Welcome to mirror list, hosted at ThFree Co, Russian Federation.

git.blender.org/blender.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorCampbell Barton <ideasman42@gmail.com>2012-05-22 19:57:42 +0400
committerCampbell Barton <ideasman42@gmail.com>2012-05-22 19:57:42 +0400
commite1eeb817927a280b91a71d25b5b5166161096a90 (patch)
treee8845e315ad56b19b2c432e32b7b9415af50e130 /source/blender/modifiers
parentd227d07112b9604f232223040a5ac4f15b1723d2 (diff)
style cleanu: skin modifier
Diffstat (limited to 'source/blender/modifiers')
-rw-r--r--source/blender/modifiers/intern/MOD_skin.c278
1 files changed, 139 insertions, 139 deletions
diff --git a/source/blender/modifiers/intern/MOD_skin.c b/source/blender/modifiers/intern/MOD_skin.c
index ed227284e0a..c20ae734828 100644
--- a/source/blender/modifiers/intern/MOD_skin.c
+++ b/source/blender/modifiers/intern/MOD_skin.c
@@ -83,7 +83,7 @@
typedef struct {
float mat[3][3];
/* Vert that edge is pointing away from, no relation to
- MEdge.v1 */
+ * MEdge.v1 */
int origin;
} EMat;
@@ -130,17 +130,17 @@ typedef struct {
} SkinOutput;
static void add_poly(SkinOutput *so,
- BMVert *v1,
- BMVert *v2,
- BMVert *v3,
- BMVert *v4);
+ BMVert *v1,
+ BMVert *v2,
+ BMVert *v3,
+ BMVert *v4);
/***************************** Convex Hull ****************************/
static int is_quad_symmetric(BMVert *quad[4],
- const SkinModifierData *smd)
+ const SkinModifierData *smd)
{
- const float threshold = 0.0001;
+ const float threshold = 0.0001f;
int axis;
for (axis = 0; axis < 3; axis++) {
@@ -170,7 +170,7 @@ static int is_quad_symmetric(BMVert *quad[4],
/* Returns true if the quad crosses the plane of symmetry, false otherwise */
static int quad_crosses_symmetry_plane(BMVert *quad[4],
- const SkinModifierData *smd)
+ const SkinModifierData *smd)
{
int axis;
@@ -196,7 +196,7 @@ static int quad_crosses_symmetry_plane(BMVert *quad[4],
/* Returns true if the frame is filled by precisely two faces (and
* outputs those faces to fill_faces), otherwise returns false. */
static int skin_frame_find_contained_faces(const Frame *frame,
- BMFace *fill_faces[2])
+ BMFace *fill_faces[2])
{
BMEdge *diag;
@@ -272,10 +272,10 @@ static int build_hull(SkinOutput *so, Frame **frames, int totframe)
Frame *frame = frames[i];
if (!frame->detached &&
- (!BM_edge_exists(frame->verts[0], frame->verts[1]) ||
- !BM_edge_exists(frame->verts[1], frame->verts[2]) ||
- !BM_edge_exists(frame->verts[2], frame->verts[3]) ||
- !BM_edge_exists(frame->verts[3], frame->verts[0])))
+ (!BM_edge_exists(frame->verts[0], frame->verts[1]) ||
+ !BM_edge_exists(frame->verts[1], frame->verts[2]) ||
+ !BM_edge_exists(frame->verts[2], frame->verts[3]) ||
+ !BM_edge_exists(frame->verts[3], frame->verts[0])))
{
frame->detached = TRUE;
}
@@ -328,7 +328,7 @@ static int build_hull(SkinOutput *so, Frame **frames, int totframe)
static float frame_len(const Frame *frame)
{
return (len_v3v3(frame->co[0], frame->co[1]) +
- len_v3v3(frame->co[1], frame->co[2])) * 0.5f;
+ len_v3v3(frame->co[1], frame->co[2])) * 0.5f;
}
static void merge_frame_corners(Frame **frames, int totframe)
@@ -373,11 +373,11 @@ static void merge_frame_corners(Frame **frames, int totframe)
* 'near' is based in the frames' minimum side
* length */
dist = len_v3v3(frames[i]->co[j],
- frames[k]->co[l]);
+ frames[k]->co[l]);
if (dist < thresh) {
mid_v3_v3v3(mid,
- frames[i]->co[j],
- frames[k]->co[l]);
+ frames[i]->co[j],
+ frames[k]->co[l]);
copy_v3_v3(frames[i]->co[j], mid);
copy_v3_v3(frames[k]->co[l], mid);
@@ -396,16 +396,16 @@ static void merge_frame_corners(Frame **frames, int totframe)
}
static Frame **collect_hull_frames(int v, SkinNode *frames,
- const MeshElemMap *emap, const MEdge *medge,
- int *tothullframe)
+ const MeshElemMap *emap, const MEdge *medge,
+ int *tothullframe)
{
SkinNode *f;
Frame **hull_frames;
int nbr, i;
(*tothullframe) = emap[v].count;
- hull_frames = MEM_callocN(sizeof(Frame **) * (*tothullframe),
- "hull_from_frames.hull_frames");
+ hull_frames = MEM_callocN(sizeof(Frame * *) * (*tothullframe),
+ "hull_from_frames.hull_frames");
i = 0;
for (nbr = 0; nbr < emap[v].count; nbr++) {
const MEdge *e = &medge[emap[v].indices[nbr]];
@@ -436,8 +436,8 @@ static void node_frames_init(SkinNode *nf, int totframe)
}
static void create_frame(Frame *frame, const float co[3],
- const float radius[2],
- float mat[3][3], float offset)
+ const float radius[2],
+ float mat[3][3], float offset)
{
float rx[3], ry[3], rz[3];
int i;
@@ -468,8 +468,8 @@ static float half_v2(const float v[2])
}
static void end_node_frames(int v, SkinNode *skin_nodes, const MVert *mvert,
- const MVertSkin *nodes, const MeshElemMap *emap,
- EMat *emat)
+ const MVertSkin *nodes, const MeshElemMap *emap,
+ EMat *emat)
{
const float *rad = nodes[v].radius;
float mat[3][3];
@@ -537,8 +537,8 @@ static int connection_node_mat(float mat[3][3], int v, const MeshElemMap *emap,
}
static void connection_node_frames(int v, SkinNode *skin_nodes, const MVert *mvert,
- const MVertSkin *nodes, const MeshElemMap *emap,
- EMat *emat)
+ const MVertSkin *nodes, const MeshElemMap *emap,
+ EMat *emat)
{
const float *rad = nodes[v].radius;
float mat[3][3];
@@ -575,8 +575,8 @@ static void connection_node_frames(int v, SkinNode *skin_nodes, const MVert *mve
}
static SkinNode *build_frames(const MVert *mvert, int totvert,
- const MVertSkin *nodes, const MeshElemMap *emap,
- EMat *emat)
+ const MVertSkin *nodes, const MeshElemMap *emap,
+ EMat *emat)
{
SkinNode *skin_nodes;
int v;
@@ -628,15 +628,15 @@ static void calc_edge_mat(float mat[3][3], const float a[3], const float b[3])
}
static void build_emats_rec(int *visited_e, EMat *emat,
- const MeshElemMap *emap, const MEdge *medge,
- const MVertSkin *vs, const MVert *mvert,
- int parent_v, float parent_mat[3][3])
+ const MeshElemMap *emap, const MEdge *medge,
+ const MVertSkin *vs, const MVert *mvert,
+ int parent_v, float parent_mat[3][3])
{
float axis[3], angle;
int i, e, v, parent_is_branch;
parent_is_branch = ((emap[parent_v].count > 2) ||
- (vs[parent_v].flag & MVERT_SKIN_ROOT));
+ (vs[parent_v].flag & MVERT_SKIN_ROOT));
for (i = 0; i < emap[parent_v].count; i++) {
e = emap[parent_v].indices[i];
@@ -651,7 +651,7 @@ static void build_emats_rec(int *visited_e, EMat *emat,
/* If parent is a branch node, start a new edge chain */
if (parent_is_branch) {
calc_edge_mat(emat[e].mat, mvert[parent_v].co,
- mvert[v].co);
+ mvert[v].co);
}
else {
/* Build edge matrix guided by parent matrix */
@@ -665,12 +665,12 @@ static void build_emats_rec(int *visited_e, EMat *emat,
}
build_emats_rec(visited_e, emat, emap, medge,
- vs, mvert, v, emat[e].mat);
+ vs, mvert, v, emat[e].mat);
}
}
static EMat *build_edge_mats(MVertSkin *vs, MVert *mvert, int totvert,
- MEdge *medge, MeshElemMap *emap, int totedge)
+ MEdge *medge, MeshElemMap *emap, int totedge)
{
EMat *emat;
float mat[3][3];
@@ -685,7 +685,7 @@ static EMat *build_edge_mats(MVertSkin *vs, MVert *mvert, int totvert,
if (emap[v].count >= 1) {
const MEdge *e = &medge[emap[v].indices[0]];
calc_edge_mat(mat, mvert[v].co,
- mvert[BKE_mesh_edge_other_vert(e, v)].co);
+ mvert[BKE_mesh_edge_other_vert(e, v)].co);
build_emats_rec(visited_e, emat, emap, medge, vs, mvert, v, mat);
}
}
@@ -705,7 +705,7 @@ static EMat *build_edge_mats(MVertSkin *vs, MVert *mvert, int totvert,
* having any special cases for dealing with sharing a frame between
* two hulls.) */
static int calc_edge_subdivisions(const MVert *mvert, const MVertSkin *nodes,
- const MEdge *e, int *degree)
+ const MEdge *e, int *degree)
{
const MVertSkin *evs[2] = {&nodes[e->v1], &nodes[e->v2]};
float edge_len, avg[2];
@@ -716,7 +716,7 @@ static int calc_edge_subdivisions(const MVert *mvert, const MVertSkin *nodes,
/* If either end is a branch node marked 'loose', don't subdivide
* the edge (or subdivide just twice if both are branches) */
if ((v1_branch && (evs[0]->flag & MVERT_SKIN_LOOSE)) ||
- (v2_branch && (evs[1]->flag & MVERT_SKIN_LOOSE))) {
+ (v2_branch && (evs[1]->flag & MVERT_SKIN_LOOSE))) {
if (v1_branch && v2_branch)
return 2;
else
@@ -769,7 +769,7 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
edge_subd = MEM_callocN(sizeof(int) * totorigedge, "edge_subd");
for (i = 0, totsubd = 0; i < totorigedge; i++) {
edge_subd[i] += calc_edge_subdivisions(origvert, orignode,
- &origedge[i], degree);
+ &origedge[i], degree);
totsubd += edge_subd[i];
}
@@ -777,9 +777,9 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
/* Allocate output derivedmesh */
dm = CDDM_from_template(orig,
- totorigvert + totsubd,
- totorigedge + totsubd,
- 0, 0, 0);
+ totorigvert + totsubd,
+ totorigedge + totsubd,
+ 0, 0, 0);
outvert = dm->getVertArray(dm);
outedge = dm->getEdgeArray(dm);
@@ -788,8 +788,8 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
/* Copy original vertex data */
CustomData_copy_data(&orig->vertData,
- &dm->vertData,
- 0, 0, totorigvert);
+ &dm->vertData,
+ 0, 0, totorigvert);
/* Subdivide edges */
for (i = 0, v = totorigvert; i < totorigedge; i++) {
@@ -828,7 +828,7 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
u = e->v1;
radrat = (half_v2(outnode[e->v2].radius) /
- half_v2(outnode[e->v1].radius));
+ half_v2(outnode[e->v1].radius));
radrat = (radrat + 1) / 2;
/* Add vertices and edge segments */
@@ -838,12 +838,12 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
/* Interpolate vertex coord */
interp_v3_v3v3(outvert[v].co, outvert[e->v1].co,
- outvert[e->v2].co, t);
+ outvert[e->v2].co, t);
/* Interpolate skin radii */
interp_v3_v3v3(outnode[v].radius,
- orignode[e->v1].radius,
- orignode[e->v2].radius, t);
+ orignode[e->v1].radius,
+ orignode[e->v2].radius, t);
/* Interpolate vertex group weights */
for (k = 0; k < totvgroup; k++) {
@@ -879,10 +879,10 @@ static DerivedMesh *subdivide_base(DerivedMesh *orig)
/* Can be either quad or triangle */
static void add_poly(SkinOutput *so,
- BMVert *v1,
- BMVert *v2,
- BMVert *v3,
- BMVert *v4)
+ BMVert *v1,
+ BMVert *v2,
+ BMVert *v3,
+ BMVert *v4)
{
BMVert *verts[4] = {v1, v2, v3, v4};
BMEdge *edges[4];
@@ -911,8 +911,8 @@ static void add_poly(SkinOutput *so,
}
static void connect_frames(SkinOutput *so,
- BMVert *frame1[4],
- BMVert *frame2[4])
+ BMVert *frame1[4],
+ BMVert *frame2[4])
{
BMVert *q[4][4] = {{frame2[0], frame2[1], frame1[1], frame1[0]},
{frame2[1], frame2[2], frame1[2], frame1[1]},
@@ -924,8 +924,8 @@ static void connect_frames(SkinOutput *so,
/* Check if frame normals need swap */
sub_v3_v3v3(p, q[3][0]->co, q[0][0]->co);
normal_quad_v3(no,
- q[0][0]->co, q[0][1]->co,
- q[0][2]->co, q[0][3]->co);
+ q[0][0]->co, q[0][1]->co,
+ q[0][2]->co, q[0][3]->co);
swap = dot_v3v3(no, p) > 0;
for (i = 0; i < 4; i++) {
@@ -937,8 +937,8 @@ static void connect_frames(SkinOutput *so,
}
static void output_frames(BMesh *bm,
- SkinNode *sn,
- const MDeformVert *input_dvert)
+ SkinNode *sn,
+ const MDeformVert *input_dvert)
{
Frame *f;
int i, j;
@@ -953,8 +953,8 @@ static void output_frames(BMesh *bm,
if (input_dvert) {
MDeformVert *dv;
dv = CustomData_bmesh_get(&bm->vdata,
- v->head.data,
- CD_MDEFORMVERT);
+ v->head.data,
+ CD_MDEFORMVERT);
BLI_assert(dv->totweight == 0);
defvert_copy(dv, input_dvert);
@@ -977,16 +977,16 @@ static void calc_frame_center(float center[3], const Frame *frame)
/* Does crappy fan triangulation of poly, may not be so accurate for
* concave faces */
static int isect_ray_poly(const float ray_start[3],
- const float ray_dir[3],
- BMFace *f,
- float *r_lambda)
+ const float ray_dir[3],
+ BMFace *f,
+ float *r_lambda)
{
BMVert *v, *v_first = NULL, *v_prev = NULL;
BMIter iter;
float best_dist = FLT_MAX;
int hit = 0;
- BM_ITER_ELEM(v, &iter, f, BM_VERTS_OF_FACE) {
+ BM_ITER_ELEM (v, &iter, f, BM_VERTS_OF_FACE) {
if (!v_first)
v_first = v;
else if (v_prev != v_first) {
@@ -994,8 +994,8 @@ static int isect_ray_poly(const float ray_start[3],
int curhit;
curhit = isect_ray_tri_v3(ray_start, ray_dir,
- v_first->co, v_prev->co, v->co,
- &dist, NULL);
+ v_first->co, v_prev->co, v->co,
+ &dist, NULL);
if (curhit && dist < best_dist) {
hit = TRUE;
best_dist = dist;
@@ -1015,7 +1015,7 @@ static int isect_ray_poly(const float ray_start[3],
* The orig_verts should contain the vertices of 'f'
*/
static BMFace *collapse_face_corners(BMesh *bm, BMFace *f, int n,
- BMVert **orig_verts)
+ BMVert **orig_verts)
{
int orig_len = f->len;
@@ -1048,14 +1048,14 @@ static BMFace *collapse_face_corners(BMesh *bm, BMFace *f, int n,
/* Find the new face */
f = NULL;
- BM_ITER_ELEM(vf, &iter, v_safe, BM_FACES_OF_VERT) {
+ BM_ITER_ELEM (vf, &iter, v_safe, BM_FACES_OF_VERT) {
int wrong_face = FALSE;
for (i = 0; i < orig_len; i++) {
if (orig_verts[i] == v_merge)
orig_verts[i] = NULL;
else if (orig_verts[i] &&
- !BM_vert_in_face(vf, orig_verts[i]))
+ !BM_vert_in_face(vf, orig_verts[i]))
{
wrong_face = TRUE;
break;
@@ -1087,13 +1087,13 @@ static BMFace *skin_hole_target_face(BMesh *bm, Frame *frame)
calc_frame_center(frame_center, frame);
normal_quad_v3(frame_normal, frame->verts[3]->co,
- frame->verts[2]->co, frame->verts[1]->co,
- frame->verts[0]->co);
+ frame->verts[2]->co, frame->verts[1]->co,
+ frame->verts[0]->co);
/* Use a line intersection test and nearest center test against
* all faces */
isect_target_face = center_target_face = NULL;
- BM_ITER_MESH(f, &iter, bm, BM_FACES_OF_MESH) {
+ BM_ITER_MESH (f, &iter, bm, BM_FACES_OF_MESH) {
float dist, poly_center[3];
int hit;
@@ -1130,7 +1130,7 @@ static BMFace *skin_hole_target_face(BMesh *bm, Frame *frame)
/* Use edge-length heuristic to choose from eight possible polygon bridges */
static void skin_choose_quad_bridge_order(BMVert *a[4], BMVert *b[4],
- int best_order[4])
+ int best_order[4])
{
int orders[8][4];
float shortest_len;
@@ -1180,7 +1180,7 @@ static void skin_fix_hole_no_good_verts(BMesh *bm, Frame *frame, BMFace *split_f
/* Update split face (should only be one new face created
* during extrusion) */
split_face = NULL;
- BMO_ITER(f, &oiter, bm, &op, "faceout", BM_FACE) {
+ BMO_ITER (f, &oiter, bm, &op, "faceout", BM_FACE) {
BLI_assert(!split_face);
split_face = f;
}
@@ -1196,19 +1196,19 @@ static void skin_fix_hole_no_good_verts(BMesh *bm, Frame *frame, BMFace *split_f
BMO_mesh_flag_disable_all(bm, NULL, BM_EDGE, 1);
BMO_elem_flag_enable(bm, longest_edge, 1);
BMO_op_callf(bm, "esubd edges=%fe numcuts=%i quadcornertype=%i",
- 1, 1, SUBD_STRAIGHT_CUT);
+ 1, 1, SUBD_STRAIGHT_CUT);
}
else if (split_face->len > 4) {
/* Maintain a dynamic vert array containing the split_face's
* vertices, avoids frequent allocs in collapse_face_corners() */
if (BLI_array_count(vert_buf) < split_face->len) {
BLI_array_grow_items(vert_buf, (split_face->len -
- BLI_array_count(vert_buf)));
+ BLI_array_count(vert_buf)));
}
/* Get split face's verts */
BM_iter_as_array(bm, BM_VERTS_OF_FACE, split_face,
- (void **)verts, split_face->len);
+ (void **)verts, split_face->len);
/* Earlier edge split operations may have turned some quads
* into higher-degree faces */
@@ -1230,7 +1230,7 @@ static void skin_fix_hole_no_good_verts(BMesh *bm, Frame *frame, BMFace *split_f
BMO_op_init(bm, &op, "weldverts");
for (i = 0; i < 4; i++) {
BMO_slot_map_ptr_insert(bm, &op, "targetmap",
- verts[i], frame->verts[best_order[i]]);
+ verts[i], frame->verts[best_order[i]]);
}
BMO_op_exec(bm, &op);
BMO_op_finish(bm, &op);
@@ -1271,8 +1271,8 @@ static void quad_from_tris(BMesh *bm, BMEdge *e, BMFace *adj[2], BMVert *ndx[4])
/* Find what the second tri has that the first doesn't */
for (i = 0; i < 3; i++) {
if (tri[1][i] != tri[0][0] &&
- tri[1][i] != tri[0][1] &&
- tri[1][i] != tri[0][2])
+ tri[1][i] != tri[0][1] &&
+ tri[1][i] != tri[0][2])
{
opp = tri[1][i];
break;
@@ -1285,7 +1285,7 @@ static void quad_from_tris(BMesh *bm, BMEdge *e, BMFace *adj[2], BMVert *ndx[4])
/* When the triangle edge cuts across our quad-to-be,
* throw in the second triangle's vertex */
if ((tri[0][i] == e->v1 || tri[0][i] == e->v2) &&
- (tri[0][(i + 1) % 3] == e->v1 || tri[0][(i + 1) % 3] == e->v2))
+ (tri[0][(i + 1) % 3] == e->v1 || tri[0][(i + 1) % 3] == e->v2))
{
j++;
ndx[j] = opp;
@@ -1311,7 +1311,7 @@ static int BM_face_share_face_count(BMFace *f1, BMFace *f2)
int count = 0;
BM_ITER_ELEM (e, &iter1, f1, BM_EDGES_OF_FACE) {
- BM_ITER_ELEM(f, &iter2, e, BM_FACES_OF_EDGE) {
+ BM_ITER_ELEM (f, &iter2, e, BM_FACES_OF_EDGE) {
if (f != f1 && f != f2 && BM_face_share_edge_count(f, f2))
count++;
}
@@ -1347,8 +1347,8 @@ static void hull_merge_triangles(SkinOutput *so, const SkinModifierData *smd)
/* Calculate a score for the quad, higher score for
* triangles being closer to coplanar */
score = ((BM_face_calc_area(adj[0]) +
- BM_face_calc_area(adj[1])) *
- dot_v3v3(adj[0]->no, adj[1]->no));
+ BM_face_calc_area(adj[1])) *
+ dot_v3v3(adj[0]->no, adj[1]->no));
/* Check if quad crosses the axis of symmetry */
if (quad_crosses_symmetry_plane(quad, smd)) {
@@ -1362,7 +1362,7 @@ static void hull_merge_triangles(SkinOutput *so, const SkinModifierData *smd)
/* Don't use the quad if it's concave */
if (!is_quad_convex_v3(quad[0]->co, quad[1]->co,
- quad[2]->co, quad[3]->co))
+ quad[2]->co, quad[3]->co))
{
continue;
}
@@ -1381,8 +1381,8 @@ static void hull_merge_triangles(SkinOutput *so, const SkinModifierData *smd)
/* If both triangles still free, and if they don't already
* share a border with another face, output as a quad */
if (!BM_elem_flag_test(adj[0], BM_ELEM_TAG) &&
- !BM_elem_flag_test(adj[1], BM_ELEM_TAG) &&
- !BM_face_share_face_count(adj[0], adj[1]))
+ !BM_elem_flag_test(adj[1], BM_ELEM_TAG) &&
+ !BM_face_share_face_count(adj[0], adj[1]))
{
add_quad_from_tris(so, e, adj);
BM_elem_flag_enable(adj[0], BM_ELEM_TAG);
@@ -1398,8 +1398,8 @@ static void hull_merge_triangles(SkinOutput *so, const SkinModifierData *smd)
}
static void skin_merge_close_frame_verts(SkinNode *skin_nodes, int totvert,
- const MeshElemMap *emap,
- const MEdge *medge)
+ const MeshElemMap *emap,
+ const MEdge *medge)
{
Frame **hull_frames;
int v, tothullframe;
@@ -1408,8 +1408,8 @@ static void skin_merge_close_frame_verts(SkinNode *skin_nodes, int totvert,
/* Only check branch nodes */
if (!skin_nodes[v].totframe) {
hull_frames = collect_hull_frames(v, skin_nodes,
- emap, medge,
- &tothullframe);
+ emap, medge,
+ &tothullframe);
merge_frame_corners(hull_frames, tothullframe);
MEM_freeN(hull_frames);
}
@@ -1440,7 +1440,7 @@ static void skin_update_merged_vertices(SkinNode *skin_nodes, int totvert)
}
static void skin_fix_hull_topology(BMesh *bm, SkinNode *skin_nodes,
- int totvert)
+ int totvert)
{
int v;
@@ -1465,7 +1465,7 @@ static void skin_fix_hull_topology(BMesh *bm, SkinNode *skin_nodes,
}
static void skin_output_end_nodes(SkinOutput *so, SkinNode *skin_nodes,
- int totvert)
+ int totvert)
{
int v;
@@ -1477,38 +1477,38 @@ static void skin_output_end_nodes(SkinOutput *so, SkinNode *skin_nodes,
int i, order[4];
skin_choose_quad_bridge_order(sn->frames[0].verts,
- sn->frames[1].verts,
- order);
+ sn->frames[1].verts,
+ order);
for (i = 0; i < 4; i++)
v_order[i] = sn->frames[1].verts[order[i]];
connect_frames(so, sn->frames[0].verts, v_order);
}
else if (sn->totframe == 2) {
connect_frames(so,
- sn->frames[0].verts,
- sn->frames[1].verts);
+ sn->frames[0].verts,
+ sn->frames[1].verts);
}
if (sn->flag & CAP_START) {
add_poly(so,
- sn->frames[0].verts[3],
- sn->frames[0].verts[2],
- sn->frames[0].verts[1],
- sn->frames[0].verts[0]);
+ sn->frames[0].verts[3],
+ sn->frames[0].verts[2],
+ sn->frames[0].verts[1],
+ sn->frames[0].verts[0]);
}
if (sn->flag & CAP_END) {
add_poly(so,
- sn->frames[1].verts[3],
- sn->frames[1].verts[2],
- sn->frames[1].verts[1],
- sn->frames[1].verts[0]);
+ sn->frames[1].verts[3],
+ sn->frames[1].verts[2],
+ sn->frames[1].verts[1],
+ sn->frames[1].verts[0]);
}
}
}
static void skin_output_connections(SkinOutput *so, SkinNode *skin_nodes,
- const MEdge *medge,
- int totedge)
+ const MEdge *medge,
+ int totedge)
{
int e;
@@ -1535,15 +1535,15 @@ static void skin_output_connections(SkinOutput *so, SkinNode *skin_nodes,
}
else {
connect_frames(so,
- a->frames[0].verts,
- b->frames[0].verts);
+ a->frames[0].verts,
+ b->frames[0].verts);
}
}
}
}
static void skin_smooth_hulls(BMesh *bm, SkinNode *skin_nodes,
- int totvert, const SkinModifierData *smd)
+ int totvert, const SkinModifierData *smd)
{
BMIter iter, eiter;
BMVert *v;
@@ -1568,7 +1568,7 @@ static void skin_smooth_hulls(BMesh *bm, SkinNode *skin_nodes,
skey = CustomData_number_of_layers(&bm->vdata, CD_SHAPEKEY) - 1;
BM_ITER_MESH (v, &iter, bm, BM_VERTS_OF_MESH) {
copy_v3_v3(CustomData_bmesh_get_n(&bm->vdata, v->head.data,
- CD_SHAPEKEY, skey), v->co);
+ CD_SHAPEKEY, skey), v->co);
}
/* Smooth vertices, weight unmarked vertices more strongly (helps
@@ -1588,8 +1588,8 @@ static void skin_smooth_hulls(BMesh *bm, SkinNode *skin_nodes,
BMVert *other = BM_edge_other_vert(e, v);
add_v3_v3(avg, CustomData_bmesh_get_n(&bm->vdata,
- other->head.data,
- CD_SHAPEKEY, skey));
+ other->head.data,
+ CD_SHAPEKEY, skey));
totv++;
}
@@ -1605,8 +1605,8 @@ static void skin_smooth_hulls(BMesh *bm, SkinNode *skin_nodes,
/* Returns TRUE if all hulls are successfully built, FALSE otherwise */
static int skin_output_branch_hulls(SkinOutput *so, SkinNode *skin_nodes,
- int totvert, const MeshElemMap *emap,
- const MEdge *medge)
+ int totvert, const MeshElemMap *emap,
+ const MEdge *medge)
{
int result = TRUE, v;
@@ -1619,8 +1619,8 @@ static int skin_output_branch_hulls(SkinOutput *so, SkinNode *skin_nodes,
int tothullframe;
hull_frames = collect_hull_frames(v, skin_nodes,
- emap, medge,
- &tothullframe);
+ emap, medge,
+ &tothullframe);
if (!build_hull(so, hull_frames, tothullframe))
result = FALSE;
@@ -1632,10 +1632,10 @@ static int skin_output_branch_hulls(SkinOutput *so, SkinNode *skin_nodes,
}
static BMesh *build_skin(SkinNode *skin_nodes,
- int totvert, const MeshElemMap *emap,
- const MEdge *medge, int totedge,
- const MDeformVert *input_dvert,
- SkinModifierData *smd)
+ int totvert, const MeshElemMap *emap,
+ const MEdge *medge, int totedge,
+ const MDeformVert *input_dvert,
+ SkinModifierData *smd)
{
SkinOutput so;
int v;
@@ -1655,7 +1655,7 @@ static BMesh *build_skin(SkinNode *skin_nodes,
for (v = 0; v < totvert; ++v) {
if (skin_nodes[v].totframe)
output_frames(so.bm, &skin_nodes[v],
- input_dvert ? &input_dvert[v] : NULL);
+ input_dvert ? &input_dvert[v] : NULL);
}
/* Update vertex pointers for merged frame corners */
@@ -1671,7 +1671,7 @@ static BMesh *build_skin(SkinNode *skin_nodes,
/* Using convex hulls may not generate a nice manifold mesh. Two
* problems can occur: an input frame's edges may be inside the
* hull, and/or an input frame's vertices may be inside the hull.
- *
+ *
* General fix to produce manifold mesh: for any frame that is
* partially detached, first detach it fully, then find a suitable
* existing face to merge with. (Note that we do this after
@@ -1695,7 +1695,7 @@ static void skin_set_orig_indices(DerivedMesh *dm)
totpoly = dm->getNumPolys(dm);
orig = CustomData_add_layer(&dm->polyData, CD_ORIGINDEX,
- CD_CALLOC, 0, totpoly);
+ CD_CALLOC, 0, totpoly);
for (i = 0; i < totpoly; i++)
orig[i] = ORIGINDEX_NONE;
}
@@ -1707,7 +1707,7 @@ static void skin_set_orig_indices(DerivedMesh *dm)
* 3) Output vertices and polygons from frames, connections, and hulls
*/
static DerivedMesh *base_skin(DerivedMesh *origdm,
- SkinModifierData *smd)
+ SkinModifierData *smd)
{
BMEditMesh fake_em;
DerivedMesh *result;
@@ -1759,7 +1759,7 @@ static DerivedMesh *base_skin(DerivedMesh *origdm,
}
static DerivedMesh *final_skin(SkinModifierData *smd,
- DerivedMesh *origdm)
+ DerivedMesh *origdm)
{
DerivedMesh *dm;
@@ -1780,7 +1780,7 @@ static DerivedMesh *final_skin(SkinModifierData *smd,
static void initData(ModifierData *md)
{
- SkinModifierData *smd = (SkinModifierData*) md;
+ SkinModifierData *smd = (SkinModifierData *) md;
/* Enable in editmode by default */
md->mode |= eModifierMode_Editmode;
@@ -1792,38 +1792,38 @@ static void initData(ModifierData *md)
static void copyData(ModifierData *md, ModifierData *target)
{
- SkinModifierData *smd = (SkinModifierData*) md;
- SkinModifierData *tsmd = (SkinModifierData*) target;
+ SkinModifierData *smd = (SkinModifierData *) md;
+ SkinModifierData *tsmd = (SkinModifierData *) target;
*tsmd = *smd;
}
static DerivedMesh *applyModifierEM(ModifierData *md,
- Object *UNUSED(ob),
- BMEditMesh *UNUSED(em),
- DerivedMesh *dm)
+ Object *UNUSED(ob),
+ BMEditMesh *UNUSED(em),
+ DerivedMesh *dm)
{
DerivedMesh *result;
- if (!(result = final_skin((SkinModifierData*)md, dm)))
+ if (!(result = final_skin((SkinModifierData *)md, dm)))
return dm;
return result;
}
static DerivedMesh *applyModifier(ModifierData *md,
- Object *UNUSED(ob),
- DerivedMesh *dm,
- ModifierApplyFlag UNUSED(flag))
+ Object *UNUSED(ob),
+ DerivedMesh *dm,
+ ModifierApplyFlag UNUSED(flag))
{
DerivedMesh *result;
- if (!(result = final_skin((SkinModifierData*)md, dm)))
+ if (!(result = final_skin((SkinModifierData *)md, dm)))
return dm;
return result;
}
static CustomDataMask requiredDataMask(Object *UNUSED(ob),
- ModifierData *UNUSED(md))
+ ModifierData *UNUSED(md))
{
return CD_MASK_MVERT_SKIN | CD_MASK_MDEFORMVERT;
}