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

bmo_fill_grid.c « operators « bmesh « blender « source - git.blender.org/blender.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
blob: 6734cc60cad117ad1fbaf6ca45879e8087d991d9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
/*
 * 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.
 */

/** \file
 * \ingroup bmesh
 *
 * Fill 2 isolated, open edge loops with a grid of quads.
 */

#include "MEM_guardedalloc.h"

#include "BLI_listbase.h"
#include "BLI_math.h"

#include "BKE_customdata.h"

#include "bmesh.h"

#include "intern/bmesh_operators_private.h" /* own include */

#include "BLI_strict_flags.h"

#define EDGE_MARK 4
#define FACE_OUT 16

#define BARYCENTRIC_INTERP

#ifdef BARYCENTRIC_INTERP
/**
 * 2 edge vectors to normal.
 */
static void quad_edges_to_normal(float no[3],
                                 const float co_a1[3],
                                 const float co_a2[3],
                                 const float co_b1[3],
                                 const float co_b2[3])
{
  float diff_a[3];
  float diff_b[3];

  sub_v3_v3v3(diff_a, co_a2, co_a1);
  sub_v3_v3v3(diff_b, co_b2, co_b1);
  normalize_v3(diff_a);
  normalize_v3(diff_b);
  add_v3_v3v3(no, diff_a, diff_b);
  normalize_v3(no);
}

static void quad_verts_to_barycentric_tri(float tri[3][3],
                                          const float co_a[3],
                                          const float co_b[3],

                                          const float co_a_next[3],
                                          const float co_b_next[3],

                                          const float co_a_prev[3],
                                          const float co_b_prev[3],
                                          const bool is_flip)
{
  float no[3];

  copy_v3_v3(tri[0], co_a);
  copy_v3_v3(tri[1], co_b);

  quad_edges_to_normal(no, co_a, co_a_next, co_b, co_b_next);

  if (co_a_prev) {
    float no_t[3];
    quad_edges_to_normal(no_t, co_a_prev, co_a, co_b_prev, co_b);
    add_v3_v3(no, no_t);
    normalize_v3(no);
  }

  if (is_flip) {
    negate_v3(no);
  }
  mul_v3_fl(no, len_v3v3(tri[0], tri[1]));

  mid_v3_v3v3(tri[2], tri[0], tri[1]);
  add_v3_v3(tri[2], no);
}

#endif

/* -------------------------------------------------------------------- */
/** \name Handle Loop Pairs
 * \{ */

/**
 * Assign a loop pair from 2 verts (which _must_ share an edge)
 */
static void bm_loop_pair_from_verts(BMVert *v_a, BMVert *v_b, BMLoop *l_pair[2])
{
  BMEdge *e = BM_edge_exists(v_a, v_b);
  if (e->l) {
    if (e->l->v == v_a) {
      l_pair[0] = e->l;
      l_pair[1] = e->l->next;
    }
    else {
      l_pair[0] = e->l->next;
      l_pair[1] = e->l;
    }
  }
  else {
    l_pair[0] = NULL;
    l_pair[1] = NULL;
  }
}

/**
 * Copy loop pair from one side to the other if either is missing,
 * this simplifies interpolation code so we only need to check if x/y are missing,
 * rather than checking each loop.
 */
static void bm_loop_pair_test_copy(BMLoop *l_pair_a[2], BMLoop *l_pair_b[2])
{
  /* if the first one is set, we know the second is too */
  if (l_pair_a[0] && l_pair_b[0] == NULL) {
    l_pair_b[0] = l_pair_a[1];
    l_pair_b[1] = l_pair_a[0];
  }
  else if (l_pair_b[0] && l_pair_a[0] == NULL) {
    l_pair_a[0] = l_pair_b[1];
    l_pair_a[1] = l_pair_b[0];
  }
}

/**
 * Interpolate from boundary loops.
 *
 * \note These weights will be calculated multiple times per vertex.
 */
static void bm_loop_interp_from_grid_boundary_4(BMesh *bm,
                                                BMLoop *l,
                                                BMLoop *l_bound[4],
                                                const float w[4])
{
  const void *l_cdata[4] = {
      l_bound[0]->head.data, l_bound[1]->head.data, l_bound[2]->head.data, l_bound[3]->head.data};

  CustomData_bmesh_interp(&bm->ldata, l_cdata, w, NULL, 4, l->head.data);
}

static void bm_loop_interp_from_grid_boundary_2(BMesh *bm,
                                                BMLoop *l,
                                                BMLoop *l_bound[2],
                                                const float t)
{
  const void *l_cdata[2] = {l_bound[0]->head.data, l_bound[1]->head.data};

  const float w[2] = {1.0f - t, t};

  CustomData_bmesh_interp(&bm->ldata, l_cdata, w, NULL, 2, l->head.data);
}

/** \} */

/**
 * Avoids calling #barycentric_weights_v2_quad often by caching weights into an array.
 */
static void barycentric_weights_v2_grid_cache(const uint xtot,
                                              const uint ytot,
                                              float (*weight_table)[4])
{
  float x_step = 1.0f / (float)(xtot - 1);
  float y_step = 1.0f / (float)(ytot - 1);
  uint i = 0;
  float xy_fl[2];

  uint x, y;
  for (y = 0; y < ytot; y++) {
    xy_fl[1] = y_step * (float)y;
    for (x = 0; x < xtot; x++) {
      xy_fl[0] = x_step * (float)x;
      {
        const float cos[4][2] = {
            {xy_fl[0], 0.0f}, {0.0f, xy_fl[1]}, {xy_fl[0], 1.0f}, {1.0f, xy_fl[1]}};
        barycentric_weights_v2_quad(UNPACK4(cos), xy_fl, weight_table[i++]);
      }
    }
  }
}

/**
 * This may be useful outside the bmesh operator.
 *
 * \param v_grid: 2d array of verts, all boundary verts must be set, we fill in the middle.
 */
static void bm_grid_fill_array(BMesh *bm,
                               BMVert **v_grid,
                               const uint xtot,
                               const uint ytot,
                               const short mat_nr,
                               const bool use_smooth,
                               const bool use_flip,
                               const bool use_interp_simple)
{
  const bool use_vert_interp = CustomData_has_interp(&bm->vdata);
  const bool use_loop_interp = CustomData_has_interp(&bm->ldata);
  uint x, y;

  /* for use_loop_interp */
  BMLoop *((*larr_x_a)[2]), *((*larr_x_b)[2]), *((*larr_y_a)[2]), *((*larr_y_b)[2]);

  float(*weight_table)[4];

#define XY(_x, _y) ((_x) + ((_y) * (xtot)))

#ifdef BARYCENTRIC_INTERP
  float tri_a[3][3];
  float tri_b[3][3];
  float tri_t[3][3]; /* temp */

  quad_verts_to_barycentric_tri(tri_a,
                                v_grid[XY(0, 0)]->co,
                                v_grid[XY(xtot - 1, 0)]->co,
                                v_grid[XY(0, 1)]->co,
                                v_grid[XY(xtot - 1, 1)]->co,
                                NULL,
                                NULL,
                                false);

  quad_verts_to_barycentric_tri(tri_b,
                                v_grid[XY(0, (ytot - 1))]->co,
                                v_grid[XY(xtot - 1, (ytot - 1))]->co,
                                v_grid[XY(0, (ytot - 2))]->co,
                                v_grid[XY(xtot - 1, (ytot - 2))]->co,
                                NULL,
                                NULL,
                                true);
#endif

  if (use_interp_simple || use_vert_interp || use_loop_interp) {
    weight_table = MEM_mallocN(sizeof(*weight_table) * (size_t)(xtot * ytot), __func__);
    barycentric_weights_v2_grid_cache(xtot, ytot, weight_table);
  }
  else {
    weight_table = NULL;
  }

  /* Store loops */
  if (use_loop_interp) {
    /* x2 because each edge connects 2 loops */
    larr_x_a = MEM_mallocN(sizeof(*larr_x_a) * (xtot - 1), __func__);
    larr_x_b = MEM_mallocN(sizeof(*larr_x_b) * (xtot - 1), __func__);

    larr_y_a = MEM_mallocN(sizeof(*larr_y_a) * (ytot - 1), __func__);
    larr_y_b = MEM_mallocN(sizeof(*larr_y_b) * (ytot - 1), __func__);

    /* fill in the loops */
    for (x = 0; x < xtot - 1; x++) {
      bm_loop_pair_from_verts(v_grid[XY(x, 0)], v_grid[XY(x + 1, 0)], larr_x_a[x]);
      bm_loop_pair_from_verts(v_grid[XY(x, ytot - 1)], v_grid[XY(x + 1, ytot - 1)], larr_x_b[x]);
      bm_loop_pair_test_copy(larr_x_a[x], larr_x_b[x]);
    }

    for (y = 0; y < ytot - 1; y++) {
      bm_loop_pair_from_verts(v_grid[XY(0, y)], v_grid[XY(0, y + 1)], larr_y_a[y]);
      bm_loop_pair_from_verts(v_grid[XY(xtot - 1, y)], v_grid[XY(xtot - 1, y + 1)], larr_y_b[y]);
      bm_loop_pair_test_copy(larr_y_a[y], larr_y_b[y]);
    }
  }

  /* Build Verts */
  for (y = 1; y < ytot - 1; y++) {
#ifdef BARYCENTRIC_INTERP
    quad_verts_to_barycentric_tri(tri_t,
                                  v_grid[XY(0, y + 0)]->co,
                                  v_grid[XY(xtot - 1, y + 0)]->co,
                                  v_grid[XY(0, y + 1)]->co,
                                  v_grid[XY(xtot - 1, y + 1)]->co,
                                  v_grid[XY(0, y - 1)]->co,
                                  v_grid[XY(xtot - 1, y - 1)]->co,
                                  false);
#endif
    for (x = 1; x < xtot - 1; x++) {
      float co[3];
      BMVert *v;
      /* we may want to allow sparse filled arrays, but for now, ensure its empty */
      BLI_assert(v_grid[(y * xtot) + x] == NULL);

      /* place the vertex */
#ifdef BARYCENTRIC_INTERP
      if (use_interp_simple == false) {
        float co_a[3], co_b[3];

        transform_point_by_tri_v3(
            co_a, v_grid[x]->co, tri_t[0], tri_t[1], tri_t[2], tri_a[0], tri_a[1], tri_a[2]);
        transform_point_by_tri_v3(co_b,
                                  v_grid[(xtot * ytot) + (x - xtot)]->co,
                                  tri_t[0],
                                  tri_t[1],
                                  tri_t[2],
                                  tri_b[0],
                                  tri_b[1],
                                  tri_b[2]);

        interp_v3_v3v3(co, co_a, co_b, (float)y / ((float)ytot - 1));
      }
      else
#endif
      {
        const float *w = weight_table[XY(x, y)];

        zero_v3(co);
        madd_v3_v3fl(co, v_grid[XY(x, 0)]->co, w[0]);
        madd_v3_v3fl(co, v_grid[XY(0, y)]->co, w[1]);
        madd_v3_v3fl(co, v_grid[XY(x, ytot - 1)]->co, w[2]);
        madd_v3_v3fl(co, v_grid[XY(xtot - 1, y)]->co, w[3]);
      }

      v = BM_vert_create(bm, co, NULL, BM_CREATE_NOP);
      v_grid[(y * xtot) + x] = v;

      /* Interpolate only along one axis, this could be changed
       * but from user POV gives predictable results since these are selected loop. */
      if (use_vert_interp) {
        const float *w = weight_table[XY(x, y)];

        const void *v_cdata[4] = {
            v_grid[XY(x, 0)]->head.data,
            v_grid[XY(0, y)]->head.data,
            v_grid[XY(x, ytot - 1)]->head.data,
            v_grid[XY(xtot - 1, y)]->head.data,
        };

        CustomData_bmesh_interp(&bm->vdata, v_cdata, w, NULL, 4, v->head.data);
      }
    }
  }

  /* Build Faces */
  for (x = 0; x < xtot - 1; x++) {
    for (y = 0; y < ytot - 1; y++) {
      BMFace *f;

      if (use_flip) {
        f = BM_face_create_quad_tri(bm,
                                    v_grid[XY(x, y + 0)],     /* BL */
                                    v_grid[XY(x, y + 1)],     /* TL */
                                    v_grid[XY(x + 1, y + 1)], /* TR */
                                    v_grid[XY(x + 1, y + 0)], /* BR */
                                    NULL,
                                    BM_CREATE_NOP);
      }
      else {
        f = BM_face_create_quad_tri(bm,
                                    v_grid[XY(x + 1, y + 0)], /* BR */
                                    v_grid[XY(x + 1, y + 1)], /* TR */
                                    v_grid[XY(x, y + 1)],     /* TL */
                                    v_grid[XY(x, y + 0)],     /* BL */
                                    NULL,
                                    BM_CREATE_NOP);
      }

      if (use_loop_interp && (larr_x_a[x][0] || larr_y_a[y][0])) {
        /* bottom/left/top/right */
        BMLoop *l_quad[4];
        BMLoop *l_bound[4];
        BMLoop *l_tmp;
        uint x_side, y_side, i;
        char interp_from;

        if (larr_x_a[x][0] && larr_y_a[y][0]) {
          interp_from = 'B'; /* B == both */
          l_tmp = larr_x_a[x][0];
        }
        else if (larr_x_a[x][0]) {
          interp_from = 'X';
          l_tmp = larr_x_a[x][0];
        }
        else {
          interp_from = 'Y';
          l_tmp = larr_y_a[y][0];
        }

        BM_elem_attrs_copy(bm, bm, l_tmp->f, f);

        BM_face_as_array_loop_quad(f, l_quad);

        l_tmp = BM_FACE_FIRST_LOOP(f);

        if (use_flip) {
          l_quad[0] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[1] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[3] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[2] = l_tmp;
        }
        else {
          l_quad[2] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[3] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[1] = l_tmp;
          l_tmp = l_tmp->next;
          l_quad[0] = l_tmp;
        }

        i = 0;

        for (x_side = 0; x_side < 2; x_side++) {
          for (y_side = 0; y_side < 2; y_side++) {
            if (interp_from == 'B') {
              const float *w = weight_table[XY(x + x_side, y + y_side)];
              l_bound[0] = larr_x_a[x][x_side]; /* B */
              l_bound[1] = larr_y_a[y][y_side]; /* L */
              l_bound[2] = larr_x_b[x][x_side]; /* T */
              l_bound[3] = larr_y_b[y][y_side]; /* R */

              bm_loop_interp_from_grid_boundary_4(bm, l_quad[i++], l_bound, w);
            }
            else if (interp_from == 'X') {
              const float t = (float)(y + y_side) / (float)(ytot - 1);
              l_bound[0] = larr_x_a[x][x_side]; /* B */
              l_bound[1] = larr_x_b[x][x_side]; /* T */

              bm_loop_interp_from_grid_boundary_2(bm, l_quad[i++], l_bound, t);
            }
            else if (interp_from == 'Y') {
              const float t = (float)(x + x_side) / (float)(xtot - 1);
              l_bound[0] = larr_y_a[y][y_side]; /* L */
              l_bound[1] = larr_y_b[y][y_side]; /* R */

              bm_loop_interp_from_grid_boundary_2(bm, l_quad[i++], l_bound, t);
            }
            else {
              BLI_assert(0);
            }
          }
        }
      }
      /* end interp */

      BMO_face_flag_enable(bm, f, FACE_OUT);
      f->mat_nr = mat_nr;
      if (use_smooth) {
        BM_elem_flag_enable(f, BM_ELEM_SMOOTH);
      }
    }
  }

  if (use_loop_interp) {
    MEM_freeN(larr_x_a);
    MEM_freeN(larr_y_a);
    MEM_freeN(larr_x_b);
    MEM_freeN(larr_y_b);
  }

  if (weight_table) {
    MEM_freeN(weight_table);
  }

#undef XY
}

static void bm_grid_fill(BMesh *bm,
                         struct BMEdgeLoopStore *estore_a,
                         struct BMEdgeLoopStore *estore_b,
                         struct BMEdgeLoopStore *estore_rail_a,
                         struct BMEdgeLoopStore *estore_rail_b,
                         const short mat_nr,
                         const bool use_smooth,
                         const bool use_interp_simple)
{
#define USE_FLIP_DETECT

  const uint xtot = (uint)BM_edgeloop_length_get(estore_a);
  const uint ytot = (uint)BM_edgeloop_length_get(estore_rail_a);
  // BMVert *v;
  uint i;
#ifdef DEBUG
  uint x, y;
#endif
  LinkData *el;
  bool use_flip = false;

  ListBase *lb_a = BM_edgeloop_verts_get(estore_a);
  ListBase *lb_b = BM_edgeloop_verts_get(estore_b);

  ListBase *lb_rail_a = BM_edgeloop_verts_get(estore_rail_a);
  ListBase *lb_rail_b = BM_edgeloop_verts_get(estore_rail_b);

  BMVert **v_grid = MEM_callocN(sizeof(BMVert *) * (size_t)(xtot * ytot), __func__);
  /**
   * <pre>
   *           estore_b
   *          +------------------+
   *       ^  |                  |
   *   end |  |                  |
   *       |  |                  |
   *       |  |estore_rail_a     |estore_rail_b
   *       |  |                  |
   * start |  |                  |
   *          |estore_a          |
   *          +------------------+
   *                --->
   *             start -> end
   * </pre>
   */

  BLI_assert(((LinkData *)lb_a->first)->data == ((LinkData *)lb_rail_a->first)->data); /* BL */
  BLI_assert(((LinkData *)lb_b->first)->data == ((LinkData *)lb_rail_a->last)->data);  /* TL */
  BLI_assert(((LinkData *)lb_b->last)->data == ((LinkData *)lb_rail_b->last)->data);   /* TR */
  BLI_assert(((LinkData *)lb_a->last)->data == ((LinkData *)lb_rail_b->first)->data);  /* BR */

  for (el = lb_a->first, i = 0; el; el = el->next, i++) {
    v_grid[i] = el->data;
  }
  for (el = lb_b->first, i = 0; el; el = el->next, i++) {
    v_grid[(ytot * xtot) + (i - xtot)] = el->data;
  }
  for (el = lb_rail_a->first, i = 0; el; el = el->next, i++) {
    v_grid[xtot * i] = el->data;
  }
  for (el = lb_rail_b->first, i = 0; el; el = el->next, i++) {
    v_grid[(xtot * i) + (xtot - 1)] = el->data;
  }
#ifdef DEBUG
  for (x = 1; x < xtot - 1; x++) {
    for (y = 1; y < ytot - 1; y++) {
      BLI_assert(v_grid[(y * xtot) + x] == NULL);
    }
  }
#endif

#ifdef USE_FLIP_DETECT
  {
    ListBase *lb_iter[4] = {lb_a, lb_b, lb_rail_a, lb_rail_b};
    const int lb_iter_dir[4] = {-1, 1, 1, -1};
    int winding_votes = 0;

    for (i = 0; i < 4; i++) {
      LinkData *el_next;
      for (el = lb_iter[i]->first; el && (el_next = el->next); el = el->next) {
        BMEdge *e = BM_edge_exists(el->data, el_next->data);
        if (BM_edge_is_boundary(e)) {
          winding_votes += (e->l->v == el->data) ? lb_iter_dir[i] : -lb_iter_dir[i];
        }
      }
    }
    use_flip = (winding_votes < 0);
  }
#endif

  bm_grid_fill_array(bm, v_grid, xtot, ytot, mat_nr, use_smooth, use_flip, use_interp_simple);
  MEM_freeN(v_grid);

#undef USE_FLIP_DETECT
}

static void bm_edgeloop_flag_set(struct BMEdgeLoopStore *estore, char hflag, bool set)
{
  /* only handle closed loops in this case */
  LinkData *link = BM_edgeloop_verts_get(estore)->first;
  link = link->next;
  while (link) {
    BMEdge *e = BM_edge_exists(link->data, link->prev->data);
    if (e) {
      BM_elem_flag_set(e, hflag, set);
    }
    link = link->next;
  }
}

static bool bm_edge_test_cb(BMEdge *e, void *bm_v)
{
  return BMO_edge_flag_test_bool((BMesh *)bm_v, e, EDGE_MARK);
}

static bool bm_edge_test_rail_cb(BMEdge *e, void *UNUSED(bm_v))
{
  /* Normally operators don't check for hidden state
   * but alternative would be to pass slot of rail edges. */
  if (BM_elem_flag_test(e, BM_ELEM_HIDDEN)) {
    return false;
  }
  return BM_edge_is_wire(e) || BM_edge_is_boundary(e);
}

void bmo_grid_fill_exec(BMesh *bm, BMOperator *op)
{
  ListBase eloops = {NULL, NULL};
  ListBase eloops_rail = {NULL, NULL};
  struct BMEdgeLoopStore *estore_a, *estore_b;
  struct BMEdgeLoopStore *estore_rail_a, *estore_rail_b;
  BMVert *v_a_first, *v_a_last;
  BMVert *v_b_first, *v_b_last;
  const short mat_nr = (short)BMO_slot_int_get(op->slots_in, "mat_nr");
  const bool use_smooth = BMO_slot_bool_get(op->slots_in, "use_smooth");
  const bool use_interp_simple = BMO_slot_bool_get(op->slots_in, "use_interp_simple");
  GSet *split_edges = NULL;

  int count;
  bool changed = false;
  BMO_slot_buffer_flag_enable(bm, op->slots_in, "edges", BM_EDGE, EDGE_MARK);

  count = BM_mesh_edgeloops_find(bm, &eloops, bm_edge_test_cb, (void *)bm);

  if (count != 2) {
    /* Note that this error message has been adjusted to make sense when called
     * from the operator 'MESH_OT_fill_grid' which has a 'prepare' pass which can
     * extract two 'rail' loops from a single edge loop, see T72075. */
    BMO_error_raise(bm,
                    op,
                    BMO_ERROR_CANCEL,
                    "Select two edge loops "
                    "or a single closed edge loop from which two edge loops can be calculated");
    goto cleanup;
  }

  estore_a = eloops.first;
  estore_b = eloops.last;

  v_a_first = ((LinkData *)BM_edgeloop_verts_get(estore_a)->first)->data;
  v_a_last = ((LinkData *)BM_edgeloop_verts_get(estore_a)->last)->data;
  v_b_first = ((LinkData *)BM_edgeloop_verts_get(estore_b)->first)->data;
  v_b_last = ((LinkData *)BM_edgeloop_verts_get(estore_b)->last)->data;

  if (BM_edgeloop_is_closed(estore_a) || BM_edgeloop_is_closed(estore_b)) {
    BMO_error_raise(bm, op, BMO_ERROR_CANCEL, "Closed loops unsupported");
    goto cleanup;
  }

  /* ok. all error checking done, now we can find the rail edges */

  /* cheat here, temp hide all edges so they won't be included in rails
   * this puts the mesh in an invalid state for a short time. */
  bm_edgeloop_flag_set(estore_a, BM_ELEM_HIDDEN, true);
  bm_edgeloop_flag_set(estore_b, BM_ELEM_HIDDEN, true);

  if (BM_mesh_edgeloops_find_path(
          bm, &eloops_rail, bm_edge_test_rail_cb, bm, v_a_first, v_b_first) &&
      BM_mesh_edgeloops_find_path(
          bm, &eloops_rail, bm_edge_test_rail_cb, bm, v_a_last, v_b_last)) {
    estore_rail_a = eloops_rail.first;
    estore_rail_b = eloops_rail.last;
  }
  else {
    BM_mesh_edgeloops_free(&eloops_rail);

    if (BM_mesh_edgeloops_find_path(
            bm, &eloops_rail, bm_edge_test_rail_cb, bm, v_a_first, v_b_last) &&
        BM_mesh_edgeloops_find_path(
            bm, &eloops_rail, bm_edge_test_rail_cb, bm, v_a_last, v_b_first)) {
      estore_rail_a = eloops_rail.first;
      estore_rail_b = eloops_rail.last;
      BM_edgeloop_flip(bm, estore_b);
    }
    else {
      BM_mesh_edgeloops_free(&eloops_rail);
    }
  }

  bm_edgeloop_flag_set(estore_a, BM_ELEM_HIDDEN, false);
  bm_edgeloop_flag_set(estore_b, BM_ELEM_HIDDEN, false);

  if (BLI_listbase_is_empty(&eloops_rail)) {
    BMO_error_raise(bm, op, BMO_ERROR_CANCEL, "Loops are not connected by wire/boundary edges");
    goto cleanup;
  }

  BLI_assert(estore_a != estore_b);
  BLI_assert(v_a_last != v_b_last);

  if (BM_edgeloop_overlap_check(estore_rail_a, estore_rail_b)) {
    BMO_error_raise(bm, op, BMO_ERROR_CANCEL, "Connecting edge loops overlap");
    goto cleanup;
  }

  /* add vertices if needed */
  {
    struct BMEdgeLoopStore *estore_pairs[2][2] = {
        {estore_a, estore_b},
        {estore_rail_a, estore_rail_b},
    };
    int i;

    for (i = 0; i < 2; i++) {
      const int len_a = BM_edgeloop_length_get(estore_pairs[i][0]);
      const int len_b = BM_edgeloop_length_get(estore_pairs[i][1]);
      if (len_a != len_b) {
        if (split_edges == NULL) {
          split_edges = BLI_gset_ptr_new(__func__);
        }

        if (len_a < len_b) {
          BM_edgeloop_expand(bm, estore_pairs[i][0], len_b, true, split_edges);
        }
        else {
          BM_edgeloop_expand(bm, estore_pairs[i][1], len_a, true, split_edges);
        }
      }
    }
  }

  /* finally we have all edge loops needed */
  bm_grid_fill(
      bm, estore_a, estore_b, estore_rail_a, estore_rail_b, mat_nr, use_smooth, use_interp_simple);

  changed = true;

  if (split_edges) {
    GSetIterator gs_iter;
    GSET_ITER (gs_iter, split_edges) {
      BMEdge *e = BLI_gsetIterator_getKey(&gs_iter);
      BM_edge_collapse(bm, e, e->v2, true, true);
    }
    BLI_gset_free(split_edges, NULL);
  }

cleanup:
  BM_mesh_edgeloops_free(&eloops);
  BM_mesh_edgeloops_free(&eloops_rail);

  if (changed) {
    BMO_slot_buffer_from_enabled_flag(bm, op, op->slots_out, "faces.out", BM_FACE, FACE_OUT);
  }
}