This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
blender-archive/source/blender/blenkernel/intern/cloth.c
Campbell Barton 8811e521c0 fix for leak in cloth modifier
- any error in cloth_build_springs wasn't freeing the edge-hash.
- was checking BLI_edgehash_haskey on matching vertices.
- was looping over setting NULL for all elements of a calloc'd array.
2013-08-24 14:32:43 +00:00

1308 lines
36 KiB
C

/*
* ***** BEGIN GPL 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.
*
* 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* The Original Code is Copyright (C) Blender Foundation
* All rights reserved.
*
* Contributor(s): Daniel Genrich
*
* ***** END GPL LICENSE BLOCK *****
*/
/** \file blender/blenkernel/intern/cloth.c
* \ingroup bke
*/
#include "MEM_guardedalloc.h"
#include "DNA_cloth_types.h"
#include "DNA_scene_types.h"
#include "DNA_object_types.h"
#include "DNA_meshdata_types.h"
#include "BLI_utildefines.h"
#include "BLI_math.h"
#include "BLI_edgehash.h"
#include "BLI_utildefines.h"
#include "BLI_linklist.h"
#include "BKE_cdderivedmesh.h"
#include "BKE_cloth.h"
#include "BKE_effect.h"
#include "BKE_global.h"
#include "BKE_modifier.h"
#include "BKE_pointcache.h"
// #include "PIL_time.h" /* timing for debug prints */
/* Our available solvers. */
// 255 is the magic reserved number, so NEVER try to put 255 solvers in here!
// 254 = MAX!
static CM_SOLVER_DEF solvers [] =
{
{ "Implicit", CM_IMPLICIT, implicit_init, implicit_solver, implicit_free },
// { "Implicit C++", CM_IMPLICITCPP, implicitcpp_init, implicitcpp_solver, implicitcpp_free },
};
/* ********** cloth engine ******* */
/* Prototypes for internal functions.
*/
static void cloth_to_object (Object *ob, ClothModifierData *clmd, float (*vertexCos)[3]);
static void cloth_from_mesh ( ClothModifierData *clmd, DerivedMesh *dm );
static int cloth_from_object(Object *ob, ClothModifierData *clmd, DerivedMesh *dm, float framenr, int first);
static void cloth_update_springs( ClothModifierData *clmd );
static int cloth_build_springs ( ClothModifierData *clmd, DerivedMesh *dm );
static void cloth_apply_vgroup ( ClothModifierData *clmd, DerivedMesh *dm );
/******************************************************************************
*
* External interface called by modifier.c clothModifier functions.
*
******************************************************************************/
/**
* cloth_init - creates a new cloth simulation.
*
* 1. create object
* 2. fill object with standard values or with the GUI settings if given
*/
void cloth_init(ClothModifierData *clmd )
{
/* Initialize our new data structure to reasonable values. */
clmd->sim_parms->gravity[0] = 0.0;
clmd->sim_parms->gravity[1] = 0.0;
clmd->sim_parms->gravity[2] = -9.81;
clmd->sim_parms->structural = 15.0;
clmd->sim_parms->shear = 15.0;
clmd->sim_parms->bending = 0.5;
clmd->sim_parms->Cdis = 5.0;
clmd->sim_parms->Cvi = 1.0;
clmd->sim_parms->mass = 0.3f;
clmd->sim_parms->stepsPerFrame = 5;
clmd->sim_parms->flags = 0;
clmd->sim_parms->solver_type = 0;
clmd->sim_parms->preroll = 0;
clmd->sim_parms->maxspringlen = 10;
clmd->sim_parms->vgroup_mass = 0;
clmd->sim_parms->avg_spring_len = 0.0;
clmd->sim_parms->presets = 2; /* cotton as start setting */
clmd->sim_parms->timescale = 1.0f; /* speed factor, describes how fast cloth moves */
clmd->sim_parms->reset = 0;
clmd->sim_parms->vel_damping = 1.0f; /* 1.0 = no damping, 0.0 = fully dampened */
clmd->coll_parms->self_friction = 5.0;
clmd->coll_parms->friction = 5.0;
clmd->coll_parms->loop_count = 2;
clmd->coll_parms->epsilon = 0.015f;
clmd->coll_parms->flags = CLOTH_COLLSETTINGS_FLAG_ENABLED;
clmd->coll_parms->collision_list = NULL;
clmd->coll_parms->self_loop_count = 1.0;
clmd->coll_parms->selfepsilon = 0.75;
clmd->coll_parms->vgroup_selfcol = 0;
/* These defaults are copied from softbody.c's
* softbody_calc_forces() function.
*/
clmd->sim_parms->eff_force_scale = 1000.0;
clmd->sim_parms->eff_wind_scale = 250.0;
// also from softbodies
clmd->sim_parms->maxgoal = 1.0f;
clmd->sim_parms->mingoal = 0.0f;
clmd->sim_parms->defgoal = 0.0f;
clmd->sim_parms->goalspring = 1.0f;
clmd->sim_parms->goalfrict = 0.0f;
clmd->sim_parms->velocity_smooth = 0.0f;
if (!clmd->sim_parms->effector_weights)
clmd->sim_parms->effector_weights = BKE_add_effector_weights(NULL);
if (clmd->point_cache)
clmd->point_cache->step = 1;
}
static BVHTree *bvhselftree_build_from_cloth (ClothModifierData *clmd, float epsilon)
{
unsigned int i;
BVHTree *bvhtree;
Cloth *cloth;
ClothVertex *verts;
float co[12];
if (!clmd)
return NULL;
cloth = clmd->clothObject;
if (!cloth)
return NULL;
verts = cloth->verts;
// in the moment, return zero if no faces there
if (!cloth->numverts)
return NULL;
// create quadtree with k=26
bvhtree = BLI_bvhtree_new(cloth->numverts, epsilon, 4, 6);
// fill tree
for (i = 0; i < cloth->numverts; i++, verts++) {
copy_v3_v3(&co[0*3], verts->xold);
BLI_bvhtree_insert(bvhtree, i, co, 1);
}
// balance tree
BLI_bvhtree_balance(bvhtree);
return bvhtree;
}
static BVHTree *bvhtree_build_from_cloth (ClothModifierData *clmd, float epsilon)
{
unsigned int i;
BVHTree *bvhtree;
Cloth *cloth;
ClothVertex *verts;
MFace *mfaces;
float co[12];
if (!clmd)
return NULL;
cloth = clmd->clothObject;
if (!cloth)
return NULL;
verts = cloth->verts;
mfaces = cloth->mfaces;
/* in the moment, return zero if no faces there */
if (!cloth->numfaces)
return NULL;
/* create quadtree with k=26 */
bvhtree = BLI_bvhtree_new(cloth->numfaces, epsilon, 4, 26);
/* fill tree */
for (i = 0; i < cloth->numfaces; i++, mfaces++) {
copy_v3_v3(&co[0*3], verts[mfaces->v1].xold);
copy_v3_v3(&co[1*3], verts[mfaces->v2].xold);
copy_v3_v3(&co[2*3], verts[mfaces->v3].xold);
if (mfaces->v4)
copy_v3_v3(&co[3*3], verts[mfaces->v4].xold);
BLI_bvhtree_insert(bvhtree, i, co, (mfaces->v4 ? 4 : 3));
}
/* balance tree */
BLI_bvhtree_balance(bvhtree);
return bvhtree;
}
void bvhtree_update_from_cloth(ClothModifierData *clmd, int moving)
{
unsigned int i = 0;
Cloth *cloth = clmd->clothObject;
BVHTree *bvhtree = cloth->bvhtree;
ClothVertex *verts = cloth->verts;
MFace *mfaces;
float co[12], co_moving[12];
int ret = 0;
if (!bvhtree)
return;
mfaces = cloth->mfaces;
// update vertex position in bvh tree
if (verts && mfaces) {
for (i = 0; i < cloth->numfaces; i++, mfaces++) {
copy_v3_v3(&co[0*3], verts[mfaces->v1].txold);
copy_v3_v3(&co[1*3], verts[mfaces->v2].txold);
copy_v3_v3(&co[2*3], verts[mfaces->v3].txold);
if (mfaces->v4)
copy_v3_v3(&co[3*3], verts[mfaces->v4].txold);
// copy new locations into array
if (moving) {
// update moving positions
copy_v3_v3(&co_moving[0*3], verts[mfaces->v1].tx);
copy_v3_v3(&co_moving[1*3], verts[mfaces->v2].tx);
copy_v3_v3(&co_moving[2*3], verts[mfaces->v3].tx);
if (mfaces->v4)
copy_v3_v3(&co_moving[3*3], verts[mfaces->v4].tx);
ret = BLI_bvhtree_update_node(bvhtree, i, co, co_moving, (mfaces->v4 ? 4 : 3));
}
else {
ret = BLI_bvhtree_update_node(bvhtree, i, co, NULL, (mfaces->v4 ? 4 : 3));
}
// check if tree is already full
if (!ret)
break;
}
BLI_bvhtree_update_tree(bvhtree);
}
}
void bvhselftree_update_from_cloth(ClothModifierData *clmd, int moving)
{
unsigned int i = 0;
Cloth *cloth = clmd->clothObject;
BVHTree *bvhtree = cloth->bvhselftree;
ClothVertex *verts = cloth->verts;
MFace *mfaces;
float co[12], co_moving[12];
int ret = 0;
if (!bvhtree)
return;
mfaces = cloth->mfaces;
// update vertex position in bvh tree
if (verts && mfaces) {
for (i = 0; i < cloth->numverts; i++, verts++) {
copy_v3_v3(&co[0*3], verts->txold);
// copy new locations into array
if (moving) {
// update moving positions
copy_v3_v3(&co_moving[0*3], verts->tx);
ret = BLI_bvhtree_update_node(bvhtree, i, co, co_moving, 1);
}
else {
ret = BLI_bvhtree_update_node(bvhtree, i, co, NULL, 1);
}
// check if tree is already full
if (!ret)
break;
}
BLI_bvhtree_update_tree(bvhtree);
}
}
void cloth_clear_cache(Object *ob, ClothModifierData *clmd, float framenr)
{
PTCacheID pid;
BKE_ptcache_id_from_cloth(&pid, ob, clmd);
// don't do anything as long as we're in editmode!
if (pid.cache->edit && ob->mode & OB_MODE_PARTICLE_EDIT)
return;
BKE_ptcache_id_clear(&pid, PTCACHE_CLEAR_AFTER, framenr);
}
static int do_init_cloth(Object *ob, ClothModifierData *clmd, DerivedMesh *result, int framenr)
{
PointCache *cache;
cache= clmd->point_cache;
/* initialize simulation data if it didn't exist already */
if (clmd->clothObject == NULL) {
if (!cloth_from_object(ob, clmd, result, framenr, 1)) {
BKE_ptcache_invalidate(cache);
modifier_setError(&(clmd->modifier), "Can't initialize cloth");
return 0;
}
if (clmd->clothObject == NULL) {
BKE_ptcache_invalidate(cache);
modifier_setError(&(clmd->modifier), "Null cloth object");
return 0;
}
implicit_set_positions(clmd);
clmd->clothObject->last_frame= MINFRAME-1;
}
return 1;
}
static int do_step_cloth(Object *ob, ClothModifierData *clmd, DerivedMesh *result, int framenr)
{
ClothVertex *verts = NULL;
Cloth *cloth;
ListBase *effectors = NULL;
MVert *mvert;
unsigned int i = 0;
int ret = 0;
/* simulate 1 frame forward */
cloth = clmd->clothObject;
verts = cloth->verts;
mvert = result->getVertArray(result);
/* force any pinned verts to their constrained location. */
for (i = 0; i < clmd->clothObject->numverts; i++, verts++) {
/* save the previous position. */
copy_v3_v3(verts->xold, verts->xconst);
copy_v3_v3(verts->txold, verts->x);
/* Get the current position. */
copy_v3_v3(verts->xconst, mvert[i].co);
mul_m4_v3(ob->obmat, verts->xconst);
}
effectors = pdInitEffectors(clmd->scene, ob, NULL, clmd->sim_parms->effector_weights);
/* Support for dynamic vertex groups, changing from frame to frame */
cloth_apply_vgroup ( clmd, result );
cloth_update_springs( clmd );
// TIMEIT_START(cloth_step)
/* call the solver. */
if (solvers [clmd->sim_parms->solver_type].solver)
ret = solvers[clmd->sim_parms->solver_type].solver(ob, framenr, clmd, effectors);
// TIMEIT_END(cloth_step)
pdEndEffectors(&effectors);
// printf ( "%f\n", ( float ) tval() );
return ret;
}
#if 0
static DerivedMesh *cloth_to_triangles(DerivedMesh *dm)
{
DerivedMesh *result = NULL;
unsigned int i = 0, j = 0;
unsigned int quads = 0, numfaces = dm->getNumTessFaces(dm);
MFace *mface = dm->getTessFaceArray(dm);
MFace *mface2 = NULL;
/* calc faces */
for (i = 0; i < numfaces; i++) {
if (mface[i].v4) {
quads++;
}
}
result = CDDM_from_template(dm, dm->getNumVerts(dm), 0, numfaces + quads, 0, 0);
DM_copy_vert_data(dm, result, 0, 0, dm->getNumVerts(dm));
DM_copy_tessface_data(dm, result, 0, 0, numfaces);
DM_ensure_tessface(result);
mface2 = result->getTessFaceArray(result);
for (i = 0, j = numfaces; i < numfaces; i++) {
// DG TODO: is this necessary?
mface2[i].v1 = mface[i].v1;
mface2[i].v2 = mface[i].v2;
mface2[i].v3 = mface[i].v3;
mface2[i].v4 = 0;
//test_index_face(&mface2[i], &result->faceData, i, 3);
if (mface[i].v4) {
DM_copy_tessface_data(dm, result, i, j, 1);
mface2[j].v1 = mface[i].v1;
mface2[j].v2 = mface[i].v3;
mface2[j].v3 = mface[i].v4;
mface2[j].v4 = 0;
//test_index_face(&mface2[j], &result->faceData, j, 3);
j++;
}
}
CDDM_calc_edges_tessface(result);
CDDM_tessfaces_to_faces(result); /* builds ngon faces from tess (mface) faces */
return result;
}
#endif
/************************************************
* clothModifier_do - main simulation function
************************************************/
void clothModifier_do(ClothModifierData *clmd, Scene *scene, Object *ob, DerivedMesh *dm, float (*vertexCos)[3])
{
PointCache *cache;
PTCacheID pid;
float timescale;
int framenr, startframe, endframe;
int cache_result;
clmd->scene= scene; /* nice to pass on later :) */
framenr= (int)scene->r.cfra;
cache= clmd->point_cache;
BKE_ptcache_id_from_cloth(&pid, ob, clmd);
BKE_ptcache_id_time(&pid, scene, framenr, &startframe, &endframe, &timescale);
clmd->sim_parms->timescale= timescale;
if (clmd->sim_parms->reset ||
(framenr == (startframe - clmd->sim_parms->preroll) && clmd->sim_parms->preroll != 0) ||
(clmd->clothObject && dm->getNumVerts(dm) != clmd->clothObject->numverts))
{
clmd->sim_parms->reset = 0;
cache->flag |= PTCACHE_OUTDATED;
BKE_ptcache_id_reset(scene, &pid, PTCACHE_RESET_OUTDATED);
BKE_ptcache_validate(cache, 0);
cache->last_exact= 0;
cache->flag &= ~PTCACHE_REDO_NEEDED;
}
// unused in the moment, calculated separately in implicit.c
clmd->sim_parms->dt = clmd->sim_parms->timescale / clmd->sim_parms->stepsPerFrame;
/* handle continuous simulation with the play button */
if ((clmd->sim_parms->preroll > 0) && (framenr > startframe - clmd->sim_parms->preroll) && (framenr < startframe)) {
BKE_ptcache_invalidate(cache);
/* do simulation */
if (!do_init_cloth(ob, clmd, dm, framenr))
return;
do_step_cloth(ob, clmd, dm, framenr);
cloth_to_object(ob, clmd, vertexCos);
clmd->clothObject->last_frame= framenr;
return;
}
/* simulation is only active during a specific period */
if (framenr < startframe) {
BKE_ptcache_invalidate(cache);
return;
}
else if (framenr > endframe) {
framenr= endframe;
}
/* initialize simulation data if it didn't exist already */
if (!do_init_cloth(ob, clmd, dm, framenr))
return;
if ((framenr == startframe) && (clmd->sim_parms->preroll == 0)) {
BKE_ptcache_id_reset(scene, &pid, PTCACHE_RESET_OUTDATED);
do_init_cloth(ob, clmd, dm, framenr);
BKE_ptcache_validate(cache, framenr);
cache->flag &= ~PTCACHE_REDO_NEEDED;
clmd->clothObject->last_frame= framenr;
return;
}
/* try to read from cache */
cache_result = BKE_ptcache_read(&pid, (float)framenr+scene->r.subframe);
if (cache_result == PTCACHE_READ_EXACT || cache_result == PTCACHE_READ_INTERPOLATED) {
implicit_set_positions(clmd);
cloth_to_object (ob, clmd, vertexCos);
BKE_ptcache_validate(cache, framenr);
if (cache_result == PTCACHE_READ_INTERPOLATED && cache->flag & PTCACHE_REDO_NEEDED)
BKE_ptcache_write(&pid, framenr);
clmd->clothObject->last_frame= framenr;
return;
}
else if (cache_result==PTCACHE_READ_OLD) {
implicit_set_positions(clmd);
}
else if ( /*ob->id.lib ||*/ (cache->flag & PTCACHE_BAKED)) { /* 2.4x disabled lib, but this can be used in some cases, testing further - campbell */
/* if baked and nothing in cache, do nothing */
BKE_ptcache_invalidate(cache);
return;
}
if (framenr!=clmd->clothObject->last_frame+1)
return;
/* if on second frame, write cache for first frame */
if (cache->simframe == startframe && (cache->flag & PTCACHE_OUTDATED || cache->last_exact==0))
BKE_ptcache_write(&pid, startframe);
clmd->sim_parms->timescale *= framenr - cache->simframe;
/* do simulation */
BKE_ptcache_validate(cache, framenr);
if (!do_step_cloth(ob, clmd, dm, framenr)) {
BKE_ptcache_invalidate(cache);
}
else
BKE_ptcache_write(&pid, framenr);
cloth_to_object (ob, clmd, vertexCos);
clmd->clothObject->last_frame= framenr;
}
/* frees all */
void cloth_free_modifier(ClothModifierData *clmd )
{
Cloth *cloth = NULL;
if ( !clmd )
return;
cloth = clmd->clothObject;
if ( cloth ) {
// If our solver provides a free function, call it
if ( solvers [clmd->sim_parms->solver_type].free ) {
solvers [clmd->sim_parms->solver_type].free ( clmd );
}
// Free the verts.
if ( cloth->verts != NULL )
MEM_freeN ( cloth->verts );
cloth->verts = NULL;
cloth->numverts = 0;
// Free the springs.
if ( cloth->springs != NULL ) {
LinkNode *search = cloth->springs;
while (search) {
ClothSpring *spring = search->link;
MEM_freeN ( spring );
search = search->next;
}
BLI_linklist_free(cloth->springs, NULL);
cloth->springs = NULL;
}
cloth->springs = NULL;
cloth->numsprings = 0;
// free BVH collision tree
if ( cloth->bvhtree )
BLI_bvhtree_free ( cloth->bvhtree );
if ( cloth->bvhselftree )
BLI_bvhtree_free ( cloth->bvhselftree );
// we save our faces for collision objects
if ( cloth->mfaces )
MEM_freeN ( cloth->mfaces );
if (cloth->edgehash)
BLI_edgehash_free ( cloth->edgehash, NULL );
/*
if (clmd->clothObject->facemarks)
MEM_freeN(clmd->clothObject->facemarks);
*/
MEM_freeN ( cloth );
clmd->clothObject = NULL;
}
}
/* frees all */
void cloth_free_modifier_extern(ClothModifierData *clmd )
{
Cloth *cloth = NULL;
if (G.debug_value > 0)
printf("cloth_free_modifier_extern\n");
if ( !clmd )
return;
cloth = clmd->clothObject;
if ( cloth ) {
if (G.debug_value > 0)
printf("cloth_free_modifier_extern in\n");
// If our solver provides a free function, call it
if ( solvers [clmd->sim_parms->solver_type].free ) {
solvers [clmd->sim_parms->solver_type].free ( clmd );
}
// Free the verts.
if ( cloth->verts != NULL )
MEM_freeN ( cloth->verts );
cloth->verts = NULL;
cloth->numverts = 0;
// Free the springs.
if ( cloth->springs != NULL ) {
LinkNode *search = cloth->springs;
while (search) {
ClothSpring *spring = search->link;
MEM_freeN ( spring );
search = search->next;
}
BLI_linklist_free(cloth->springs, NULL);
cloth->springs = NULL;
}
cloth->springs = NULL;
cloth->numsprings = 0;
// free BVH collision tree
if ( cloth->bvhtree )
BLI_bvhtree_free ( cloth->bvhtree );
if ( cloth->bvhselftree )
BLI_bvhtree_free ( cloth->bvhselftree );
// we save our faces for collision objects
if ( cloth->mfaces )
MEM_freeN ( cloth->mfaces );
if (cloth->edgehash)
BLI_edgehash_free ( cloth->edgehash, NULL );
/*
if (clmd->clothObject->facemarks)
MEM_freeN(clmd->clothObject->facemarks);
*/
MEM_freeN ( cloth );
clmd->clothObject = NULL;
}
}
/******************************************************************************
*
* Internal functions.
*
******************************************************************************/
/**
* cloth_to_object - copies the deformed vertices to the object.
*
**/
static void cloth_to_object (Object *ob, ClothModifierData *clmd, float (*vertexCos)[3])
{
unsigned int i = 0;
Cloth *cloth = clmd->clothObject;
if (clmd->clothObject) {
/* inverse matrix is not uptodate... */
invert_m4_m4(ob->imat, ob->obmat);
for (i = 0; i < cloth->numverts; i++) {
copy_v3_v3 (vertexCos[i], cloth->verts[i].x);
mul_m4_v3(ob->imat, vertexCos[i]); /* cloth is in global coords */
}
}
}
int cloth_uses_vgroup(ClothModifierData *clmd)
{
return (((clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_SCALING ) ||
(clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_GOAL ) ||
(clmd->coll_parms->flags & CLOTH_COLLSETTINGS_FLAG_SELF)) &&
((clmd->sim_parms->vgroup_mass>0) ||
(clmd->sim_parms->vgroup_struct>0)||
(clmd->sim_parms->vgroup_bend>0) ||
(clmd->coll_parms->vgroup_selfcol>0)));
}
/**
* cloth_apply_vgroup - applies a vertex group as specified by type
*
**/
/* can be optimized to do all groups in one loop */
static void cloth_apply_vgroup ( ClothModifierData *clmd, DerivedMesh *dm )
{
int i = 0;
int j = 0;
MDeformVert *dvert = NULL;
Cloth *clothObj = NULL;
int numverts;
/* float goalfac = 0; */ /* UNUSED */
ClothVertex *verts = NULL;
if (!clmd || !dm) return;
clothObj = clmd->clothObject;
numverts = dm->getNumVerts (dm);
verts = clothObj->verts;
if (cloth_uses_vgroup(clmd)) {
for ( i = 0; i < numverts; i++, verts++ ) {
/* Reset Goal values to standard */
if ( clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_GOAL )
verts->goal= clmd->sim_parms->defgoal;
else
verts->goal= 0.0f;
/* Reset vertex flags */
verts->flags &= ~CLOTH_VERT_FLAG_PINNED;
verts->flags &= ~CLOTH_VERT_FLAG_NOSELFCOLL;
dvert = dm->getVertData ( dm, i, CD_MDEFORMVERT );
if ( dvert ) {
for ( j = 0; j < dvert->totweight; j++ ) {
if (( dvert->dw[j].def_nr == (clmd->sim_parms->vgroup_mass-1)) && (clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_GOAL )) {
verts->goal = dvert->dw [j].weight;
/* goalfac= 1.0f; */ /* UNUSED */
// Kicking goal factor to simplify things...who uses that anyway?
// ABS ( clmd->sim_parms->maxgoal - clmd->sim_parms->mingoal );
verts->goal = powf(verts->goal, 4.0f);
if ( verts->goal >= SOFTGOALSNAP )
verts->flags |= CLOTH_VERT_FLAG_PINNED;
}
if (clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_SCALING ) {
if ( dvert->dw[j].def_nr == (clmd->sim_parms->vgroup_struct-1)) {
verts->struct_stiff = dvert->dw [j].weight;
verts->shear_stiff = dvert->dw [j].weight;
}
if ( dvert->dw[j].def_nr == (clmd->sim_parms->vgroup_bend-1)) {
verts->bend_stiff = dvert->dw [j].weight;
}
}
if (clmd->coll_parms->flags & CLOTH_COLLSETTINGS_FLAG_SELF ) {
if ( dvert->dw[j].def_nr == (clmd->coll_parms->vgroup_selfcol-1)) {
if (dvert->dw [j].weight > 0.0f) {
verts->flags |= CLOTH_VERT_FLAG_NOSELFCOLL;
}
}
}
}
}
}
}
}
static int cloth_from_object(Object *ob, ClothModifierData *clmd, DerivedMesh *dm, float UNUSED(framenr), int first)
{
int i = 0;
MVert *mvert = NULL;
ClothVertex *verts = NULL;
float (*shapekey_rest)[3] = NULL;
float tnull[3] = {0, 0, 0};
Cloth *cloth = NULL;
float maxdist = 0;
// If we have a clothObject, free it.
if ( clmd->clothObject != NULL ) {
cloth_free_modifier ( clmd );
if (G.debug_value > 0)
printf("cloth_free_modifier cloth_from_object\n");
}
// Allocate a new cloth object.
clmd->clothObject = MEM_callocN ( sizeof ( Cloth ), "cloth" );
if ( clmd->clothObject ) {
clmd->clothObject->old_solver_type = 255;
// clmd->clothObject->old_collision_type = 255;
cloth = clmd->clothObject;
clmd->clothObject->edgehash = NULL;
}
else if (!clmd->clothObject) {
modifier_setError(&(clmd->modifier), "Out of memory on allocating clmd->clothObject");
return 0;
}
// mesh input objects need DerivedMesh
if ( !dm )
return 0;
cloth_from_mesh ( clmd, dm );
// create springs
clmd->clothObject->springs = NULL;
clmd->clothObject->numsprings = -1;
if ( clmd->sim_parms->shapekey_rest )
shapekey_rest = dm->getVertDataArray ( dm, CD_CLOTH_ORCO );
mvert = dm->getVertArray (dm);
verts = clmd->clothObject->verts;
// set initial values
for ( i = 0; i < dm->getNumVerts(dm); i++, verts++ ) {
if (first) {
copy_v3_v3(verts->x, mvert[i].co);
mul_m4_v3(ob->obmat, verts->x);
if ( shapekey_rest ) {
verts->xrest= shapekey_rest[i];
mul_m4_v3(ob->obmat, verts->xrest);
}
else
verts->xrest = verts->x;
}
/* no GUI interface yet */
verts->mass = clmd->sim_parms->mass;
verts->impulse_count = 0;
if ( clmd->sim_parms->flags & CLOTH_SIMSETTINGS_FLAG_GOAL )
verts->goal= clmd->sim_parms->defgoal;
else
verts->goal= 0.0f;
verts->flags = 0;
copy_v3_v3 ( verts->xold, verts->x );
copy_v3_v3 ( verts->xconst, verts->x );
copy_v3_v3 ( verts->txold, verts->x );
copy_v3_v3 ( verts->tx, verts->x );
mul_v3_fl(verts->v, 0.0f);
verts->impulse_count = 0;
copy_v3_v3 ( verts->impulse, tnull );
}
// apply / set vertex groups
// has to be happen before springs are build!
cloth_apply_vgroup (clmd, dm);
if ( !cloth_build_springs ( clmd, dm ) ) {
cloth_free_modifier ( clmd );
modifier_setError(&(clmd->modifier), "Cannot build springs");
printf("cloth_free_modifier cloth_build_springs\n");
return 0;
}
for ( i = 0; i < dm->getNumVerts(dm); i++) {
if ((!(cloth->verts[i].flags & CLOTH_VERT_FLAG_PINNED)) && (cloth->verts[i].goal > ALMOST_ZERO)) {
cloth_add_spring (clmd, i, i, 0.0, CLOTH_SPRING_TYPE_GOAL);
}
}
// init our solver
if ( solvers [clmd->sim_parms->solver_type].init ) {
solvers [clmd->sim_parms->solver_type].init ( ob, clmd );
}
if (!first)
implicit_set_positions(clmd);
clmd->clothObject->bvhtree = bvhtree_build_from_cloth ( clmd, MAX2(clmd->coll_parms->epsilon, clmd->coll_parms->distance_repel) );
for (i = 0; i < dm->getNumVerts(dm); i++) {
maxdist = MAX2(maxdist, clmd->coll_parms->selfepsilon* ( cloth->verts[i].avg_spring_len*2.0f));
}
clmd->clothObject->bvhselftree = bvhselftree_build_from_cloth ( clmd, maxdist );
return 1;
}
static void cloth_from_mesh ( ClothModifierData *clmd, DerivedMesh *dm )
{
unsigned int numverts = dm->getNumVerts (dm);
unsigned int numfaces = dm->getNumTessFaces (dm);
MFace *mface = dm->getTessFaceArray(dm);
unsigned int i = 0;
/* Allocate our vertices. */
clmd->clothObject->numverts = numverts;
clmd->clothObject->verts = MEM_callocN ( sizeof ( ClothVertex ) * clmd->clothObject->numverts, "clothVertex" );
if ( clmd->clothObject->verts == NULL ) {
cloth_free_modifier ( clmd );
modifier_setError(&(clmd->modifier), "Out of memory on allocating clmd->clothObject->verts");
printf("cloth_free_modifier clmd->clothObject->verts\n");
return;
}
// save face information
clmd->clothObject->numfaces = numfaces;
clmd->clothObject->mfaces = MEM_callocN ( sizeof ( MFace ) * clmd->clothObject->numfaces, "clothMFaces" );
if ( clmd->clothObject->mfaces == NULL ) {
cloth_free_modifier ( clmd );
modifier_setError(&(clmd->modifier), "Out of memory on allocating clmd->clothObject->mfaces");
printf("cloth_free_modifier clmd->clothObject->mfaces\n");
return;
}
for ( i = 0; i < numfaces; i++ )
memcpy ( &clmd->clothObject->mfaces[i], &mface[i], sizeof ( MFace ) );
/* Free the springs since they can't be correct if the vertices
* changed.
*/
if ( clmd->clothObject->springs != NULL )
MEM_freeN ( clmd->clothObject->springs );
}
/***************************************************************************************
* SPRING NETWORK BUILDING IMPLEMENTATION BEGIN
***************************************************************************************/
// be careful: implicit solver has to be resettet when using this one!
// --> only for implicit handling of this spring!
int cloth_add_spring(ClothModifierData *clmd, unsigned int indexA, unsigned int indexB, float restlength, int spring_type)
{
Cloth *cloth = clmd->clothObject;
ClothSpring *spring = NULL;
if (cloth) {
// TODO: look if this spring is already there
spring = (ClothSpring *)MEM_callocN ( sizeof ( ClothSpring ), "cloth spring" );
if (!spring)
return 0;
spring->ij = indexA;
spring->kl = indexB;
spring->restlen = restlength;
spring->type = spring_type;
spring->flags = 0;
spring->stiffness = 0;
cloth->numsprings++;
BLI_linklist_prepend ( &cloth->springs, spring );
return 1;
}
return 0;
}
static void cloth_free_edgelist(LinkNode **edgelist, unsigned int numverts)
{
if (edgelist) {
unsigned int i;
for (i = 0; i < numverts; i++) {
BLI_linklist_free(edgelist[i], NULL);
}
MEM_freeN(edgelist);
}
}
static void cloth_free_errorsprings(Cloth *cloth, LinkNode **edgelist)
{
if ( cloth->springs != NULL ) {
LinkNode *search = cloth->springs;
while (search) {
ClothSpring *spring = search->link;
MEM_freeN ( spring );
search = search->next;
}
BLI_linklist_free(cloth->springs, NULL);
cloth->springs = NULL;
}
cloth_free_edgelist(edgelist, cloth->numverts);
if (cloth->edgehash) {
BLI_edgehash_free(cloth->edgehash, NULL);
cloth->edgehash = NULL;
}
}
/* update stiffness if vertex group values are changing from frame to frame */
static void cloth_update_springs( ClothModifierData *clmd )
{
Cloth *cloth = clmd->clothObject;
LinkNode *search = NULL;
search = cloth->springs;
while (search) {
ClothSpring *spring = search->link;
spring->stiffness = 0.0f;
if (spring->type == CLOTH_SPRING_TYPE_STRUCTURAL) {
spring->stiffness = (cloth->verts[spring->kl].struct_stiff + cloth->verts[spring->ij].struct_stiff) / 2.0f;
}
else if (spring->type == CLOTH_SPRING_TYPE_SHEAR) {
spring->stiffness = (cloth->verts[spring->kl].shear_stiff + cloth->verts[spring->ij].shear_stiff) / 2.0f;
}
else if (spring->type == CLOTH_SPRING_TYPE_BENDING) {
spring->stiffness = (cloth->verts[spring->kl].bend_stiff + cloth->verts[spring->ij].bend_stiff) / 2.0f;
}
else if (spring->type == CLOTH_SPRING_TYPE_GOAL) {
/* Warning: Appending NEW goal springs does not work because implicit solver would need reset! */
/* Activate / Deactivate existing springs */
if ((!(cloth->verts[spring->ij].flags & CLOTH_VERT_FLAG_PINNED)) &&
(cloth->verts[spring->ij].goal > ALMOST_ZERO))
{
spring->flags &= ~CLOTH_SPRING_FLAG_DEACTIVATE;
}
else {
spring->flags |= CLOTH_SPRING_FLAG_DEACTIVATE;
}
}
search = search->next;
}
}
static int cloth_build_springs ( ClothModifierData *clmd, DerivedMesh *dm )
{
Cloth *cloth = clmd->clothObject;
ClothSpring *spring = NULL, *tspring = NULL, *tspring2 = NULL;
unsigned int struct_springs = 0, shear_springs=0, bend_springs = 0;
unsigned int i = 0;
unsigned int numverts = (unsigned int)dm->getNumVerts (dm);
unsigned int numedges = (unsigned int)dm->getNumEdges (dm);
unsigned int numfaces = (unsigned int)dm->getNumTessFaces (dm);
MEdge *medge = dm->getEdgeArray (dm);
MFace *mface = dm->getTessFaceArray (dm);
int index2 = 0; // our second vertex index
LinkNode **edgelist = NULL;
EdgeHash *edgehash = NULL;
LinkNode *search = NULL, *search2 = NULL;
// error handling
if ( numedges==0 )
return 0;
/* NOTE: handling ownership of sptings and edgehash is quite sloppy
* currenlty they are never initialized but assert just to be sure */
BLI_assert(cloth->springs == NULL);
BLI_assert(cloth->edgehash == NULL);
cloth->springs = NULL;
cloth->edgehash = NULL;
edgelist = MEM_callocN ( sizeof (LinkNode *) * numverts, "cloth_edgelist_alloc" );
if (!edgelist)
return 0;
// structural springs
for ( i = 0; i < numedges; i++ ) {
spring = (ClothSpring *)MEM_callocN ( sizeof ( ClothSpring ), "cloth spring" );
if ( spring ) {
spring->ij = MIN2(medge[i].v1, medge[i].v2);
spring->kl = MAX2(medge[i].v2, medge[i].v1);
spring->restlen = len_v3v3(cloth->verts[spring->kl].xrest, cloth->verts[spring->ij].xrest);
clmd->sim_parms->avg_spring_len += spring->restlen;
cloth->verts[spring->ij].avg_spring_len += spring->restlen;
cloth->verts[spring->kl].avg_spring_len += spring->restlen;
cloth->verts[spring->ij].spring_count++;
cloth->verts[spring->kl].spring_count++;
spring->type = CLOTH_SPRING_TYPE_STRUCTURAL;
spring->flags = 0;
spring->stiffness = (cloth->verts[spring->kl].struct_stiff + cloth->verts[spring->ij].struct_stiff) / 2.0f;
struct_springs++;
BLI_linklist_prepend ( &cloth->springs, spring );
}
else {
cloth_free_errorsprings(cloth, edgelist);
return 0;
}
}
if (struct_springs > 0)
clmd->sim_parms->avg_spring_len /= struct_springs;
for (i = 0; i < numverts; i++) {
cloth->verts[i].avg_spring_len = cloth->verts[i].avg_spring_len * 0.49f / ((float)cloth->verts[i].spring_count);
}
// shear springs
for ( i = 0; i < numfaces; i++ ) {
// triangle faces already have shear springs due to structural geometry
if ( !mface[i].v4 )
continue;
spring = (ClothSpring *)MEM_callocN(sizeof(ClothSpring), "cloth spring");
if (!spring) {
cloth_free_errorsprings(cloth, edgelist);
return 0;
}
spring->ij = MIN2(mface[i].v1, mface[i].v3);
spring->kl = MAX2(mface[i].v3, mface[i].v1);
spring->restlen = len_v3v3(cloth->verts[spring->kl].xrest, cloth->verts[spring->ij].xrest);
spring->type = CLOTH_SPRING_TYPE_SHEAR;
spring->stiffness = (cloth->verts[spring->kl].shear_stiff + cloth->verts[spring->ij].shear_stiff) / 2.0f;
BLI_linklist_append ( &edgelist[spring->ij], spring );
BLI_linklist_append ( &edgelist[spring->kl], spring );
shear_springs++;
BLI_linklist_prepend ( &cloth->springs, spring );
// if ( mface[i].v4 ) --> Quad face
spring = (ClothSpring *)MEM_callocN ( sizeof ( ClothSpring ), "cloth spring" );
if (!spring) {
cloth_free_errorsprings(cloth, edgelist);
return 0;
}
spring->ij = MIN2(mface[i].v2, mface[i].v4);
spring->kl = MAX2(mface[i].v4, mface[i].v2);
spring->restlen = len_v3v3(cloth->verts[spring->kl].xrest, cloth->verts[spring->ij].xrest);
spring->type = CLOTH_SPRING_TYPE_SHEAR;
spring->stiffness = (cloth->verts[spring->kl].shear_stiff + cloth->verts[spring->ij].shear_stiff) / 2.0f;
BLI_linklist_append ( &edgelist[spring->ij], spring );
BLI_linklist_append ( &edgelist[spring->kl], spring );
shear_springs++;
BLI_linklist_prepend ( &cloth->springs, spring );
}
edgehash = BLI_edgehash_new_ex(__func__, numedges);
cloth->edgehash = edgehash;
if (numfaces) {
// bending springs
search2 = cloth->springs;
for ( i = struct_springs; i < struct_springs+shear_springs; i++ ) {
if ( !search2 )
break;
tspring2 = search2->link;
search = edgelist[tspring2->kl];
while ( search ) {
tspring = search->link;
index2 = ( ( tspring->ij==tspring2->kl ) ? ( tspring->kl ) : ( tspring->ij ) );
// check for existing spring
// check also if startpoint is equal to endpoint
if ((index2 != tspring2->ij) &&
!BLI_edgehash_haskey(edgehash, MIN2(tspring2->ij, index2), MAX2(tspring2->ij, index2)))
{
spring = (ClothSpring *)MEM_callocN ( sizeof ( ClothSpring ), "cloth spring" );
if (!spring) {
cloth_free_errorsprings(cloth, edgelist);
return 0;
}
spring->ij = MIN2(tspring2->ij, index2);
spring->kl = MAX2(tspring2->ij, index2);
spring->restlen = len_v3v3(cloth->verts[spring->kl].xrest, cloth->verts[spring->ij].xrest);
spring->type = CLOTH_SPRING_TYPE_BENDING;
spring->stiffness = (cloth->verts[spring->kl].bend_stiff + cloth->verts[spring->ij].bend_stiff) / 2.0f;
BLI_edgehash_insert(edgehash, spring->ij, spring->kl, NULL);
bend_springs++;
BLI_linklist_prepend ( &cloth->springs, spring );
}
search = search->next;
}
search2 = search2->next;
}
}
else if (struct_springs > 2) {
/* bending springs for hair strands */
/* The current algorightm only goes through the edges in order of the mesh edges list */
/* and makes springs between the outer vert of edges sharing a vertice. This works just */
/* fine for hair, but not for user generated string meshes. This could/should be later */
/* extended to work with non-ordered edges so that it can be used for general "rope */
/* dynamics" without the need for the vertices or edges to be ordered through the length*/
/* of the strands. -jahka */
search = cloth->springs;
search2 = search->next;
while (search && search2) {
tspring = search->link;
tspring2 = search2->link;
if (tspring->ij == tspring2->kl) {
spring = (ClothSpring *)MEM_callocN ( sizeof ( ClothSpring ), "cloth spring" );
if (!spring) {
cloth_free_errorsprings(cloth, edgelist);
return 0;
}
spring->ij = tspring2->ij;
spring->kl = tspring->kl;
spring->restlen = len_v3v3(cloth->verts[spring->kl].xrest, cloth->verts[spring->ij].xrest);
spring->type = CLOTH_SPRING_TYPE_BENDING;
spring->stiffness = (cloth->verts[spring->kl].bend_stiff + cloth->verts[spring->ij].bend_stiff) / 2.0f;
bend_springs++;
BLI_linklist_prepend ( &cloth->springs, spring );
}
search = search->next;
search2 = search2->next;
}
}
/* insert other near springs in edgehash AFTER bending springs are calculated (for selfcolls) */
for (i = 0; i < numedges; i++) { /* struct springs */
BLI_edgehash_insert(edgehash, MIN2(medge[i].v1, medge[i].v2), MAX2(medge[i].v2, medge[i].v1), NULL);
}
for (i = 0; i < numfaces; i++) { /* edge springs */
if (mface[i].v4) {
BLI_edgehash_insert(edgehash, MIN2(mface[i].v1, mface[i].v3), MAX2(mface[i].v3, mface[i].v1), NULL);
BLI_edgehash_insert(edgehash, MIN2(mface[i].v2, mface[i].v4), MAX2(mface[i].v2, mface[i].v4), NULL);
}
}
cloth->numsprings = struct_springs + shear_springs + bend_springs;
cloth_free_edgelist(edgelist, numverts);
#if 0
if (G.debug_value > 0)
printf("avg_len: %f\n", clmd->sim_parms->avg_spring_len);
#endif
return 1;
} /* cloth_build_springs */
/***************************************************************************************
* SPRING NETWORK BUILDING IMPLEMENTATION END
***************************************************************************************/