summaryrefslogtreecommitdiffstats
path: root/gfx/qcms/src/transform.rs
blob: cfca37be4c6e88a49e722d82b962ae07b6eba42a (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
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
//  qcms
//  Copyright (C) 2009 Mozilla Foundation
//  Copyright (C) 1998-2007 Marti Maria
//
// Permission is hereby granted, free of charge, to any person obtaining
// a copy of this software and associated documentation files (the "Software"),
// to deal in the Software without restriction, including without limitation
// the rights to use, copy, modify, merge, publish, distribute, sublicense,
// and/or sell copies of the Software, and to permit persons to whom the Software
// is furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
// EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
// THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
// LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
// OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
// WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

#![allow(clippy::missing_safety_doc)]
#[cfg(all(any(target_arch = "arm", target_arch = "aarch64"), feature = "neon"))]
use crate::transform_neon::{
    qcms_transform_data_bgra_out_lut_neon, qcms_transform_data_rgb_out_lut_neon,
    qcms_transform_data_rgba_out_lut_neon,
};
use crate::{
    chain::chain_transform,
    double_to_s15Fixed16Number,
    iccread::SUPPORTS_ICCV4,
    matrix::*,
    transform_util::{
        build_colorant_matrix, build_input_gamma_table, build_output_lut, compute_precache,
        lut_interp_linear,
    },
};
use crate::{
    iccread::{qcms_CIE_xyY, qcms_CIE_xyYTRIPLE, Profile, GRAY_SIGNATURE, RGB_SIGNATURE},
    transform_util::clamp_float,
    Intent,
};
#[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
use crate::{
    transform_avx::{
        qcms_transform_data_bgra_out_lut_avx, qcms_transform_data_rgb_out_lut_avx,
        qcms_transform_data_rgba_out_lut_avx,
    },
    transform_sse2::{
        qcms_transform_data_bgra_out_lut_sse2, qcms_transform_data_rgb_out_lut_sse2,
        qcms_transform_data_rgba_out_lut_sse2,
    },
};

use std::sync::atomic::Ordering;
use std::sync::Arc;
#[cfg(all(target_arch = "arm", feature = "neon"))]
use std::arch::is_arm_feature_detected;
#[cfg(all(target_arch = "aarch64", feature = "neon"))]
use std::arch::is_aarch64_feature_detected;

pub const PRECACHE_OUTPUT_SIZE: usize = 8192;
pub const PRECACHE_OUTPUT_MAX: usize = PRECACHE_OUTPUT_SIZE - 1;
pub const FLOATSCALE: f32 = PRECACHE_OUTPUT_SIZE as f32;
pub const CLAMPMAXVAL: f32 = ((PRECACHE_OUTPUT_SIZE - 1) as f32) / PRECACHE_OUTPUT_SIZE as f32;

#[repr(C)]
#[derive(Debug)]
pub struct PrecacheOuput {
    /* We previously used a count of 65536 here but that seems like more
     * precision than we actually need.  By reducing the size we can
     * improve startup performance and reduce memory usage. ColorSync on
     * 10.5 uses 4097 which is perhaps because they use a fixed point
     * representation where 1. is represented by 0x1000. */
    pub data: [u8; PRECACHE_OUTPUT_SIZE],
}

impl Default for PrecacheOuput {
    fn default() -> PrecacheOuput {
        PrecacheOuput {
            data: [0; PRECACHE_OUTPUT_SIZE],
        }
    }
}

/* used as a lookup table for the output transformation.
 * we refcount them so we only need to have one around per output
 * profile, instead of duplicating them per transform */

#[repr(C)]
#[repr(align(16))]
#[derive(Clone, Default)]
pub struct qcms_transform {
    pub matrix: [[f32; 4]; 3],
    pub input_gamma_table_r: Option<Box<[f32; 256]>>,
    pub input_gamma_table_g: Option<Box<[f32; 256]>>,
    pub input_gamma_table_b: Option<Box<[f32; 256]>>,
    pub input_clut_table_length: u16,
    pub clut: Option<Vec<f32>>,
    pub grid_size: u16,
    pub output_clut_table_length: u16,
    pub input_gamma_table_gray: Option<Box<[f32; 256]>>,
    pub out_gamma_r: f32,
    pub out_gamma_g: f32,
    pub out_gamma_b: f32,
    pub out_gamma_gray: f32,
    pub output_gamma_lut_r: Option<Vec<u16>>,
    pub output_gamma_lut_g: Option<Vec<u16>>,
    pub output_gamma_lut_b: Option<Vec<u16>>,
    pub output_gamma_lut_gray: Option<Vec<u16>>,
    pub output_gamma_lut_r_length: usize,
    pub output_gamma_lut_g_length: usize,
    pub output_gamma_lut_b_length: usize,
    pub output_gamma_lut_gray_length: usize,
    pub output_table_r: Option<Arc<PrecacheOuput>>,
    pub output_table_g: Option<Arc<PrecacheOuput>>,
    pub output_table_b: Option<Arc<PrecacheOuput>>,
    pub transform_fn: transform_fn_t,
}

pub type transform_fn_t =
    Option<unsafe fn(_: &qcms_transform, _: *const u8, _: *mut u8, _: usize) -> ()>;
/// The format of pixel data
#[repr(u32)]
#[derive(PartialEq, Eq, Clone, Copy)]
#[allow(clippy::upper_case_acronyms)]
pub enum DataType {
    RGB8 = 0,
    RGBA8 = 1,
    BGRA8 = 2,
    Gray8 = 3,
    GrayA8 = 4,
    CMYK = 5,
}

impl DataType {
    pub fn bytes_per_pixel(&self) -> usize {
        match self {
            RGB8 => 3,
            RGBA8 => 4,
            BGRA8 => 4,
            Gray8 => 1,
            GrayA8 => 2,
            CMYK => 4,
        }
    }
}

use DataType::*;

#[repr(C)]
#[derive(Copy, Clone)]
#[allow(clippy::upper_case_acronyms)]
pub struct CIE_XYZ {
    pub X: f64,
    pub Y: f64,
    pub Z: f64,
}

pub trait Format {
    const kRIndex: usize;
    const kGIndex: usize;
    const kBIndex: usize;
    const kAIndex: usize;
}

#[allow(clippy::upper_case_acronyms)]
pub struct BGRA;
impl Format for BGRA {
    const kBIndex: usize = 0;
    const kGIndex: usize = 1;
    const kRIndex: usize = 2;
    const kAIndex: usize = 3;
}

#[allow(clippy::upper_case_acronyms)]
pub struct RGBA;
impl Format for RGBA {
    const kRIndex: usize = 0;
    const kGIndex: usize = 1;
    const kBIndex: usize = 2;
    const kAIndex: usize = 3;
}

#[allow(clippy::upper_case_acronyms)]
pub struct RGB;
impl Format for RGB {
    const kRIndex: usize = 0;
    const kGIndex: usize = 1;
    const kBIndex: usize = 2;
    const kAIndex: usize = 0xFF;
}

pub trait GrayFormat {
    const has_alpha: bool;
}

pub struct Gray;
impl GrayFormat for Gray {
    const has_alpha: bool = false;
}

pub struct GrayAlpha;
impl GrayFormat for GrayAlpha {
    const has_alpha: bool = true;
}

#[inline]
fn clamp_u8(v: f32) -> u8 {
    if v > 255. {
        255
    } else if v < 0. {
        0
    } else {
        (v + 0.5).floor() as u8
    }
}

// Build a White point, primary chromas transfer matrix from RGB to CIE XYZ
// This is just an approximation, I am not handling all the non-linear
// aspects of the RGB to XYZ process, and assumming that the gamma correction
// has transitive property in the tranformation chain.
//
// the alghoritm:
//
//            - First I build the absolute conversion matrix using
//              primaries in XYZ. This matrix is next inverted
//            - Then I eval the source white point across this matrix
//              obtaining the coeficients of the transformation
//            - Then, I apply these coeficients to the original matrix
fn build_RGB_to_XYZ_transfer_matrix(
    white: qcms_CIE_xyY,
    primrs: qcms_CIE_xyYTRIPLE,
) -> Option<Matrix> {
    let mut primaries: Matrix = Matrix { m: [[0.; 3]; 3] };

    let mut result: Matrix = Matrix { m: [[0.; 3]; 3] };
    let mut white_point: Vector = Vector { v: [0.; 3] };

    let xn: f64 = white.x;
    let yn: f64 = white.y;
    if yn == 0.0f64 {
        return None;
    }

    let xr: f64 = primrs.red.x;
    let yr: f64 = primrs.red.y;
    let xg: f64 = primrs.green.x;
    let yg: f64 = primrs.green.y;
    let xb: f64 = primrs.blue.x;
    let yb: f64 = primrs.blue.y;
    primaries.m[0][0] = xr as f32;
    primaries.m[0][1] = xg as f32;
    primaries.m[0][2] = xb as f32;
    primaries.m[1][0] = yr as f32;
    primaries.m[1][1] = yg as f32;
    primaries.m[1][2] = yb as f32;
    primaries.m[2][0] = (1f64 - xr - yr) as f32;
    primaries.m[2][1] = (1f64 - xg - yg) as f32;
    primaries.m[2][2] = (1f64 - xb - yb) as f32;
    white_point.v[0] = (xn / yn) as f32;
    white_point.v[1] = 1.;
    white_point.v[2] = ((1.0f64 - xn - yn) / yn) as f32;
    let primaries_invert: Matrix = primaries.invert()?;

    let coefs: Vector = primaries_invert.eval(white_point);
    result.m[0][0] = (coefs.v[0] as f64 * xr) as f32;
    result.m[0][1] = (coefs.v[1] as f64 * xg) as f32;
    result.m[0][2] = (coefs.v[2] as f64 * xb) as f32;
    result.m[1][0] = (coefs.v[0] as f64 * yr) as f32;
    result.m[1][1] = (coefs.v[1] as f64 * yg) as f32;
    result.m[1][2] = (coefs.v[2] as f64 * yb) as f32;
    result.m[2][0] = (coefs.v[0] as f64 * (1.0f64 - xr - yr)) as f32;
    result.m[2][1] = (coefs.v[1] as f64 * (1.0f64 - xg - yg)) as f32;
    result.m[2][2] = (coefs.v[2] as f64 * (1.0f64 - xb - yb)) as f32;
    Some(result)
}
/* CIE Illuminant D50 */
const D50_XYZ: CIE_XYZ = CIE_XYZ {
    X: 0.9642f64,
    Y: 1.0000f64,
    Z: 0.8249f64,
};
/* from lcms: xyY2XYZ()
 * corresponds to argyll: icmYxy2XYZ() */
fn xyY2XYZ(source: qcms_CIE_xyY) -> CIE_XYZ {
    let mut dest: CIE_XYZ = CIE_XYZ {
        X: 0.,
        Y: 0.,
        Z: 0.,
    };
    dest.X = source.x / source.y * source.Y;
    dest.Y = source.Y;
    dest.Z = (1f64 - source.x - source.y) / source.y * source.Y;
    dest
}
/* from lcms: ComputeChromaticAdaption */
// Compute chromatic adaption matrix using chad as cone matrix
fn compute_chromatic_adaption(
    source_white_point: CIE_XYZ,
    dest_white_point: CIE_XYZ,
    chad: Matrix,
) -> Option<Matrix> {
    let mut cone_source_XYZ: Vector = Vector { v: [0.; 3] };

    let mut cone_dest_XYZ: Vector = Vector { v: [0.; 3] };

    let mut cone: Matrix = Matrix { m: [[0.; 3]; 3] };

    let chad_inv: Matrix = chad.invert()?;
    cone_source_XYZ.v[0] = source_white_point.X as f32;
    cone_source_XYZ.v[1] = source_white_point.Y as f32;
    cone_source_XYZ.v[2] = source_white_point.Z as f32;
    cone_dest_XYZ.v[0] = dest_white_point.X as f32;
    cone_dest_XYZ.v[1] = dest_white_point.Y as f32;
    cone_dest_XYZ.v[2] = dest_white_point.Z as f32;

    let cone_source_rgb: Vector = chad.eval(cone_source_XYZ);
    let cone_dest_rgb: Vector = chad.eval(cone_dest_XYZ);
    cone.m[0][0] = cone_dest_rgb.v[0] / cone_source_rgb.v[0];
    cone.m[0][1] = 0.;
    cone.m[0][2] = 0.;
    cone.m[1][0] = 0.;
    cone.m[1][1] = cone_dest_rgb.v[1] / cone_source_rgb.v[1];
    cone.m[1][2] = 0.;
    cone.m[2][0] = 0.;
    cone.m[2][1] = 0.;
    cone.m[2][2] = cone_dest_rgb.v[2] / cone_source_rgb.v[2];
    // Normalize
    Some(Matrix::multiply(chad_inv, Matrix::multiply(cone, chad)))
}
/* from lcms: cmsAdaptionMatrix */
// Returns the final chrmatic adaptation from illuminant FromIll to Illuminant ToIll
// Bradford is assumed
fn adaption_matrix(source_illumination: CIE_XYZ, target_illumination: CIE_XYZ) -> Option<Matrix> {
    let lam_rigg: Matrix = {
        Matrix {
            m: [
                [0.8951, 0.2664, -0.1614],
                [-0.7502, 1.7135, 0.0367],
                [0.0389, -0.0685, 1.0296],
            ],
        }
    };
    compute_chromatic_adaption(source_illumination, target_illumination, lam_rigg)
}
/* from lcms: cmsAdaptMatrixToD50 */
fn adapt_matrix_to_D50(r: Option<Matrix>, source_white_pt: qcms_CIE_xyY) -> Option<Matrix> {
    if source_white_pt.y == 0.0f64 {
        return None;
    }

    let Dn: CIE_XYZ = xyY2XYZ(source_white_pt);
    let Bradford: Matrix = adaption_matrix(Dn, D50_XYZ)?;
    Some(Matrix::multiply(Bradford, r?))
}
pub(crate) fn set_rgb_colorants(
    mut profile: &mut Profile,
    white_point: qcms_CIE_xyY,
    primaries: qcms_CIE_xyYTRIPLE,
) -> bool {
    let colorants = build_RGB_to_XYZ_transfer_matrix(white_point, primaries);
    let colorants = match adapt_matrix_to_D50(colorants, white_point) {
        Some(colorants) => colorants,
        None => return false,
    };

    /* note: there's a transpose type of operation going on here */
    profile.redColorant.X = double_to_s15Fixed16Number(colorants.m[0][0] as f64);
    profile.redColorant.Y = double_to_s15Fixed16Number(colorants.m[1][0] as f64);
    profile.redColorant.Z = double_to_s15Fixed16Number(colorants.m[2][0] as f64);
    profile.greenColorant.X = double_to_s15Fixed16Number(colorants.m[0][1] as f64);
    profile.greenColorant.Y = double_to_s15Fixed16Number(colorants.m[1][1] as f64);
    profile.greenColorant.Z = double_to_s15Fixed16Number(colorants.m[2][1] as f64);
    profile.blueColorant.X = double_to_s15Fixed16Number(colorants.m[0][2] as f64);
    profile.blueColorant.Y = double_to_s15Fixed16Number(colorants.m[1][2] as f64);
    profile.blueColorant.Z = double_to_s15Fixed16Number(colorants.m[2][2] as f64);
    true
}
pub(crate) fn get_rgb_colorants(
    white_point: qcms_CIE_xyY,
    primaries: qcms_CIE_xyYTRIPLE,
) -> Option<Matrix> {
    let colorants = build_RGB_to_XYZ_transfer_matrix(white_point, primaries);
    adapt_matrix_to_D50(colorants, white_point)
}
/* Alpha is not corrected.
   A rationale for this is found in Alvy Ray's "Should Alpha Be Nonlinear If
   RGB Is?" Tech Memo 17 (December 14, 1998).
    See: ftp://ftp.alvyray.com/Acrobat/17_Nonln.pdf
*/
unsafe extern "C" fn qcms_transform_data_gray_template_lut<I: GrayFormat, F: Format>(
    transform: &qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32;
    let input_gamma_table_gray = transform.input_gamma_table_gray.as_ref().unwrap();

    let mut i: u32 = 0;
    while (i as usize) < length {
        let fresh0 = src;
        src = src.offset(1);
        let device: u8 = *fresh0;
        let mut alpha: u8 = 0xffu8;
        if I::has_alpha {
            let fresh1 = src;
            src = src.offset(1);
            alpha = *fresh1
        }
        let linear: f32 = input_gamma_table_gray[device as usize];

        let out_device_r: f32 = lut_interp_linear(
            linear as f64,
            &(*transform).output_gamma_lut_r.as_ref().unwrap(),
        );
        let out_device_g: f32 = lut_interp_linear(
            linear as f64,
            &(*transform).output_gamma_lut_g.as_ref().unwrap(),
        );
        let out_device_b: f32 = lut_interp_linear(
            linear as f64,
            &(*transform).output_gamma_lut_b.as_ref().unwrap(),
        );
        *dest.add(F::kRIndex) = clamp_u8(out_device_r * 255f32);
        *dest.add(F::kGIndex) = clamp_u8(out_device_g * 255f32);
        *dest.add(F::kBIndex) = clamp_u8(out_device_b * 255f32);
        if F::kAIndex != 0xff {
            *dest.add(F::kAIndex) = alpha
        }
        dest = dest.offset(components as isize);
        i += 1
    }
}
unsafe fn qcms_transform_data_gray_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_lut::<Gray, RGB>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_gray_rgba_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_lut::<Gray, RGBA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_gray_bgra_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_lut::<Gray, BGRA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_graya_rgba_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_lut::<GrayAlpha, RGBA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_graya_bgra_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_lut::<GrayAlpha, BGRA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_gray_template_precache<I: GrayFormat, F: Format>(
    transform: *const qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32;
    let output_table_r = ((*transform).output_table_r).as_deref().unwrap();
    let output_table_g = ((*transform).output_table_g).as_deref().unwrap();
    let output_table_b = ((*transform).output_table_b).as_deref().unwrap();

    let input_gamma_table_gray = (*transform)
        .input_gamma_table_gray
        .as_ref()
        .unwrap()
        .as_ptr();

    let mut i: u32 = 0;
    while (i as usize) < length {
        let fresh2 = src;
        src = src.offset(1);
        let device: u8 = *fresh2;
        let mut alpha: u8 = 0xffu8;
        if I::has_alpha {
            let fresh3 = src;
            src = src.offset(1);
            alpha = *fresh3
        }

        let linear: f32 = *input_gamma_table_gray.offset(device as isize);
        /* we could round here... */
        let gray: u16 = (linear * PRECACHE_OUTPUT_MAX as f32) as u16;
        *dest.add(F::kRIndex) = (output_table_r).data[gray as usize];
        *dest.add(F::kGIndex) = (output_table_g).data[gray as usize];
        *dest.add(F::kBIndex) = (output_table_b).data[gray as usize];
        if F::kAIndex != 0xff {
            *dest.add(F::kAIndex) = alpha
        }
        dest = dest.offset(components as isize);
        i += 1
    }
}
unsafe fn qcms_transform_data_gray_out_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_precache::<Gray, RGB>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_gray_rgba_out_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_precache::<Gray, RGBA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_gray_bgra_out_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_precache::<Gray, BGRA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_graya_rgba_out_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_precache::<GrayAlpha, RGBA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_graya_bgra_out_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_gray_template_precache::<GrayAlpha, BGRA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_template_lut_precache<F: Format>(
    transform: &qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32;
    let output_table_r = ((*transform).output_table_r).as_deref().unwrap();
    let output_table_g = ((*transform).output_table_g).as_deref().unwrap();
    let output_table_b = ((*transform).output_table_b).as_deref().unwrap();
    let input_gamma_table_r = (*transform).input_gamma_table_r.as_ref().unwrap().as_ptr();
    let input_gamma_table_g = (*transform).input_gamma_table_g.as_ref().unwrap().as_ptr();
    let input_gamma_table_b = (*transform).input_gamma_table_b.as_ref().unwrap().as_ptr();

    let mat = &transform.matrix;
    let mut i: u32 = 0;
    while (i as usize) < length {
        let device_r: u8 = *src.add(F::kRIndex);
        let device_g: u8 = *src.add(F::kGIndex);
        let device_b: u8 = *src.add(F::kBIndex);
        let mut alpha: u8 = 0;
        if F::kAIndex != 0xff {
            alpha = *src.add(F::kAIndex)
        }
        src = src.offset(components as isize);

        let linear_r: f32 = *input_gamma_table_r.offset(device_r as isize);
        let linear_g: f32 = *input_gamma_table_g.offset(device_g as isize);
        let linear_b: f32 = *input_gamma_table_b.offset(device_b as isize);
        let mut out_linear_r = mat[0][0] * linear_r + mat[1][0] * linear_g + mat[2][0] * linear_b;
        let mut out_linear_g = mat[0][1] * linear_r + mat[1][1] * linear_g + mat[2][1] * linear_b;
        let mut out_linear_b = mat[0][2] * linear_r + mat[1][2] * linear_g + mat[2][2] * linear_b;
        out_linear_r = clamp_float(out_linear_r);
        out_linear_g = clamp_float(out_linear_g);
        out_linear_b = clamp_float(out_linear_b);
        /* we could round here... */

        let r: u16 = (out_linear_r * PRECACHE_OUTPUT_MAX as f32) as u16;
        let g: u16 = (out_linear_g * PRECACHE_OUTPUT_MAX as f32) as u16;
        let b: u16 = (out_linear_b * PRECACHE_OUTPUT_MAX as f32) as u16;
        *dest.add(F::kRIndex) = (output_table_r).data[r as usize];
        *dest.add(F::kGIndex) = (output_table_g).data[g as usize];
        *dest.add(F::kBIndex) = (output_table_b).data[b as usize];
        if F::kAIndex != 0xff {
            *dest.add(F::kAIndex) = alpha
        }
        dest = dest.offset(components as isize);
        i += 1
    }
}
#[no_mangle]
pub unsafe fn qcms_transform_data_rgb_out_lut_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut_precache::<RGB>(transform, src, dest, length);
}
#[no_mangle]
pub unsafe fn qcms_transform_data_rgba_out_lut_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut_precache::<RGBA>(transform, src, dest, length);
}
#[no_mangle]
pub unsafe fn qcms_transform_data_bgra_out_lut_precache(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut_precache::<BGRA>(transform, src, dest, length);
}
// Not used
/*
static void qcms_transform_data_clut(const qcms_transform *transform, const unsigned char *src, unsigned char *dest, size_t length) {
    unsigned int i;
    int xy_len = 1;
    int x_len = transform->grid_size;
    int len = x_len * x_len;
    const float* r_table = transform->r_clut;
    const float* g_table = transform->g_clut;
    const float* b_table = transform->b_clut;

    for (i = 0; i < length; i++) {
        unsigned char in_r = *src++;
        unsigned char in_g = *src++;
        unsigned char in_b = *src++;
        float linear_r = in_r/255.0f, linear_g=in_g/255.0f, linear_b = in_b/255.0f;

        int x = floorf(linear_r * (transform->grid_size-1));
        int y = floorf(linear_g * (transform->grid_size-1));
        int z = floorf(linear_b * (transform->grid_size-1));
        int x_n = ceilf(linear_r * (transform->grid_size-1));
        int y_n = ceilf(linear_g * (transform->grid_size-1));
        int z_n = ceilf(linear_b * (transform->grid_size-1));
        float x_d = linear_r * (transform->grid_size-1) - x;
        float y_d = linear_g * (transform->grid_size-1) - y;
        float z_d = linear_b * (transform->grid_size-1) - z;

        float r_x1 = lerp(CLU(r_table,x,y,z), CLU(r_table,x_n,y,z), x_d);
        float r_x2 = lerp(CLU(r_table,x,y_n,z), CLU(r_table,x_n,y_n,z), x_d);
        float r_y1 = lerp(r_x1, r_x2, y_d);
        float r_x3 = lerp(CLU(r_table,x,y,z_n), CLU(r_table,x_n,y,z_n), x_d);
        float r_x4 = lerp(CLU(r_table,x,y_n,z_n), CLU(r_table,x_n,y_n,z_n), x_d);
        float r_y2 = lerp(r_x3, r_x4, y_d);
        float clut_r = lerp(r_y1, r_y2, z_d);

        float g_x1 = lerp(CLU(g_table,x,y,z), CLU(g_table,x_n,y,z), x_d);
        float g_x2 = lerp(CLU(g_table,x,y_n,z), CLU(g_table,x_n,y_n,z), x_d);
        float g_y1 = lerp(g_x1, g_x2, y_d);
        float g_x3 = lerp(CLU(g_table,x,y,z_n), CLU(g_table,x_n,y,z_n), x_d);
        float g_x4 = lerp(CLU(g_table,x,y_n,z_n), CLU(g_table,x_n,y_n,z_n), x_d);
        float g_y2 = lerp(g_x3, g_x4, y_d);
        float clut_g = lerp(g_y1, g_y2, z_d);

        float b_x1 = lerp(CLU(b_table,x,y,z), CLU(b_table,x_n,y,z), x_d);
        float b_x2 = lerp(CLU(b_table,x,y_n,z), CLU(b_table,x_n,y_n,z), x_d);
        float b_y1 = lerp(b_x1, b_x2, y_d);
        float b_x3 = lerp(CLU(b_table,x,y,z_n), CLU(b_table,x_n,y,z_n), x_d);
        float b_x4 = lerp(CLU(b_table,x,y_n,z_n), CLU(b_table,x_n,y_n,z_n), x_d);
        float b_y2 = lerp(b_x3, b_x4, y_d);
        float clut_b = lerp(b_y1, b_y2, z_d);

        *dest++ = clamp_u8(clut_r*255.0f);
        *dest++ = clamp_u8(clut_g*255.0f);
        *dest++ = clamp_u8(clut_b*255.0f);
    }
}
*/
fn int_div_ceil(value: i32, div: i32) -> i32 {
    (value + div - 1) / div
}
// Using lcms' tetra interpolation algorithm.
unsafe extern "C" fn qcms_transform_data_tetra_clut_template<F: Format>(
    transform: *const qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32;

    let xy_len: i32 = 1;
    let x_len: i32 = (*transform).grid_size as i32;
    let len: i32 = x_len * x_len;
    let table = (*transform).clut.as_ref().unwrap().as_ptr();
    let r_table: *const f32 = table;
    let g_table: *const f32 = table.offset(1);
    let b_table: *const f32 = table.offset(2);

    let mut i: u32 = 0;
    while (i as usize) < length {
        let c0_r: f32;
        let c1_r: f32;
        let c2_r: f32;
        let c3_r: f32;
        let c0_g: f32;
        let c1_g: f32;
        let c2_g: f32;
        let c3_g: f32;
        let c0_b: f32;
        let c1_b: f32;
        let c2_b: f32;
        let c3_b: f32;
        let in_r: u8 = *src.add(F::kRIndex);
        let in_g: u8 = *src.add(F::kGIndex);
        let in_b: u8 = *src.add(F::kBIndex);
        let mut in_a: u8 = 0;
        if F::kAIndex != 0xff {
            in_a = *src.add(F::kAIndex)
        }
        src = src.offset(components as isize);
        let linear_r: f32 = in_r as i32 as f32 / 255.0;
        let linear_g: f32 = in_g as i32 as f32 / 255.0;
        let linear_b: f32 = in_b as i32 as f32 / 255.0;
        let x: i32 = in_r as i32 * ((*transform).grid_size as i32 - 1) / 255;
        let y: i32 = in_g as i32 * ((*transform).grid_size as i32 - 1) / 255;
        let z: i32 = in_b as i32 * ((*transform).grid_size as i32 - 1) / 255;
        let x_n: i32 = int_div_ceil(in_r as i32 * ((*transform).grid_size as i32 - 1), 255);
        let y_n: i32 = int_div_ceil(in_g as i32 * ((*transform).grid_size as i32 - 1), 255);
        let z_n: i32 = int_div_ceil(in_b as i32 * ((*transform).grid_size as i32 - 1), 255);
        let rx: f32 = linear_r * ((*transform).grid_size as i32 - 1) as f32 - x as f32;
        let ry: f32 = linear_g * ((*transform).grid_size as i32 - 1) as f32 - y as f32;
        let rz: f32 = linear_b * ((*transform).grid_size as i32 - 1) as f32 - z as f32;
        let CLU = |table: *const f32, x, y, z| {
            *table.offset(((x * len + y * x_len + z * xy_len) * 3) as isize)
        };

        c0_r = CLU(r_table, x, y, z);
        c0_g = CLU(g_table, x, y, z);
        c0_b = CLU(b_table, x, y, z);
        if rx >= ry {
            if ry >= rz {
                //rx >= ry && ry >= rz
                c1_r = CLU(r_table, x_n, y, z) - c0_r;
                c2_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x_n, y, z);
                c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
                c1_g = CLU(g_table, x_n, y, z) - c0_g;
                c2_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x_n, y, z);
                c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
                c1_b = CLU(b_table, x_n, y, z) - c0_b;
                c2_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x_n, y, z);
                c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
            } else if rx >= rz {
                //rx >= rz && rz >= ry
                c1_r = CLU(r_table, x_n, y, z) - c0_r;
                c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
                c3_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x_n, y, z);
                c1_g = CLU(g_table, x_n, y, z) - c0_g;
                c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
                c3_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x_n, y, z);
                c1_b = CLU(b_table, x_n, y, z) - c0_b;
                c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
                c3_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x_n, y, z);
            } else {
                //rz > rx && rx >= ry
                c1_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x, y, z_n);
                c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
                c3_r = CLU(r_table, x, y, z_n) - c0_r;
                c1_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x, y, z_n);
                c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
                c3_g = CLU(g_table, x, y, z_n) - c0_g;
                c1_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x, y, z_n);
                c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
                c3_b = CLU(b_table, x, y, z_n) - c0_b;
            }
        } else if rx >= rz {
            //ry > rx && rx >= rz
            c1_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x, y_n, z);
            c2_r = CLU(r_table, x, y_n, z) - c0_r;
            c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
            c1_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x, y_n, z);
            c2_g = CLU(g_table, x, y_n, z) - c0_g;
            c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
            c1_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x, y_n, z);
            c2_b = CLU(b_table, x, y_n, z) - c0_b;
            c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
        } else if ry >= rz {
            //ry >= rz && rz > rx
            c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
            c2_r = CLU(r_table, x, y_n, z) - c0_r;
            c3_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y_n, z);
            c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
            c2_g = CLU(g_table, x, y_n, z) - c0_g;
            c3_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y_n, z);
            c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
            c2_b = CLU(b_table, x, y_n, z) - c0_b;
            c3_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y_n, z);
        } else {
            //rz > ry && ry > rx
            c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
            c2_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y, z_n);
            c3_r = CLU(r_table, x, y, z_n) - c0_r;
            c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
            c2_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y, z_n);
            c3_g = CLU(g_table, x, y, z_n) - c0_g;
            c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
            c2_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y, z_n);
            c3_b = CLU(b_table, x, y, z_n) - c0_b;
        }
        let clut_r = c0_r + c1_r * rx + c2_r * ry + c3_r * rz;
        let clut_g = c0_g + c1_g * rx + c2_g * ry + c3_g * rz;
        let clut_b = c0_b + c1_b * rx + c2_b * ry + c3_b * rz;
        *dest.add(F::kRIndex) = clamp_u8(clut_r * 255.0);
        *dest.add(F::kGIndex) = clamp_u8(clut_g * 255.0);
        *dest.add(F::kBIndex) = clamp_u8(clut_b * 255.0);
        if F::kAIndex != 0xff {
            *dest.add(F::kAIndex) = in_a
        }
        dest = dest.offset(components as isize);
        i += 1
    }
}

unsafe fn tetra(
    transform: &qcms_transform,
    table: *const f32,
    in_r: u8,
    in_g: u8,
    in_b: u8,
) -> (f32, f32, f32) {
    let r_table: *const f32 = table;
    let g_table: *const f32 = table.offset(1);
    let b_table: *const f32 = table.offset(2);
    let linear_r: f32 = in_r as i32 as f32 / 255.0;
    let linear_g: f32 = in_g as i32 as f32 / 255.0;
    let linear_b: f32 = in_b as i32 as f32 / 255.0;
    let xy_len: i32 = 1;
    let x_len: i32 = (*transform).grid_size as i32;
    let len: i32 = x_len * x_len;
    let x: i32 = in_r as i32 * ((*transform).grid_size as i32 - 1) / 255;
    let y: i32 = in_g as i32 * ((*transform).grid_size as i32 - 1) / 255;
    let z: i32 = in_b as i32 * ((*transform).grid_size as i32 - 1) / 255;
    let x_n: i32 = int_div_ceil(in_r as i32 * ((*transform).grid_size as i32 - 1), 255);
    let y_n: i32 = int_div_ceil(in_g as i32 * ((*transform).grid_size as i32 - 1), 255);
    let z_n: i32 = int_div_ceil(in_b as i32 * ((*transform).grid_size as i32 - 1), 255);
    let rx: f32 = linear_r * ((*transform).grid_size as i32 - 1) as f32 - x as f32;
    let ry: f32 = linear_g * ((*transform).grid_size as i32 - 1) as f32 - y as f32;
    let rz: f32 = linear_b * ((*transform).grid_size as i32 - 1) as f32 - z as f32;
    let CLU = |table: *const f32, x, y, z| {
        *table.offset(((x * len + y * x_len + z * xy_len) * 3) as isize)
    };
    let c0_r: f32;
    let c1_r: f32;
    let c2_r: f32;
    let c3_r: f32;
    let c0_g: f32;
    let c1_g: f32;
    let c2_g: f32;
    let c3_g: f32;
    let c0_b: f32;
    let c1_b: f32;
    let c2_b: f32;
    let c3_b: f32;
    c0_r = CLU(r_table, x, y, z);
    c0_g = CLU(g_table, x, y, z);
    c0_b = CLU(b_table, x, y, z);
    if rx >= ry {
        if ry >= rz {
            //rx >= ry && ry >= rz
            c1_r = CLU(r_table, x_n, y, z) - c0_r;
            c2_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x_n, y, z);
            c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
            c1_g = CLU(g_table, x_n, y, z) - c0_g;
            c2_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x_n, y, z);
            c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
            c1_b = CLU(b_table, x_n, y, z) - c0_b;
            c2_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x_n, y, z);
            c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
        } else if rx >= rz {
            //rx >= rz && rz >= ry
            c1_r = CLU(r_table, x_n, y, z) - c0_r;
            c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
            c3_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x_n, y, z);
            c1_g = CLU(g_table, x_n, y, z) - c0_g;
            c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
            c3_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x_n, y, z);
            c1_b = CLU(b_table, x_n, y, z) - c0_b;
            c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
            c3_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x_n, y, z);
        } else {
            //rz > rx && rx >= ry
            c1_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x, y, z_n);
            c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
            c3_r = CLU(r_table, x, y, z_n) - c0_r;
            c1_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x, y, z_n);
            c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
            c3_g = CLU(g_table, x, y, z_n) - c0_g;
            c1_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x, y, z_n);
            c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
            c3_b = CLU(b_table, x, y, z_n) - c0_b;
        }
    } else if rx >= rz {
        //ry > rx && rx >= rz
        c1_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x, y_n, z);
        c2_r = CLU(r_table, x, y_n, z) - c0_r;
        c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
        c1_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x, y_n, z);
        c2_g = CLU(g_table, x, y_n, z) - c0_g;
        c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
        c1_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x, y_n, z);
        c2_b = CLU(b_table, x, y_n, z) - c0_b;
        c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
    } else if ry >= rz {
        //ry >= rz && rz > rx
        c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
        c2_r = CLU(r_table, x, y_n, z) - c0_r;
        c3_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y_n, z);
        c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
        c2_g = CLU(g_table, x, y_n, z) - c0_g;
        c3_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y_n, z);
        c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
        c2_b = CLU(b_table, x, y_n, z) - c0_b;
        c3_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y_n, z);
    } else {
        //rz > ry && ry > rx
        c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
        c2_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y, z_n);
        c3_r = CLU(r_table, x, y, z_n) - c0_r;
        c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
        c2_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y, z_n);
        c3_g = CLU(g_table, x, y, z_n) - c0_g;
        c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
        c2_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y, z_n);
        c3_b = CLU(b_table, x, y, z_n) - c0_b;
    }
    let clut_r = c0_r + c1_r * rx + c2_r * ry + c3_r * rz;
    let clut_g = c0_g + c1_g * rx + c2_g * ry + c3_g * rz;
    let clut_b = c0_b + c1_b * rx + c2_b * ry + c3_b * rz;
    (clut_r, clut_g, clut_b)
}

#[inline]
fn lerp(a: f32, b: f32, t: f32) -> f32 {
    a * (1.0 - t) + b * t
}

// lerp between two tetrahedral interpolations
// See lcms:Eval4InputsFloat
#[allow(clippy::many_single_char_names)]
unsafe fn qcms_transform_data_tetra_clut_cmyk(
    transform: &qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let table = (*transform).clut.as_ref().unwrap().as_ptr();
    assert!(
        (*transform).clut.as_ref().unwrap().len()
            >= ((transform.grid_size as i32).pow(4) * 3) as usize
    );
    for _ in 0..length {
        let c: u8 = *src.add(0);
        let m: u8 = *src.add(1);
        let y: u8 = *src.add(2);
        let k: u8 = *src.add(3);
        src = src.offset(4);
        let linear_k: f32 = k as i32 as f32 / 255.0;
        let grid_size = (*transform).grid_size as i32;
        let w: i32 = k as i32 * ((*transform).grid_size as i32 - 1) / 255;
        let w_n: i32 = int_div_ceil(k as i32 * ((*transform).grid_size as i32 - 1), 255);
        let t: f32 = linear_k * ((*transform).grid_size as i32 - 1) as f32 - w as f32;

        let table1 = table.offset((w * grid_size * grid_size * grid_size * 3) as isize);
        let table2 = table.offset((w_n * grid_size * grid_size * grid_size * 3) as isize);

        let (r1, g1, b1) = tetra(transform, table1, c, m, y);
        let (r2, g2, b2) = tetra(transform, table2, c, m, y);
        let r = lerp(r1, r2, t);
        let g = lerp(g1, g2, t);
        let b = lerp(b1, b2, t);
        *dest.add(0) = clamp_u8(r * 255.0);
        *dest.add(1) = clamp_u8(g * 255.0);
        *dest.add(2) = clamp_u8(b * 255.0);
        dest = dest.offset(3);
    }
}

unsafe fn qcms_transform_data_tetra_clut_rgb(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_tetra_clut_template::<RGB>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_tetra_clut_rgba(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_tetra_clut_template::<RGBA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_tetra_clut_bgra(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_tetra_clut_template::<BGRA>(transform, src, dest, length);
}
unsafe fn qcms_transform_data_template_lut<F: Format>(
    transform: &qcms_transform,
    mut src: *const u8,
    mut dest: *mut u8,
    length: usize,
) {
    let components: u32 = if F::kAIndex == 0xff { 3 } else { 4 } as u32;

    let mat = &transform.matrix;
    let mut i: u32 = 0;
    let input_gamma_table_r = (*transform).input_gamma_table_r.as_ref().unwrap().as_ptr();
    let input_gamma_table_g = (*transform).input_gamma_table_g.as_ref().unwrap().as_ptr();
    let input_gamma_table_b = (*transform).input_gamma_table_b.as_ref().unwrap().as_ptr();
    while (i as usize) < length {
        let device_r: u8 = *src.add(F::kRIndex);
        let device_g: u8 = *src.add(F::kGIndex);
        let device_b: u8 = *src.add(F::kBIndex);
        let mut alpha: u8 = 0;
        if F::kAIndex != 0xff {
            alpha = *src.add(F::kAIndex)
        }
        src = src.offset(components as isize);

        let linear_r: f32 = *input_gamma_table_r.offset(device_r as isize);
        let linear_g: f32 = *input_gamma_table_g.offset(device_g as isize);
        let linear_b: f32 = *input_gamma_table_b.offset(device_b as isize);
        let mut out_linear_r = mat[0][0] * linear_r + mat[1][0] * linear_g + mat[2][0] * linear_b;
        let mut out_linear_g = mat[0][1] * linear_r + mat[1][1] * linear_g + mat[2][1] * linear_b;
        let mut out_linear_b = mat[0][2] * linear_r + mat[1][2] * linear_g + mat[2][2] * linear_b;
        out_linear_r = clamp_float(out_linear_r);
        out_linear_g = clamp_float(out_linear_g);
        out_linear_b = clamp_float(out_linear_b);

        let out_device_r: f32 = lut_interp_linear(
            out_linear_r as f64,
            &(*transform).output_gamma_lut_r.as_ref().unwrap(),
        );
        let out_device_g: f32 = lut_interp_linear(
            out_linear_g as f64,
            (*transform).output_gamma_lut_g.as_ref().unwrap(),
        );
        let out_device_b: f32 = lut_interp_linear(
            out_linear_b as f64,
            (*transform).output_gamma_lut_b.as_ref().unwrap(),
        );
        *dest.add(F::kRIndex) = clamp_u8(out_device_r * 255f32);
        *dest.add(F::kGIndex) = clamp_u8(out_device_g * 255f32);
        *dest.add(F::kBIndex) = clamp_u8(out_device_b * 255f32);
        if F::kAIndex != 0xff {
            *dest.add(F::kAIndex) = alpha
        }
        dest = dest.offset(components as isize);
        i += 1
    }
}
#[no_mangle]
pub unsafe fn qcms_transform_data_rgb_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut::<RGB>(transform, src, dest, length);
}
#[no_mangle]
pub unsafe fn qcms_transform_data_rgba_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut::<RGBA>(transform, src, dest, length);
}
#[no_mangle]
pub unsafe fn qcms_transform_data_bgra_out_lut(
    transform: &qcms_transform,
    src: *const u8,
    dest: *mut u8,
    length: usize,
) {
    qcms_transform_data_template_lut::<BGRA>(transform, src, dest, length);
}

fn precache_create() -> Arc<PrecacheOuput> {
    Arc::new(PrecacheOuput::default())
}

#[no_mangle]
pub unsafe extern "C" fn qcms_transform_release(t: *mut qcms_transform) {
    drop(Box::from_raw(t));
}

const bradford_matrix: Matrix = Matrix {
    m: [
        [0.8951, 0.2664, -0.1614],
        [-0.7502, 1.7135, 0.0367],
        [0.0389, -0.0685, 1.0296],
    ],
};

const bradford_matrix_inv: Matrix = Matrix {
    m: [
        [0.9869929, -0.1470543, 0.1599627],
        [0.4323053, 0.5183603, 0.0492912],
        [-0.0085287, 0.0400428, 0.9684867],
    ],
};

// See ICCv4 E.3
fn compute_whitepoint_adaption(X: f32, Y: f32, Z: f32) -> Matrix {
    let p: f32 = (0.96422 * bradford_matrix.m[0][0]
        + 1.000 * bradford_matrix.m[1][0]
        + 0.82521 * bradford_matrix.m[2][0])
        / (X * bradford_matrix.m[0][0] + Y * bradford_matrix.m[1][0] + Z * bradford_matrix.m[2][0]);
    let y: f32 = (0.96422 * bradford_matrix.m[0][1]
        + 1.000 * bradford_matrix.m[1][1]
        + 0.82521 * bradford_matrix.m[2][1])
        / (X * bradford_matrix.m[0][1] + Y * bradford_matrix.m[1][1] + Z * bradford_matrix.m[2][1]);
    let b: f32 = (0.96422 * bradford_matrix.m[0][2]
        + 1.000 * bradford_matrix.m[1][2]
        + 0.82521 * bradford_matrix.m[2][2])
        / (X * bradford_matrix.m[0][2] + Y * bradford_matrix.m[1][2] + Z * bradford_matrix.m[2][2]);
    let white_adaption = Matrix {
        m: [[p, 0., 0.], [0., y, 0.], [0., 0., b]],
    };
    Matrix::multiply(
        bradford_matrix_inv,
        Matrix::multiply(white_adaption, bradford_matrix),
    )
}
#[no_mangle]
pub extern "C" fn qcms_profile_precache_output_transform(mut profile: &mut Profile) {
    /* we only support precaching on rgb profiles */
    if profile.color_space != RGB_SIGNATURE {
        return;
    }
    if SUPPORTS_ICCV4.load(Ordering::Relaxed) {
        /* don't precache since we will use the B2A LUT */
        if profile.B2A0.is_some() {
            return;
        }
        /* don't precache since we will use the mBA LUT */
        if profile.mBA.is_some() {
            return;
        }
    }
    /* don't precache if we do not have the TRC curves */
    if profile.redTRC.is_none() || profile.greenTRC.is_none() || profile.blueTRC.is_none() {
        return;
    }
    if profile.output_table_r.is_none() {
        let mut output_table_r = precache_create();
        if compute_precache(
            profile.redTRC.as_deref().unwrap(),
            &mut Arc::get_mut(&mut output_table_r).unwrap().data,
        ) {
            profile.output_table_r = Some(output_table_r);
        }
    }
    if profile.output_table_g.is_none() {
        let mut output_table_g = precache_create();
        if compute_precache(
            profile.greenTRC.as_deref().unwrap(),
            &mut Arc::get_mut(&mut output_table_g).unwrap().data,
        ) {
            profile.output_table_g = Some(output_table_g);
        }
    }
    if profile.output_table_b.is_none() {
        let mut output_table_b = precache_create();
        if compute_precache(
            profile.blueTRC.as_deref().unwrap(),
            &mut Arc::get_mut(&mut output_table_b).unwrap().data,
        ) {
            profile.output_table_b = Some(output_table_b);
        }
    };
}
/* Replace the current transformation with a LUT transformation using a given number of sample points */
fn transform_precacheLUT_float(
    mut transform: Box<qcms_transform>,
    input: &Profile,
    output: &Profile,
    samples: i32,
    in_type: DataType,
) -> Option<Box<qcms_transform>> {
    /* The range between which 2 consecutive sample points can be used to interpolate */
    let lutSize: u32 = (3 * samples * samples * samples) as u32;

    let mut src = Vec::with_capacity(lutSize as usize);
    let dest = vec![0.; lutSize as usize];
    /* Prepare a list of points we want to sample */
    for x in 0..samples {
        for y in 0..samples {
            for z in 0..samples {
                src.push(x as f32 / (samples - 1) as f32);
                src.push(y as f32 / (samples - 1) as f32);
                src.push(z as f32 / (samples - 1) as f32);
            }
        }
    }
    let lut = chain_transform(input, output, src, dest, lutSize as usize);
    if let Some(lut) = lut {
        (*transform).clut = Some(lut);
        (*transform).grid_size = samples as u16;
        if in_type == RGBA8 {
            (*transform).transform_fn = Some(qcms_transform_data_tetra_clut_rgba)
        } else if in_type == BGRA8 {
            (*transform).transform_fn = Some(qcms_transform_data_tetra_clut_bgra)
        } else if in_type == RGB8 {
            (*transform).transform_fn = Some(qcms_transform_data_tetra_clut_rgb)
        }
        debug_assert!((*transform).transform_fn.is_some());
    } else {
        return None;
    }

    Some(transform)
}

fn transform_precacheLUT_cmyk_float(
    mut transform: Box<qcms_transform>,
    input: &Profile,
    output: &Profile,
    samples: i32,
    in_type: DataType,
) -> Option<Box<qcms_transform>> {
    /* The range between which 2 consecutive sample points can be used to interpolate */
    let lutSize: u32 = (4 * samples * samples * samples * samples) as u32;

    let mut src = Vec::with_capacity(lutSize as usize);
    let dest = vec![0.; lutSize as usize];
    /* Prepare a list of points we want to sample */
    for k in 0..samples {
        for c in 0..samples {
            for m in 0..samples {
                for y in 0..samples {
                    src.push(c as f32 / (samples - 1) as f32);
                    src.push(m as f32 / (samples - 1) as f32);
                    src.push(y as f32 / (samples - 1) as f32);
                    src.push(k as f32 / (samples - 1) as f32);
                }
            }
        }
    }
    let lut = chain_transform(input, output, src, dest, lutSize as usize);
    if let Some(lut) = lut {
        transform.clut = Some(lut);
        transform.grid_size = samples as u16;
        assert!(in_type == DataType::CMYK);
        transform.transform_fn = Some(qcms_transform_data_tetra_clut_cmyk)
    } else {
        return None;
    }

    Some(transform)
}

pub fn transform_create(
    input: &Profile,
    in_type: DataType,
    output: &Profile,
    out_type: DataType,
    _intent: Intent,
) -> Option<Box<qcms_transform>> {
    // Ensure the requested input and output types make sense.
    let matching_format = match (in_type, out_type) {
        (RGB8, RGB8) => true,
        (RGBA8, RGBA8) => true,
        (BGRA8, BGRA8) => true,
        (Gray8, out_type) => matches!(out_type, RGB8 | RGBA8 | BGRA8),
        (GrayA8, out_type) => matches!(out_type, RGBA8 | BGRA8),
        (CMYK, RGB8) => true,
        _ => false,
    };
    if !matching_format {
        debug_assert!(false, "input/output type");
        return None;
    }
    let mut transform: Box<qcms_transform> = Box::new(Default::default());
    let mut precache: bool = false;
    if output.output_table_r.is_some()
        && output.output_table_g.is_some()
        && output.output_table_b.is_some()
    {
        precache = true
    }
    // This precache assumes RGB_SIGNATURE (fails on GRAY_SIGNATURE, for instance)
    if SUPPORTS_ICCV4.load(Ordering::Relaxed)
        && (in_type == RGB8 || in_type == RGBA8 || in_type == BGRA8 || in_type == CMYK)
        && (input.A2B0.is_some()
            || output.B2A0.is_some()
            || input.mAB.is_some()
            || output.mAB.is_some())
    {
        if in_type == CMYK {
            return transform_precacheLUT_cmyk_float(transform, input, output, 17, in_type);
        }
        // Precache the transformation to a CLUT 33x33x33 in size.
        // 33 is used by many profiles and works well in pratice.
        // This evenly divides 256 into blocks of 8x8x8.
        // TODO For transforming small data sets of about 200x200 or less
        // precaching should be avoided.
        let result = transform_precacheLUT_float(transform, input, output, 33, in_type);
        debug_assert!(result.is_some(), "precacheLUT failed");
        return result;
    }
    if precache {
        transform.output_table_r = Some(Arc::clone(output.output_table_r.as_ref().unwrap()));
        transform.output_table_g = Some(Arc::clone(output.output_table_g.as_ref().unwrap()));
        transform.output_table_b = Some(Arc::clone(output.output_table_b.as_ref().unwrap()));
    } else {
        if output.redTRC.is_none() || output.greenTRC.is_none() || output.blueTRC.is_none() {
            return None;
        }
        transform.output_gamma_lut_r = build_output_lut(output.redTRC.as_deref().unwrap());
        transform.output_gamma_lut_g = build_output_lut(output.greenTRC.as_deref().unwrap());
        transform.output_gamma_lut_b = build_output_lut(output.blueTRC.as_deref().unwrap());

        if transform.output_gamma_lut_r.is_none()
            || transform.output_gamma_lut_g.is_none()
            || transform.output_gamma_lut_b.is_none()
        {
            return None;
        }
    }
    if input.color_space == RGB_SIGNATURE {
        if precache {
            #[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
            if is_x86_feature_detected!("avx") {
                if in_type == RGB8 {
                    transform.transform_fn = Some(qcms_transform_data_rgb_out_lut_avx)
                } else if in_type == RGBA8 {
                    transform.transform_fn = Some(qcms_transform_data_rgba_out_lut_avx)
                } else if in_type == BGRA8 {
                    transform.transform_fn = Some(qcms_transform_data_bgra_out_lut_avx)
                }
            } else if cfg!(not(miri)) && is_x86_feature_detected!("sse2") {
                if in_type == RGB8 {
                    transform.transform_fn = Some(qcms_transform_data_rgb_out_lut_sse2)
                } else if in_type == RGBA8 {
                    transform.transform_fn = Some(qcms_transform_data_rgba_out_lut_sse2)
                } else if in_type == BGRA8 {
                    transform.transform_fn = Some(qcms_transform_data_bgra_out_lut_sse2)
                }
            }

            #[cfg(all(target_arch = "arm", feature = "neon"))]
            let neon_supported = is_arm_feature_detected!("neon");
            #[cfg(all(target_arch = "aarch64", feature = "neon"))]
            let neon_supported = is_aarch64_feature_detected!("neon");

            #[cfg(all(any(target_arch = "arm", target_arch = "aarch64"), feature = "neon"))]
            if neon_supported {
                if in_type == RGB8 {
                    transform.transform_fn = Some(qcms_transform_data_rgb_out_lut_neon)
                } else if in_type == RGBA8 {
                    transform.transform_fn = Some(qcms_transform_data_rgba_out_lut_neon)
                } else if in_type == BGRA8 {
                    transform.transform_fn = Some(qcms_transform_data_bgra_out_lut_neon)
                }
            }

            if transform.transform_fn.is_none() {
                if in_type == RGB8 {
                    transform.transform_fn = Some(qcms_transform_data_rgb_out_lut_precache)
                } else if in_type == RGBA8 {
                    transform.transform_fn = Some(qcms_transform_data_rgba_out_lut_precache)
                } else if in_type == BGRA8 {
                    transform.transform_fn = Some(qcms_transform_data_bgra_out_lut_precache)
                }
            }
        } else if in_type == RGB8 {
            transform.transform_fn = Some(qcms_transform_data_rgb_out_lut)
        } else if in_type == RGBA8 {
            transform.transform_fn = Some(qcms_transform_data_rgba_out_lut)
        } else if in_type == BGRA8 {
            transform.transform_fn = Some(qcms_transform_data_bgra_out_lut)
        }
        //XXX: avoid duplicating tables if we can
        transform.input_gamma_table_r = build_input_gamma_table(input.redTRC.as_deref());
        transform.input_gamma_table_g = build_input_gamma_table(input.greenTRC.as_deref());
        transform.input_gamma_table_b = build_input_gamma_table(input.blueTRC.as_deref());
        if transform.input_gamma_table_r.is_none()
            || transform.input_gamma_table_g.is_none()
            || transform.input_gamma_table_b.is_none()
        {
            return None;
        }
        /* build combined colorant matrix */

        let in_matrix: Matrix = build_colorant_matrix(input);
        let mut out_matrix: Matrix = build_colorant_matrix(output);
        out_matrix = out_matrix.invert()?;

        let result_0: Matrix = Matrix::multiply(out_matrix, in_matrix);
        /* check for NaN values in the matrix and bail if we find any */
        let mut i: u32 = 0;
        while i < 3 {
            let mut j: u32 = 0;
            while j < 3 {
                #[allow(clippy::eq_op, clippy::float_cmp)]
                if result_0.m[i as usize][j as usize].is_nan() {
                    return None;
                }
                j += 1
            }
            i += 1
        }
        /* store the results in column major mode
         * this makes doing the multiplication with sse easier */
        transform.matrix[0][0] = result_0.m[0][0];
        transform.matrix[1][0] = result_0.m[0][1];
        transform.matrix[2][0] = result_0.m[0][2];
        transform.matrix[0][1] = result_0.m[1][0];
        transform.matrix[1][1] = result_0.m[1][1];
        transform.matrix[2][1] = result_0.m[1][2];
        transform.matrix[0][2] = result_0.m[2][0];
        transform.matrix[1][2] = result_0.m[2][1];
        transform.matrix[2][2] = result_0.m[2][2]
    } else if input.color_space == GRAY_SIGNATURE {
        transform.input_gamma_table_gray = build_input_gamma_table(input.grayTRC.as_deref());
        transform.input_gamma_table_gray.as_ref()?;
        if precache {
            if out_type == RGB8 {
                transform.transform_fn = Some(qcms_transform_data_gray_out_precache)
            } else if out_type == RGBA8 {
                if in_type == Gray8 {
                    transform.transform_fn = Some(qcms_transform_data_gray_rgba_out_precache)
                } else {
                    transform.transform_fn = Some(qcms_transform_data_graya_rgba_out_precache)
                }
            } else if out_type == BGRA8 {
                if in_type == Gray8 {
                    transform.transform_fn = Some(qcms_transform_data_gray_bgra_out_precache)
                } else {
                    transform.transform_fn = Some(qcms_transform_data_graya_bgra_out_precache)
                }
            }
        } else if out_type == RGB8 {
            transform.transform_fn = Some(qcms_transform_data_gray_out_lut)
        } else if out_type == RGBA8 {
            if in_type == Gray8 {
                transform.transform_fn = Some(qcms_transform_data_gray_rgba_out_lut)
            } else {
                transform.transform_fn = Some(qcms_transform_data_graya_rgba_out_lut)
            }
        } else if out_type == BGRA8 {
            if in_type == Gray8 {
                transform.transform_fn = Some(qcms_transform_data_gray_bgra_out_lut)
            } else {
                transform.transform_fn = Some(qcms_transform_data_graya_bgra_out_lut)
            }
        }
    } else {
        debug_assert!(false, "unexpected colorspace");
        return None;
    }
    debug_assert!(transform.transform_fn.is_some());
    Some(transform)
}
/// A transform from an input profile to an output one.
pub struct Transform {
    src_ty: DataType,
    dst_ty: DataType,
    xfm: Box<qcms_transform>,
}

impl Transform {
    /// Create a new transform from `input` to `output` for pixels of `DataType` `ty` with `intent`
    pub fn new(input: &Profile, output: &Profile, ty: DataType, intent: Intent) -> Option<Self> {
        transform_create(input, ty, output, ty, intent).map(|xfm| Transform {
            src_ty: ty,
            dst_ty: ty,
            xfm,
        })
    }

    /// Create a new transform from `input` to `output` for pixels of `DataType` `ty` with `intent`
    pub fn new_to(
        input: &Profile,
        output: &Profile,
        src_ty: DataType,
        dst_ty: DataType,
        intent: Intent,
    ) -> Option<Self> {
        transform_create(input, src_ty, output, dst_ty, intent).map(|xfm| Transform {
            src_ty,
            dst_ty,
            xfm,
        })
    }

    /// Apply the color space transform to `data`
    pub fn apply(&self, data: &mut [u8]) {
        if data.len() % self.src_ty.bytes_per_pixel() != 0 {
            panic!(
                "incomplete pixels: should be a multiple of {} got {}",
                self.src_ty.bytes_per_pixel(),
                data.len()
            )
        }
        unsafe {
            self.xfm.transform_fn.expect("non-null function pointer")(
                &*self.xfm,
                data.as_ptr(),
                data.as_mut_ptr(),
                data.len() / self.src_ty.bytes_per_pixel(),
            );
        }
    }

    /// Apply the color space transform to `data`
    pub fn convert(&self, src: &[u8], dst: &mut [u8]) {
        if src.len() % self.src_ty.bytes_per_pixel() != 0 {
            panic!(
                "incomplete pixels: should be a multiple of {} got {}",
                self.src_ty.bytes_per_pixel(),
                src.len()
            )
        }
        if dst.len() % self.dst_ty.bytes_per_pixel() != 0 {
            panic!(
                "incomplete pixels: should be a multiple of {} got {}",
                self.dst_ty.bytes_per_pixel(),
                dst.len()
            )
        }
        assert_eq!(
            src.len() / self.src_ty.bytes_per_pixel(),
            dst.len() / self.dst_ty.bytes_per_pixel()
        );
        unsafe {
            self.xfm.transform_fn.expect("non-null function pointer")(
                &*self.xfm,
                src.as_ptr(),
                dst.as_mut_ptr(),
                src.len() / self.src_ty.bytes_per_pixel(),
            );
        }
    }
}

#[no_mangle]
pub extern "C" fn qcms_enable_iccv4() {
    SUPPORTS_ICCV4.store(true, Ordering::Relaxed);
}