
|
case OB_DSIZE_Y:
|
|
|
|
tdi->sizey= &(bezt->vec[1][1]); break;
|
|
|
|
case OB_SIZE_Z:
|
|
|
|
case OB_DSIZE_Z:
|
|
|
|
tdi->sizez= &(bezt->vec[1][1]); break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* oldvals for e.g. undo */
|
|
|
|
if(tdi->locx) set_tdi_old(tdi->oldloc, tdi->locx);
|
|
|
|
if(tdi->locy) set_tdi_old(tdi->oldloc+1, tdi->locy);
|
|
|
|
if(tdi->locz) set_tdi_old(tdi->oldloc+2, tdi->locz);
|
|
|
|
|
|
|
|
/* remember, for mapping curves ('1'=10 degrees) */
|
|
|
|
if(tdi->rotx) set_tdi_old(tdi->oldrot, tdi->rotx);
|
|
|
|
if(tdi->roty) set_tdi_old(tdi->oldrot+1, tdi->roty);
|
|
|
|
if(tdi->rotz) set_tdi_old(tdi->oldrot+2, tdi->rotz);
|
|
|
|
|
|
|
|
/* this is not allowed to be dsize! */
|
|
|
|
if(tdi->sizex) set_tdi_old(tdi->oldsize, tdi->sizex);
|
|
|
|
if(tdi->sizey) set_tdi_old(tdi->oldsize+1, tdi->sizey);
|
|
|
|
if(tdi->sizez) set_tdi_old(tdi->oldsize+2, tdi->sizez);
|
|
|
|
|
|
|
|
tdi->flag= TOB_IPO;
|
|
|
|
if(delta) tdi->flag |= TOB_IPODROT;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* *************************** Object Transform data ******************* */
|
|
|
|
|
|
|
|
static void ObjectToTransData(TransData *td, Object *ob)
|
|
|
|
{
|
|
|
|
float obmtx[3][3];
|
|
|
|
Object *tr;
|
|
|
|
void *cfirst, *clast;
|
|
|
|
|
2005-06-06 12:23:47 +00:00
|
|
|
/* set axismtx BEFORE clearing constraints to have the real orientation */
|
|
|
|
Mat3CpyMat4(td->axismtx, ob->obmat);
|
|
|
|
Mat3Ortho(td->axismtx);
|
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* then why are constraints and track disabled here?
|
|
|
|
they dont alter loc/rot/size itself (ton) */
|
2005-04-10 18:33:19 +00:00
|
|
|
cfirst = ob->constraints.first;
|
|
|
|
clast = ob->constraints.last;
|
|
|
|
ob->constraints.first=ob->constraints.last=NULL;
|
|
|
|
|
|
|
|
tr= ob->track;
|
|
|
|
ob->track= NULL;
|
|
|
|
|
|
|
|
where_is_object(ob);
|
|
|
|
|
|
|
|
ob->track= tr;
|
|
|
|
|
|
|
|
ob->constraints.first = cfirst;
|
|
|
|
ob->constraints.last = clast;
|
|
|
|
|
|
|
|
td->ob = ob;
|
|
|
|
|
|
|
|
td->loc = ob->loc;
|
|
|
|
VECCOPY(td->iloc, td->loc);
|
|
|
|
|
|
|
|
td->ext->rot = ob->rot;
|
|
|
|
VECCOPY(td->ext->irot, ob->rot);
|
|
|
|
VECCOPY(td->ext->drot, ob->drot);
|
|
|
|
|
|
|
|
td->ext->size = ob->size;
|
|
|
|
VECCOPY(td->ext->isize, ob->size);
|
|
|
|
VECCOPY(td->ext->dsize, ob->dsize);
|
|
|
|
|
|
|
|
VECCOPY(td->center, ob->obmat[3]);
|
|
|
|
|
|
|
|
if (ob->parent)
|
|
|
|
{
|
|
|
|
float totmat[3][3], obinv[3][3];
|
|
|
|
|
|
|
|
/* we calculate smtx without obmat: so a parmat */
|
|
|
|
object_to_mat3(ob, obmtx);
|
|
|
|
Mat3CpyMat4(totmat, ob->obmat);
|
|
|
|
Mat3Inv(obinv, totmat);
|
|
|
|
Mat3MulMat3(td->smtx, obmtx, obinv);
|
|
|
|
Mat3Inv(td->mtx, td->smtx);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
Mat3One(td->smtx);
|
|
|
|
Mat3One(td->mtx);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* sets flags in Bases to define whether they take part in transform */
|
|
|
|
/* it deselects Bases, so we have to call the clear function always after */
|
|
|
|
static void set_trans_object_base_flags(TransInfo *t)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
if Base selected and has parent selected:
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
base->flag= BA_WAS_SEL
|
2005-04-10 18:33:19 +00:00
|
|
|
*/
|
|
|
|
Base *base;
|
|
|
|
|
|
|
|
/* makes sure base flags and object flags are identical */
|
|
|
|
copy_baseflags();
|
|
|
|
|
2005-09-26 15:34:21 +00:00
|
|
|
/* handle pending update events, otherwise they got copied below */
|
|
|
|
for (base= FIRSTBASE; base; base= base->next) {
|
|
|
|
if(base->object->recalc)
|
|
|
|
object_handle_update(base->object);
|
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
for (base= FIRSTBASE; base; base= base->next) {
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
base->flag &= ~BA_WAS_SEL;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
if(TESTBASELIB(base)) {
|
2005-04-10 18:33:19 +00:00
|
|
|
Object *ob= base->object;
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
Object *parsel= ob->parent;
|
2005-04-10 18:33:19 +00:00
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* if parent selected, deselect */
|
|
|
|
while(parsel) {
|
|
|
|
if(parsel->flag & SELECT) break;
|
|
|
|
parsel= parsel->parent;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
if(parsel) {
|
|
|
|
base->flag &= ~SELECT;
|
|
|
|
base->flag |= BA_WAS_SEL;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* used for flush, depgraph will change recalcs if needed :) */
|
|
|
|
ob->recalc |= OB_RECALC_OB;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* all recalc flags get flushed */
|
2005-10-03 09:04:57 +00:00
|
|
|
DAG_scene_flush_update(G.scene, screen_view3d_layers());
|
2005-04-10 18:33:19 +00:00
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* and we store them temporal in base (only used for transform code) */
|
|
|
|
/* this because after doing updates, the object->recalc is cleared */
|
|
|
|
for (base= FIRSTBASE; base; base= base->next) {
|
|
|
|
if(base->object->recalc & OB_RECALC_OB)
|
|
|
|
base->flag |= BA_HAS_RECALC_OB;
|
|
|
|
if(base->object->recalc & OB_RECALC_DATA)
|
|
|
|
base->flag |= BA_HAS_RECALC_DATA;
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|
2005-07-19 11:21:01 +00:00
|
|
|
static void clear_trans_object_base_flags(void)
|
2005-04-10 18:33:19 +00:00
|
|
|
{
|
|
|
|
Base *base;
|
|
|
|
|
|
|
|
base= FIRSTBASE;
|
|
|
|
while(base) {
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
if(base->flag & BA_WAS_SEL) base->flag |= SELECT;
|
|
|
|
base->flag &= ~(BA_WAS_SEL|BA_HAS_RECALC_OB|BA_HAS_RECALC_DATA|BA_DO_IPO);
|
2005-04-10 18:33:19 +00:00
|
|
|
|
2005-07-04 05:16:54 +00:00
|
|
|
base = base->next;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
/* inserting keys, refresh ipo-keys, softbody, redraw events... (ton) */
|
2005-08-16 10:17:02 +00:00
|
|
|
/* note; transdata has been freed already! */
|
|
|
|
void special_aftertrans_update(TransInfo *t)
|
2005-05-04 12:53:35 +00:00
|
|
|
{
|
2005-08-16 10:17:02 +00:00
|
|
|
Object *ob;
|
2005-05-04 12:53:35 +00:00
|
|
|
Base *base;
|
2005-12-18 20:14:22 +00:00
|
|
|
IpoCurve *icu;
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
int redrawipo=0;
|
2005-08-16 10:17:02 +00:00
|
|
|
int cancelled= (t->state == TRANS_CANCEL);
|
|
|
|
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
if(G.obedit) {
|
|
|
|
if(t->mode==TFM_BONESIZE || t->mode==TFM_BONE_ENVELOPE)
|
|
|
|
allqueue(REDRAWBUTSEDIT, 0);
|
2005-12-01 19:04:57 +00:00
|
|
|
|
|
|
|
/* table needs to be created for each edit command, since vertices can move etc */
|
|
|
|
mesh_octree_table(G.obedit, NULL, 'e');
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
}
|
2005-08-16 10:17:02 +00:00
|
|
|
else if( (t->flag & T_POSE) && t->poseobj) {
|
|
|
|
bArmature *arm;
|
2005-05-04 12:53:35 +00:00
|
|
|
bAction *act;
|
|
|
|
bPose *pose;
|
|
|
|
bPoseChannel *pchan;
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
|
2005-08-16 10:17:02 +00:00
|
|
|
ob= t->poseobj;
|
|
|
|
arm= ob->data;
|
2005-10-30 13:50:42 +00:00
|
|
|
pose= ob->pose;
|
2005-08-16 10:17:02 +00:00
|
|
|
|
2005-10-22 18:47:38 +00:00
|
|
|
/* this signal does one recalc on pose, then unlocks, so ESC or edit will work */
|
2005-10-30 13:50:42 +00:00
|
|
|
pose->flag |= POSE_DO_UNLOCK;
|
2005-11-01 20:58:37 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* if target-less IK grabbing, we calculate the pchan transforms and clear flag */
|
|
|
|
if(!cancelled && t->mode==TFM_TRANSLATION)
|
|
|
|
apply_targetless_ik(ob);
|
2005-11-15 15:54:57 +00:00
|
|
|
else {
|
|
|
|
/* not forget to clear the auto flag */
|
|
|
|
for (pchan=ob->pose->chanbase.first; pchan; pchan=pchan->next) {
|
|
|
|
bKinematicConstraint *data= has_targetless_ik(pchan);
|
|
|
|
if(data) data->flag &= ~CONSTRAINT_IK_AUTO;
|
|
|
|
}
|
|
|
|
}
|
2005-10-30 13:50:42 +00:00
|
|
|
|
|
|
|
if(t->mode==TFM_TRANSLATION)
|
|
|
|
pose_grab_with_ik_clear(ob);
|
2005-10-03 14:40:42 +00:00
|
|
|
|
2005-10-30 13:50:42 +00:00
|
|
|
/* automatic inserting of keys */
|
Orange's buglist!
- Action Editor: hotkeys V and H for handles were invisible, added menus
- NLA strips: when current frame is exactly on the strip end, it didn't
include that action... needs a rounding correction for it.
- Action/NLA: deleting keys in Action, which results in only 1 key left,
resulted in zero sized strip length. Now the strips are defaulted to be
1 frame in size minimal.
- NLA editor: ALT+C "Convert to strip" didn't increment Action user count
- 3D Window: CTRL+P make parent to Bone still gave the insane menu with all
bone names. With unified PoseMode select it can just parent to the
active Bone. Note; this now requires the Armature to be in PoseMode to
work.
- Rotation Constraint; the new options to only map to X,Y,Z rotation, did
set the not mapped rotation axes to zero. These should remain unchanged.
- AutoKey optionn for Actions; should not insert action keys on ESC
And added a fix myself:
- When SHIFT+selecting a Bone in PoseMode, and the Armature was not selected
or active yet, it doesn't extend-select/deselect the Bone anymore.
This case is only useful when you try to add IK or Constraint, so the
shift+selection should only activate the clicked Bone.
2005-10-29 10:15:36 +00:00
|
|
|
if((G.flags & G_RECORDKEYS) && (!cancelled)) {
|
2005-07-23 18:52:31 +00:00
|
|
|
act= ob->action;
|
2005-05-04 12:53:35 +00:00
|
|
|
|
|
|
|
if (!act)
|
2005-10-10 17:42:48 +00:00
|
|
|
act= ob->action= add_empty_action(ID_PO);
|
2005-05-04 12:53:35 +00:00
|
|
|
|
|
|
|
for (pchan=pose->chanbase.first; pchan; pchan=pchan->next){
|
2005-11-17 22:07:06 +00:00
|
|
|
if (pchan->bone->flag & BONE_TRANSFORM){
|
2005-12-18 20:14:22 +00:00
|
|
|
|
|
|
|
if(U.uiflag & USER_KEYINSERTAVAI) {
|
|
|
|
bActionChannel *achan;
|
|
|
|
|
|
|
|
for (achan = act->chanbase.first; achan; achan=achan->next){
|
|
|
|
|
|
|
|
if (achan->ipo && !strcmp (achan->name, pchan->name)){
|
|
|
|
for (icu = achan->ipo->curve.first; icu; icu=icu->next){
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, icu->adrcode);
|
|
|
|
}
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else{
|
|
|
|
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_SIZE_X);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_SIZE_Y);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_SIZE_Z);
|
|
|
|
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_QUAT_W);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_QUAT_X);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_QUAT_Y);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_QUAT_Z);
|
|
|
|
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_LOC_X);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_LOC_Y);
|
|
|
|
insertkey(&ob->id, ID_PO, pchan->name, NULL, AC_LOC_Z);
|
|
|
|
}
|
2005-05-04 12:53:35 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
remake_action_ipos (act);
|
|
|
|
allspace(REMAKEIPO, 0);
|
|
|
|
allqueue(REDRAWACTION, 0);
|
|
|
|
allqueue(REDRAWIPO, 0);
|
|
|
|
allqueue(REDRAWNLA, 0);
|
2006-01-20 14:38:09 +00:00
|
|
|
allqueue(REDRAWTIME, 0);
|
2005-07-11 13:59:02 +00:00
|
|
|
|
2005-07-23 18:52:31 +00:00
|
|
|
DAG_object_flush_update(G.scene, ob, OB_RECALC_DATA);
|
2005-05-04 12:53:35 +00:00
|
|
|
}
|
2005-07-23 18:52:31 +00:00
|
|
|
else if(arm->flag & ARM_DELAYDEFORM) {
|
2005-07-11 13:59:02 +00:00
|
|
|
/* old optimize trick... this enforces to bypass the depgraph */
|
2005-07-23 18:52:31 +00:00
|
|
|
DAG_object_flush_update(G.scene, ob, OB_RECALC_DATA);
|
|
|
|
ob->recalc= 0; // is set on OK position already by recalcData()
|
2005-07-11 13:59:02 +00:00
|
|
|
}
|
2005-10-22 18:47:38 +00:00
|
|
|
else
|
|
|
|
DAG_object_flush_update(G.scene, ob, OB_RECALC_DATA);
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
|
|
|
|
if(t->mode==TFM_BONESIZE || t->mode==TFM_BONE_ENVELOPE)
|
|
|
|
allqueue(REDRAWBUTSEDIT, 0);
|
|
|
|
|
2005-05-04 12:53:35 +00:00
|
|
|
}
|
|
|
|
else {
|
|
|
|
base= FIRSTBASE;
|
|
|
|
while(base) {
|
|
|
|
|
|
|
|
if(base->flag & BA_DO_IPO) redrawipo= 1;
|
2005-07-19 11:21:01 +00:00
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
ob= base->object;
|
2005-08-11 02:23:52 +00:00
|
|
|
|
2005-08-26 17:48:46 +00:00
|
|
|
if(modifiers_isSoftbodyEnabled(ob)) ob->softflag |= OB_SB_REDO;
|
2005-05-04 12:53:35 +00:00
|
|
|
|
|
|
|
/* Set autokey if necessary */
|
2005-07-06 08:14:13 +00:00
|
|
|
if ((G.flags & G_RECORDKEYS) && (!cancelled) && (base->flag & SELECT)){
|
2005-11-01 22:48:07 +00:00
|
|
|
char *actname="";
|
2005-12-18 20:14:22 +00:00
|
|
|
|
2005-11-01 22:48:07 +00:00
|
|
|
if(ob->ipoflag & OB_ACTION_OB)
|
|
|
|
actname= "Object";
|
2005-12-18 20:14:22 +00:00
|
|
|
|
|
|
|
if(U.uiflag & USER_KEYINSERTAVAI) {
|
|
|
|
if(base->object->ipo) {
|
|
|
|
ID* id= (ID *)(base->object);
|
|
|
|
icu= base->object->ipo->curve.first;
|
|
|
|
while(icu) {
|
|
|
|
icu->flag &= ~IPO_SELECT;
|
|
|
|
insertkey(id, ID_OB, actname, NULL, icu->adrcode);
|
|
|
|
icu= icu->next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_ROT_X);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_ROT_Y);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_ROT_Z);
|
|
|
|
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_LOC_X);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_LOC_Y);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_LOC_Z);
|
|
|
|
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_SIZE_X);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_SIZE_Y);
|
|
|
|
insertkey(&base->object->id, ID_OB, actname, NULL, OB_SIZE_Z);
|
|
|
|
}
|
|
|
|
|
2005-05-04 12:53:35 +00:00
|
|
|
remake_object_ipos (ob);
|
|
|
|
allqueue(REDRAWIPO, 0);
|
|
|
|
allspace(REMAKEIPO, 0);
|
|
|
|
allqueue(REDRAWVIEW3D, 0);
|
|
|
|
allqueue(REDRAWNLA, 0);
|
2006-01-20 14:38:09 +00:00
|
|
|
allqueue(REDRAWTIME, 0);
|
2005-05-04 12:53:35 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
base= base->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2005-07-19 11:21:01 +00:00
|
|
|
clear_trans_object_base_flags();
|
|
|
|
|
2005-05-04 12:53:35 +00:00
|
|
|
if(redrawipo) {
|
|
|
|
allqueue(REDRAWNLA, 0);
|
|
|
|
allqueue(REDRAWACTION, 0);
|
|
|
|
allqueue(REDRAWIPO, 0);
|
|
|
|
}
|
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
reset_slowparents();
|
|
|
|
|
|
|
|
/* note; should actually only be done for all objects when a lamp is moved... (ton) */
|
2005-08-23 18:13:30 +00:00
|
|
|
if(t->spacetype==SPACE_VIEW3D && G.vd->drawtype == OB_SHADED)
|
|
|
|
reshadeall_displist();
|
2005-05-04 12:53:35 +00:00
|
|
|
}
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
static void createTransObject(TransInfo *t)
|
|
|
|
{
|
|
|
|
TransData *td = NULL;
|
|
|
|
TransDataExtension *tx;
|
|
|
|
Object *ob;
|
|
|
|
Base *base;
|
|
|
|
IpoKey *ik;
|
|
|
|
ListBase elems;
|
|
|
|
|
|
|
|
set_trans_object_base_flags(t);
|
|
|
|
|
|
|
|
/* count */
|
|
|
|
for(base= FIRSTBASE; base; base= base->next) {
|
|
|
|
if TESTBASELIB(base) {
|
|
|
|
ob= base->object;
|
|
|
|
|
|
|
|
/* store ipo keys? */
|
|
|
|
if(ob->ipo && ob->ipo->showkey && (ob->ipoflag & OB_DRAWKEY)) {
|
|
|
|
elems.first= elems.last= NULL;
|
|
|
|
make_ipokey_transform(ob, &elems, 1); /* '1' only selected keys */
|
|
|
|
|
|
|
|
pushdata(&elems, sizeof(ListBase));
|
|
|
|
|
|
|
|
for(ik= elems.first; ik; ik= ik->next) t->total++;
|
|
|
|
|
|
|
|
if(elems.first==NULL) t->total++;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
t->total++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if(!t->total) {
|
|
|
|
/* clear here, main transform function escapes too */
|
|
|
|
clear_trans_object_base_flags();
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2005-05-29 22:12:12 +00:00
|
|
|
td = t->data = MEM_callocN(t->total*sizeof(TransData), "TransOb");
|
|
|
|
tx = t->ext = MEM_callocN(t->total*sizeof(TransDataExtension), "TransObExtension");
|
2005-04-10 18:33:19 +00:00
|
|
|
|
|
|
|
for(base= FIRSTBASE; base; base= base->next) {
|
|
|
|
if TESTBASELIB(base) {
|
|
|
|
ob= base->object;
|
|
|
|
|
|
|
|
td->flag= TD_SELECTED;
|
2005-10-29 20:08:25 +00:00
|
|
|
td->protectflag= ob->protectflag;
|
2005-04-10 18:33:19 +00:00
|
|
|
td->ext = tx;
|
|
|
|
|
|
|
|
/* store ipo keys? */
|
|
|
|
if(ob->ipo && ob->ipo->showkey && (ob->ipoflag & OB_DRAWKEY)) {
|
|
|
|
|
|
|
|
popfirst(&elems); // bring back pushed listbase
|
|
|
|
|
|
|
|
if(elems.first) {
|
|
|
|
float cfraont;
|
|
|
|
int ipoflag;
|
|
|
|
|
Result of 2 weeks of quiet coding work in Greece :)
Aim was to get a total refresh of the animation system. This
is needed because;
- we need to upgrade it with 21st century features
- current code is spaghetti/hack combo, and hides good design
- it should become lag-free with using dependency graphs
A full log, with complete code API/structure/design explanation
will follow, that's a load of work... so here below the list with
hot changes;
- The entire object update system (matrices, geometry) is now
centralized. Calls to where_is_object and makeDispList are
forbidden, instead we tag objects 'changed' and let the
depgraph code sort it out
- Removed all old "Ika" code
- Depgraph is aware of all relationships, including meta balls,
constraints, bevelcurve, and so on.
- Made depgraph aware of relation types and layers, to do smart
flushing of 'changed' events. Nothing gets calculated too often!
- Transform uses depgraph to detect changes
- On frame-advance, depgraph flushes animated changes
Armatures;
Almost all armature related code has been fully built from scratch.
It now reveils the original design much better, with a very clean
implementation, lag free without even calculating each Bone more than
once. Result is quite a speedup yes!
Important to note is;
1) Armature is data containing the 'rest position'
2) Pose is the changes of rest position, and always on object level.
That way more Objects can use same Pose. Also constraints are in Pose
3) Actions only contain the Ipos to change values in Poses.
- Bones draw unrotated now
- Drawing bones speedup enormously (10-20 times)
- Bone selecting in EditMode, selection state is saved for PoseMode,
and vice-versa
- Undo in editmode
- Bone renaming does vertexgroups, constraints, posechannels, actions,
for all users of Armature in entire file
- Added Bone renaming in NKey panel
- Nkey PoseMode shows eulers now
- EditMode and PoseMode now have 'active' bone too (last clicked)
- Parenting in EditMode' CTRL+P, ALT+P, with nice options!
- Pose is added in Outliner now, with showing that constraints are in
the Pose, not Armature
- Disconnected IK solving from constraints. It's a separate phase now,
on top of the full Pose calculations
- Pose itself has a dependency graph too, so evaluation order is lag free.
TODO NOW;
- Rotating in Posemode has incorrect inverse transform (Martin will fix)
- Python Bone/Armature/Pose API disabled... needs full recode too
(wait for my doc!)
- Game engine will need upgrade too
- Depgraph code needs revision, cleanup, can be much faster!
(But, compliments for Jean-Luc, it works like a charm!)
- IK changed, it now doesnt use previous position to advance to next
position anymore. That system looks nice (no flips) but is not well
suited for NLA and background render.
TODO LATER;
We now can do loadsa new nifty features as well; like:
- Kill PoseMode (can be option for armatures itself)
- Make B-Bones (Bezier, Bspline, like for spines)
- Move all silly button level edit to 3d window (like CTRL+I = add
IK)
- Much better & informative drawing
- Fix action/nla editors
- Put all ipos in Actions (object, mesh key, lamp color)
- Add hooks
- Null bones
- Much more advanced constraints...
Bugfixes;
- OGL render (view3d header) had wrong first frame on anim render
- Ipo 'recording' mode had wrong playback speed
- Vertex-key mode now sticks to show 'active key', until frame change
-Ton-
2005-07-03 17:35:38 +00:00
|
|
|
base->flag |= BA_DO_IPO+BA_WAS_SEL;
|
2005-04-10 18:33:19 +00:00
|
|
|
base->flag &= ~SELECT;
|
|
|
|
|
|
|
|
cfraont= CFRA;
|
|
|
|
set_no_parent_ipo(1);
|
|
|
|
ipoflag= ob->ipoflag;
|
|
|
|
ob->ipoflag &= ~OB_OFFS_OB;
|
|
|
|
|
|
|
|
pushdata(ob->loc, 7*3*4); // tsk! tsk!
|
|
|
|
|
|
|
|
for(ik= elems.first; ik; ik= ik->next) {
|
|
|
|
|
|
|
|
/* weak... this doesn't correct for floating values, giving small errors */
|
|
|
|
CFRA= (short)(ik->val/G.scene->r.framelen);
|
|
|
|
|
|
|
|
do_ob_ipo(ob);
|
|
|
|
ObjectToTransData(td, ob); // does where_is_object()
|
|
|
|
|
2005-05-29 23:21:56 +00:00
|
|
|
td->flag= TD_SELECTED;
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
td->tdi= MEM_callocN(sizeof(TransDataIpokey), "TransDataIpokey");
|
|
|
|
/* also does tdi->flag and oldvals, needs to be after ob_to_transob()! */
|
|
|
|
ipokey_to_transdata(ik, td);
|
|
|
|
|
|
|
|
td++;
|
|
|
|
tx++;
|
|
|
|
if(ik->next) td->ext= tx; // prevent corrupting mem!
|
|
|
|
}
|
|
|
|
free_ipokey(&elems);
|
|
|
|
|
|
|
|
poplast(ob->loc);
|
|
|
|
set_no_parent_ipo(0);
|
|
|
|
|
|
|
|
CFRA= (short)cfraont;
|
|
|
|
ob->ipoflag= ipoflag;
|
2005-05-29 23:21:56 +00:00
|
|
|
|
|
|
|
where_is_object(ob); // restore
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
else {
|
|
|
|
ObjectToTransData(td, ob);
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
2005-05-29 22:39:49 +00:00
|
|
|
td++;
|
|
|
|
tx++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
ObjectToTransData(td, ob);
|
|
|
|
td->tdi = NULL;
|
|
|
|
td->val = NULL;
|
2005-05-29 22:39:49 +00:00
|
|
|
td++;
|
|
|
|
tx++;
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void createTransData(TransInfo *t)
|
|
|
|
{
|
2005-07-23 18:52:31 +00:00
|
|
|
Object *ob= OBACT;
|
|
|
|
|
2005-04-10 18:33:19 +00:00
|
|
|
if (t->context == CTX_TEXTURE) {
|
|
|
|
t->flag |= T_TEXTURE;
|
|
|
|
createTransTexspace(t);
|
|
|
|
}
|
|
|
|
else if (t->context == CTX_EDGE) {
|
|
|
|
t->ext = NULL;
|
|
|
|
t->flag |= T_EDIT;
|
|
|
|
createTransEdge(t);
|
|
|
|
if(t->data && t->flag & T_PROP_EDIT) {
|
|
|
|
sort_trans_data(t); // makes selected become first in array
|
|
|
|
set_prop_dist(t, 1);
|
|
|
|
sort_trans_data_dist(t);
|
|
|
|
}
|
|
|
|
}
|
2005-08-23 18:13:30 +00:00
|
|
|
else if (t->spacetype == SPACE_IMAGE) {
|
|
|
|
t->flag |= T_POINTS|T_2D_EDIT;
|
|
|
|
createTransUVs(t);
|
|
|
|
if(t->data && (t->flag & T_PROP_EDIT)) {
|
|
|
|
sort_trans_data(t); // makes selected become first in array
|
|
|
|
set_prop_dist(t, 1);
|
|
|
|
sort_trans_data_dist(t);
|
|
|
|
}
|
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
else if (G.obedit) {
|
|
|
|
t->ext = NULL;
|
|
|
|
if (G.obedit->type == OB_MESH) {
|
|
|
|
createTransEditVerts(t);
|
|
|
|
}
|
|
|
|
else if ELEM(G.obedit->type, OB_CURVE, OB_SURF) {
|
|
|
|
createTransCurveVerts(t);
|
|
|
|
}
|
|
|
|
else if (G.obedit->type==OB_LATTICE) {
|
|
|
|
createTransLatticeVerts(t);
|
|
|
|
}
|
|
|
|
else if (G.obedit->type==OB_MBALL) {
|
|
|
|
createTransMBallVerts(t);
|
|
|
|
}
|
|
|
|
else if (G.obedit->type==OB_ARMATURE) {
|
2005-12-19 16:45:11 +00:00
|
|
|
t->flag &= ~T_PROP_EDIT;
|
2005-04-10 18:33:19 +00:00
|
|
|
createTransArmatureVerts(t);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
printf("not done yet! only have mesh surface curve\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
if(t->data && t->flag & T_PROP_EDIT) {
|
|
|
|
if (ELEM(G.obedit->type, OB_CURVE, OB_MESH)) {
|
|
|
|
sort_trans_data(t); // makes selected become first in array
|
|
|
|
set_prop_dist(t, 0);
|
|
|
|
sort_trans_data_dist(t);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
sort_trans_data(t); // makes selected become first in array
|
|
|
|
set_prop_dist(t, 1);
|
|
|
|
sort_trans_data_dist(t);
|
|
|
|
}
|
|
|
|
}
|
2005-08-23 18:13:30 +00:00
|
|
|
|
|
|
|
t->flag |= T_EDIT|T_POINTS;
|
2005-07-16 19:07:02 +00:00
|
|
|
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
/* exception... hackish, we want bonesize to use bone orientation matrix (ton) */
|
2005-07-16 19:07:02 +00:00
|
|
|
if(t->mode==TFM_BONESIZE) {
|
2005-08-23 18:13:30 +00:00
|
|
|
t->flag &= ~(T_EDIT|T_POINTS);
|
2005-07-16 19:07:02 +00:00
|
|
|
t->flag |= T_POSE;
|
Armature "Envelope" editing.
For defining the deformation distances of Bones, three values are being
used now. The bone tip and root radius define the bone-shape itself and the
"dist" defines the soft area around it. A full (user) doc is in CMS here;
http://www.blender3d.org/cms/Armature_Envelopes.647.0.html
Note: todo still is allowing both Vertex Deform Groups and these Envelopes
together (and or per Bone).
Also part of this commit is:
- New: Hiding bones in EditMode. This is a separate 'hide flag', so you can
keep the PoseMode hidden Bones separate from EditMode.
(In the future we should do some kind of bone-grouping or so)
- While transform(), the hotkeys G,R,S only switch mode when the previous
mode was compatible. Caused conflicts with Crease/BoneDist/etc.
- Deleting the last VertexGroup now also deletes the entire Mesh 'dvert'
data. Sounds logical, but remember that VertexGroups are partial on a
Mesh, partial on Object. Weird design decision though...
Anyhoo, at this moment the only way to have Bone Envelopes deform, is
by deleting all VertexGroups!
- In PoseMode, the hotkey ALT+S now does both B-Bone size or Envelope,
depending draw type.
- In EditMode, Extrude now also works when only Root points were selected.
- Weight editing is also symmetrical btw, with the "X-axis Mirror" option
set.
2005-08-19 12:35:15 +00:00
|
|
|
t->poseobj= ob; /* <- tsk tsk, this is going to give issues one day */
|
2005-07-16 19:07:02 +00:00
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
2005-07-23 20:56:40 +00:00
|
|
|
else if (ob && (ob->flag & OB_POSEMODE)) {
|
2005-10-30 13:50:42 +00:00
|
|
|
createTransPose(t, OBACT);
|
2005-08-16 10:17:02 +00:00
|
|
|
}
|
|
|
|
else if (G.f & G_WEIGHTPAINT) {
|
|
|
|
/* exception, we look for the one selected armature */
|
|
|
|
Base *base;
|
|
|
|
for(base=FIRSTBASE; base; base= base->next) {
|
|
|
|
if(TESTBASELIB(base)) {
|
|
|
|
if(base->object->type==OB_ARMATURE)
|
|
|
|
if(base->object->flag & OB_POSEMODE)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if(base) {
|
2005-10-30 13:50:42 +00:00
|
|
|
createTransPose(t, base->object);
|
2005-08-16 10:17:02 +00:00
|
|
|
}
|
2005-07-23 20:56:40 +00:00
|
|
|
}
|
2005-04-10 18:33:19 +00:00
|
|
|
else {
|
|
|
|
createTransObject(t);
|
|
|
|
t->flag |= T_OBJECT;
|
|
|
|
}
|
|
|
|
|
|
|
|
if((t->flag & T_OBJECT) && G.vd->camera==OBACT && G.vd->persp>1) {
|
|
|
|
t->flag |= T_CAMERA;
|
|
|
|
}
|
Giant commit!
A full detailed description of this will be done later... is several days
of work. Here's a summary:
Render:
- Full cleanup of render code, removing *all* globals and bad level calls
all over blender. Render module is now not called abusive anymore
- API-fied calls to rendering
- Full recode of internal render pipeline. Is now rendering tiles by
default, prepared for much smarter 'bucket' render later.
- Each thread now can render a full part
- Renders were tested with 4 threads, goes fine, apart from some lookup
tables in softshadow and AO still
- Rendering is prepared to do multiple layers and passes
- No single 32 bits trick in render code anymore, all 100% floats now.
Writing images/movies
- moved writing images to blender kernel (bye bye 'schrijfplaatje'!)
- made a new Movie handle system, also in kernel. This will enable much
easier use of movies in Blender
PreviewRender:
- Using new render API, previewrender (in buttons) now uses regular render
code to generate images.
- new datafile 'preview.blend.c' has the preview scenes in it
- previews get rendered in exact displayed size (1 pixel = 1 pixel)
3D Preview render
- new; press Pkey in 3d window, for a panel that continuously renders
(pkey is for games, i know... but we dont do that in orange now!)
- this render works nearly identical to buttons-preview render, so it stops
rendering on any event (mouse, keyboard, etc)
- on moving/scaling the panel, the render code doesn't recreate all geometry
- same for shifting/panning view
- all other operations (now) regenerate the full render database still.
- this is WIP... but big fun, especially for simple scenes!
Compositor
- Using same node system as now in use for shaders, you can composit images
- works pretty straightforward... needs much more options/tools and integration
with rendering still
- is not threaded yet, nor is so smart to only recalculate changes... will be
done soon!
- the "Render Result" node will get all layers/passes as output sockets
- The "Output" node renders to a builtin image, which you can view in the Image
window. (yes, output nodes to render-result, and to files, is on the list!)
The Bad News
- "Unified Render" is removed. It might come back in some stage, but this
system should be built from scratch. I can't really understand this code...
I expect it is not much needed, especially with advanced layer/passes
control
- Panorama render, Field render, Motion blur, is not coded yet... (I had to
recode every single feature in render, so...!)
- Lens Flare is also not back... needs total revision, might become composit
effect though (using zbuffer for visibility)
- Part render is gone! (well, thats obvious, its default now).
- The render window is only restored with limited functionality... I am going
to check first the option to render to a Image window, so Blender can become
a true single-window application. :)
For example, the 'Spare render buffer' (jkey) doesnt work.
- Render with border, now default creates a smaller image
- No zbuffers are written yet... on the todo!
- Scons files and MSVC will need work to get compiling again
OK... thats what I can quickly recall. Now go compiling!
2006-01-23 22:05:47 +00:00
|
|
|
|
|
|
|
/* temporal...? */
|
|
|
|
G.scene->recalc |= SCE_PRV_CHANGED; /* test for 3d preview */
|
2005-04-10 18:33:19 +00:00
|
|
|
}
|
|
|
|
|