summaryrefslogtreecommitdiff
path: root/Master/texmf-dist/tex/context/base/char-ini.lua
blob: 5c4a40bad516fd5839e90ad62177118ae21e408b (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
if not modules then modules = { } end modules ['char-ini'] = {
    version   = 1.001,
    comment   = "companion to char-ini.mkiv",
    author    = "Hans Hagen, PRAGMA-ADE, Hasselt NL",
    copyright = "PRAGMA ADE / ConTeXt Development Team",
    license   = "see context related readme files"
}

tex = tex or { }
xml = xml or { }

local utf = unicode.utf8

local utfchar, utfbyte, utfvalues = utf.char, utf.byte, string.utfvalues
local concat = table.concat
local next, tonumber = next, tonumber
local texsprint, texprint = tex.sprint, tex.print
local format, lower, gsub, match, gmatch = string.format, string.lower, string.gsub, string.match, string.match, string.gmatch

local ctxcatcodes = tex.ctxcatcodes
local texcatcodes = tex.texcatcodes

--[[ldx--
<p>This module implements some methods and creates additional datastructured
from the big character table that we use for all kind of purposes:
<type>char-def.lua</type>.</p>

<p>We assume that at this point <type>characters.data</type> is already
loaded!</p>
--ldx]]--

characters      = characters      or { }
characters.data = characters.data or { }

local data = characters.data

if not characters.ranges then
    characters.ranges = { }
    for k, v in next, data do
        characters.ranges[#characters.ranges+1] = k
    end
end

storage.register("characters/ranges",characters.ranges,"characters.ranges")

local ranges = characters.ranges

setmetatable(data, {
    __index = function(t,k)
        for r=1,#ranges do
            local rr = ranges[r] -- first in range
            if k > rr and k <= data[rr].range then
                t[k] = t[rr]
                return t[k]
            end
        end
        return nil
    end
})

characters.blocks = {
    ["aegeannumbers"]                        = { 0x10100, 0x1013F, "Aegean Numbers" },
    ["alphabeticpresentationforms"]          = { 0x0FB00, 0x0FB4F, "Alphabetic Presentation Forms" },
    ["ancientgreekmusicalnotation"]          = { 0x1D200, 0x1D24F, "Ancient Greek Musical Notation" },
    ["ancientgreeknumbers"]                  = { 0x10140, 0x1018F, "Ancient Greek Numbers" },
    ["ancientsymbols"]                       = { 0x10190, 0x101CF, "Ancient Symbols" },
    ["arabic"]                               = { 0x00600, 0x006FF, "Arabic" },
    ["arabicpresentationformsa"]             = { 0x0FB50, 0x0FDFF, "Arabic Presentation Forms-A" },
    ["arabicpresentationformsb"]             = { 0x0FE70, 0x0FEFF, "Arabic Presentation Forms-B" },
    ["arabicsupplement"]                     = { 0x00750, 0x0077F, "Arabic Supplement" },
    ["armenian"]                             = { 0x00530, 0x0058F, "Armenian" },
    ["arrows"]                               = { 0x02190, 0x021FF, "Arrows" },
    ["balinese"]                             = { 0x01B00, 0x01B7F, "Balinese" },
    ["basiclatin"]                           = { 0x00000, 0x0007F, "Basic Latin" },
    ["bengali"]                              = { 0x00980, 0x009FF, "Bengali" },
    ["blockelements"]                        = { 0x02580, 0x0259F, "Block Elements" },
    ["bopomofo"]                             = { 0x03100, 0x0312F, "Bopomofo" },
    ["bopomofoextended"]                     = { 0x031A0, 0x031BF, "Bopomofo Extended" },
    ["boxdrawing"]                           = { 0x02500, 0x0257F, "Box Drawing" },
    ["braillepatterns"]                      = { 0x02800, 0x028FF, "Braille Patterns" },
    ["buginese"]                             = { 0x01A00, 0x01A1F, "Buginese" },
    ["buhid"]                                = { 0x01740, 0x0175F, "Buhid" },
    ["byzantinemusicalsymbols"]              = { 0x1D000, 0x1D0FF, "Byzantine Musical Symbols" },
    ["carian"]                               = { 0x102A0, 0x102DF, "Carian" },
    ["cham"]                                 = { 0x0AA00, 0x0AA5F, "Cham" },
    ["cherokee"]                             = { 0x013A0, 0x013FF, "Cherokee" },
    ["cjkcompatibility"]                     = { 0x03300, 0x033FF, "CJK Compatibility" },
    ["cjkcompatibilityforms"]                = { 0x0FE30, 0x0FE4F, "CJK Compatibility Forms" },
    ["cjkcompatibilityideographs"]           = { 0x0F900, 0x0FAFF, "CJK Compatibility Ideographs" },
    ["cjkcompatibilityideographssupplement"] = { 0x2F800, 0x2FA1F, "CJK Compatibility Ideographs Supplement" },
    ["cjkradicalssupplement"]                = { 0x02E80, 0x02EFF, "CJK Radicals Supplement" },
    ["cjkstrokes"]                           = { 0x031C0, 0x031EF, "CJK Strokes" },
    ["cjksymbolsandpunctuation"]             = { 0x03000, 0x0303F, "CJK Symbols and Punctuation" },
    ["cjkunifiedideographs"]                 = { 0x04E00, 0x09FFF, "CJK Unified Ideographs" },
    ["cjkunifiedideographsextensiona"]       = { 0x03400, 0x04DBF, "CJK Unified Ideographs Extension A" },
    ["cjkunifiedideographsextensionb"]       = { 0x20000, 0x2A6DF, "CJK Unified Ideographs Extension B" },
    ["combiningdiacriticalmarks"]            = { 0x00300, 0x0036F, "Combining Diacritical Marks" },
    ["combiningdiacriticalmarksforsymbols"]  = { 0x020D0, 0x020FF, "Combining Diacritical Marks for Symbols" },
    ["combiningdiacriticalmarkssupplement"]  = { 0x01DC0, 0x01DFF, "Combining Diacritical Marks Supplement" },
    ["combininghalfmarks"]                   = { 0x0FE20, 0x0FE2F, "Combining Half Marks" },
    ["controlpictures"]                      = { 0x02400, 0x0243F, "Control Pictures" },
    ["coptic"]                               = { 0x02C80, 0x02CFF, "Coptic" },
    ["countingrodnumerals"]                  = { 0x1D360, 0x1D37F, "Counting Rod Numerals" },
    ["cuneiform"]                            = { 0x12000, 0x123FF, "Cuneiform" },
    ["cuneiformnumbersandpunctuation"]       = { 0x12400, 0x1247F, "Cuneiform Numbers and Punctuation" },
    ["currencysymbols"]                      = { 0x020A0, 0x020CF, "Currency Symbols" },
    ["cypriotsyllabary"]                     = { 0x10800, 0x1083F, "Cypriot Syllabary" },
    ["cyrillic"]                             = { 0x00400, 0x004FF, "Cyrillic" },
    ["cyrillicextendeda"]                    = { 0x02DE0, 0x02DFF, "Cyrillic Extended-A" },
    ["cyrillicextendedb"]                    = { 0x0A640, 0x0A69F, "Cyrillic Extended-B" },
    ["cyrillicsupplement"]                   = { 0x00500, 0x0052F, "Cyrillic Supplement" },
    ["deseret"]                              = { 0x10400, 0x1044F, "Deseret" },
    ["devanagari"]                           = { 0x00900, 0x0097F, "Devanagari" },
    ["dingbats"]                             = { 0x02700, 0x027BF, "Dingbats" },
    ["dominotiles"]                          = { 0x1F030, 0x1F09F, "Domino Tiles" },
    ["enclosedalphanumerics"]                = { 0x02460, 0x024FF, "Enclosed Alphanumerics" },
    ["enclosedcjklettersandmonths"]          = { 0x03200, 0x032FF, "Enclosed CJK Letters and Months" },
    ["ethiopic"]                             = { 0x01200, 0x0137F, "Ethiopic" },
    ["ethiopicextended"]                     = { 0x02D80, 0x02DDF, "Ethiopic Extended" },
    ["ethiopicsupplement"]                   = { 0x01380, 0x0139F, "Ethiopic Supplement" },
    ["generalpunctuation"]                   = { 0x02000, 0x0206F, "General Punctuation" },
    ["geometricshapes"]                      = { 0x025A0, 0x025FF, "Geometric Shapes" },
    ["georgian"]                             = { 0x010A0, 0x010FF, "Georgian" },
    ["georgiansupplement"]                   = { 0x02D00, 0x02D2F, "Georgian Supplement" },
    ["glagolitic"]                           = { 0x02C00, 0x02C5F, "Glagolitic" },
    ["gothic"]                               = { 0x10330, 0x1034F, "Gothic" },
    ["greekandcoptic"]                       = { 0x00370, 0x003FF, "Greek and Coptic" },
    ["greekextended"]                        = { 0x01F00, 0x01FFF, "Greek Extended" },
    ["gujarati"]                             = { 0x00A80, 0x00AFF, "Gujarati" },
    ["gurmukhi"]                             = { 0x00A00, 0x00A7F, "Gurmukhi" },
    ["halfwidthandfullwidthforms"]           = { 0x0FF00, 0x0FFEF, "Halfwidth and Fullwidth Forms" },
    ["hangulcompatibilityjamo"]              = { 0x03130, 0x0318F, "Hangul Compatibility Jamo" },
    ["hanguljamo"]                           = { 0x01100, 0x011FF, "Hangul Jamo" },
    ["hangulsyllables"]                      = { 0x0AC00, 0x0D7AF, "Hangul Syllables" },
    ["hanunoo"]                              = { 0x01720, 0x0173F, "Hanunoo" },
    ["hebrew"]                               = { 0x00590, 0x005FF, "Hebrew" },
    ["highprivateusesurrogates"]             = { 0x0DB80, 0x0DBFF, "High Private Use Surrogates" },
    ["highsurrogates"]                       = { 0x0D800, 0x0DB7F, "High Surrogates" },
    ["hiragana"]                             = { 0x03040, 0x0309F, "Hiragana" },
    ["ideographicdescriptioncharacters"]     = { 0x02FF0, 0x02FFF, "Ideographic Description Characters" },
    ["ipaextensions"]                        = { 0x00250, 0x02AF, "IPA Extensions" },
    ["kanbun"]                               = { 0x03190, 0x0319F, "Kanbun" },
    ["kangxiradicals"]                       = { 0x02F00, 0x02FDF, "Kangxi Radicals" },
    ["kannada"]                              = { 0x00C80, 0x00CFF, "Kannada" },
    ["katakana"]                             = { 0x030A0, 0x030FF, "Katakana" },
    ["katakanaphoneticextensions"]           = { 0x031F0, 0x031FF, "Katakana Phonetic Extensions" },
    ["kayahli"]                              = { 0x0A900, 0x0A92F, "Kayah Li" },
    ["kharoshthi"]                           = { 0x10A00, 0x10A5F, "Kharoshthi" },
    ["khmer"]                                = { 0x01780, 0x017FF, "Khmer" },
    ["khmersymbols"]                         = { 0x019E0, 0x019FF, "Khmer Symbols" },
    ["lao"]                                  = { 0x00E80, 0x00EFF, "Lao" },
    ["latinextendeda"]                       = { 0x00100, 0x0017F, "Latin Extended-A" },
    ["latinextendedadditional"]              = { 0x01E00, 0x01EFF, "Latin Extended Additional" },
    ["latinextendedb"]                       = { 0x00180, 0x0024F, "Latin Extended-B" },
    ["latinextendedc"]                       = { 0x02C60, 0x02C7F, "Latin Extended-C" },
    ["latinextendedd"]                       = { 0x0A720, 0x0A7FF, "Latin Extended-D" },
    ["latinsupplement"]                      = { 0x00080, 0x000FF, "Latin-1 Supplement" },
    ["lepcha"]                               = { 0x01C00, 0x01C4F, "Lepcha" },
    ["letterlikesymbols"]                    = { 0x02100, 0x0214F, "Letterlike Symbols" },
    ["limbu"]                                = { 0x01900, 0x0194F, "Limbu" },
    ["linearbideograms"]                     = { 0x10080, 0x100FF, "Linear B Ideograms" },
    ["linearbsyllabary"]                     = { 0x10000, 0x1007F, "Linear B Syllabary" },
    ["lowsurrogates"]                        = { 0x0DC00, 0x0DFFF, "Low Surrogates" },
    ["lycian"]                               = { 0x10280, 0x1029F, "Lycian" },
    ["lydian"]                               = { 0x10920, 0x1093F, "Lydian" },
    ["mahjongtiles"]                         = { 0x1F000, 0x1F02F, "Mahjong Tiles" },
    ["malayalam"]                            = { 0x00D00, 0x00D7F, "Malayalam" },
    ["mathematicalalphanumericsymbols"]      = { 0x1D400, 0x1D7FF, "Mathematical Alphanumeric Symbols" },
    ["mathematicaloperators"]                = { 0x02200, 0x022FF, "Mathematical Operators" },
    ["miscellaneousmathematicalsymbolsa"]    = { 0x027C0, 0x027EF, "Miscellaneous Mathematical Symbols-A" },
    ["miscellaneousmathematicalsymbolsb"]    = { 0x02980, 0x029FF, "Miscellaneous Mathematical Symbols-B" },
    ["miscellaneoussymbols"]                 = { 0x02600, 0x026FF, "Miscellaneous Symbols" },
    ["miscellaneoussymbolsandarrows"]        = { 0x02B00, 0x02BFF, "Miscellaneous Symbols and Arrows" },
    ["miscellaneoustechnical"]               = { 0x02300, 0x023FF, "Miscellaneous Technical" },
    ["modifiertoneletters"]                  = { 0x0A700, 0x0A71F, "Modifier Tone Letters" },
    ["mongolian"]                            = { 0x01800, 0x018AF, "Mongolian" },
    ["musicalsymbols"]                       = { 0x1D100, 0x1D1FF, "Musical Symbols" },
    ["myanmar"]                              = { 0x01000, 0x0109F, "Myanmar" },
    ["newtailue"]                            = { 0x01980, 0x019DF, "New Tai Lue" },
    ["nko"]                                  = { 0x007C0, 0x007FF, "NKo" },
    ["numberforms"]                          = { 0x02150, 0x0218F, "Number Forms" },
    ["ogham"]                                = { 0x01680, 0x0169F, "Ogham" },
    ["olchiki"]                              = { 0x01C50, 0x01C7F, "Ol Chiki" },
    ["olditalic"]                            = { 0x10300, 0x1032F, "Old Italic" },
    ["oldpersian"]                           = { 0x103A0, 0x103DF, "Old Persian" },
    ["opticalcharacterrecognition"]          = { 0x02440, 0x0245F, "Optical Character Recognition" },
    ["oriya"]                                = { 0x00B00, 0x00B7F, "Oriya" },
    ["osmanya"]                              = { 0x10480, 0x104AF, "Osmanya" },
    ["phagspa"]                              = { 0x0A840, 0x0A87F, "Phags-pa" },
    ["phaistosdisc"]                         = { 0x101D0, 0x101FF, "Phaistos Disc" },
    ["phoenician"]                           = { 0x10900, 0x1091F, "Phoenician" },
    ["phoneticextensions"]                   = { 0x01D00, 0x01D7F, "Phonetic Extensions" },
    ["phoneticextensionssupplement"]         = { 0x01D80, 0x01DBF, "Phonetic Extensions Supplement" },
    ["privateusearea"]                       = { 0x0E000, 0x0F8FF, "Private Use Area" },
    ["rejang"]                               = { 0x0A930, 0x0A95F, "Rejang" },
    ["runic"]                                = { 0x016A0, 0x016FF, "Runic" },
    ["saurashtra"]                           = { 0x0A880, 0x0A8DF, "Saurashtra" },
    ["shavian"]                              = { 0x10450, 0x1047F, "Shavian" },
    ["sinhala"]                              = { 0x00D80, 0x00DFF, "Sinhala" },
    ["smallformvariants"]                    = { 0x0FE50, 0x0FE6F, "Small Form Variants" },
    ["spacingmodifierletters"]               = { 0x002B0, 0x002FF, "Spacing Modifier Letters" },
    ["specials"]                             = { 0x0FFF0, 0x0FFFF, "Specials" },
    ["sundanese"]                            = { 0x01B80, 0x01BBF, "Sundanese" },
    ["superscriptsandsubscripts"]            = { 0x02070, 0x0209F, "Superscripts and Subscripts" },
    ["supplementalarrowsa"]                  = { 0x027F0, 0x027FF, "Supplemental Arrows-A" },
    ["supplementalarrowsb"]                  = { 0x02900, 0x0297F, "Supplemental Arrows-B" },
    ["supplementalmathematicaloperators"]    = { 0x02A00, 0x02AFF, "Supplemental Mathematical Operators" },
    ["supplementalpunctuation"]              = { 0x02E00, 0x02E7F, "Supplemental Punctuation" },
    ["supplementaryprivateuseareaa"]         = { 0xF0000, 0xFFFFF, "Supplementary Private Use Area-A" },
    ["supplementaryprivateuseareab"]         = { 0x100000,0x10FFFF,"Supplementary Private Use Area-B" },
    ["sylotinagri"]                          = { 0x0A800, 0x0A82F, "Syloti Nagri" },
    ["syriac"]                               = { 0x00700, 0x0074F, "Syriac" },
    ["tagalog"]                              = { 0x01700, 0x0171F, "Tagalog" },
    ["tagbanwa"]                             = { 0x01760, 0x0177F, "Tagbanwa" },
    ["tags"]                                 = { 0xE0000, 0xE007F, "Tags" },
    ["taile"]                                = { 0x01950, 0x0197F, "Tai Le" },
    ["taixuanjingsymbols"]                   = { 0x1D300, 0x1D35F, "Tai Xuan Jing Symbols" },
    ["tamil"]                                = { 0x00B80, 0x00BFF, "Tamil" },
    ["telugu"]                               = { 0x00C00, 0x00C7F, "Telugu" },
    ["thaana"]                               = { 0x00780, 0x007BF, "Thaana" },
    ["thai"]                                 = { 0x00E00, 0x00E7F, "Thai" },
    ["tibetan"]                              = { 0x00F00, 0x00FFF, "Tibetan" },
    ["tifinagh"]                             = { 0x02D30, 0x02D7F, "Tifinagh" },
    ["ugaritic"]                             = { 0x10380, 0x1039F, "Ugaritic" },
    ["unifiedcanadianaboriginalsyllabics"]   = { 0x01400, 0x0167F, "Unified Canadian Aboriginal Syllabics" },
    ["vai"]                                  = { 0x0A500, 0x0A63F, "Vai" },
    ["variationselectors"]                   = { 0x0FE00, 0x0FE0F, "Variation Selectors" },
    ["variationselectorssupplement"]         = { 0xE0100, 0xE01EF, "Variation Selectors Supplement" },
    ["verticalforms"]                        = { 0x0FE10, 0x0FE1F, "Vertical Forms" },
    ["yijinghexagramsymbols"]                = { 0x04DC0, 0x04DFF, "Yijing Hexagram Symbols" },
    ["yiradicals"]                           = { 0x0A490, 0x0A4CF, "Yi Radicals" },
    ["yisyllables"]                          = { 0x0A000, 0x0A48F, "Yi Syllables" },
}

function characters.getrange(name)
    local tag = lower(name)
    tag = gsub(name,"[^a-z]", "")
    local range = characters.blocks[tag]
    if range then
        return range[1], range[2], range[3]
    end
    name = gsub(name,'"',"0x") -- goodie: tex hex notation
    local start, stop = match(name,"^(.-)[%-%:](.-)$")
    if start and stop then
        start, stop = tonumber(start,16) or tonumber(start), tonumber(stop,16) or tonumber(stop)
        if start and stop then
            return start, stop, nil
        end
    end
    local slot = tonumber(name,16) or tonumber(name)
    return slot, slot, nil
end

characters.categories = {
    lu = "Letter Uppercase",
    ll = "Letter Lowercase",
    lt = "Letter Titlecase",
    lm = "Letter Modifier",
    lo = "Letter Other",
    mn = "Mark Nonspacing",
    mc = "Mark Spacing Combining",
    me = "Mark Enclosing",
    nd = "Number Decimal Digit",
    nl = "Number Letter",
    no = "Number Other",
    pc = "Punctuation Connector",
    pd = "Punctuation Dash",
    ps = "Punctuation Open",
    pe = "Punctuation Close",
    pi = "Punctuation Initial Quote",
    pf = "Punctuation Final Quote",
    po = "Punctuation Other",
    sm = "Symbol Math",
    sc = "Symbol Currency",
    sk = "Symbol Modifier",
    so = "Symbol Other",
    zs = "Separator Space",
    zl = "Separator Line",
    zp = "Separator Paragraph",
    cc = "Other Control",
    cf = "Other Format",
    cs = "Other Surrogate",
    co = "Other Private Use",
    cn = "Other Not Assigned",
}

--~ special   : cf (softhyphen) zs (emspace)
--~ characters: ll lm lo lt lu mn nl no pc pd pe pf pi po ps sc sk sm so

characters.is_character = table.tohash {
    "lu","ll","lt","lm","lo",
    "nd","nl","no",
    "mn",
    "nl","no",
    "pc","pd","ps","pe","pi","pf","po",
    "sm","sc","sk","so"
}

characters.is_letter = table.tohash {
    "ll","lm","lo","lt","lu"
}

characters.is_command = table.tohash {
    "cf","zs"
}

-- linebreak: todo: hash
--
-- normative   : BK CR LF CM SG GL CB SP ZW NL WJ JL JV JT H2 H3
-- informative : XX OP CL QU NS EX SY IS PR PO NU AL ID IN HY BB BA SA AI B2

-- east asian width:
--
-- N A H W F Na

characters.bidi = {
    l   = "Left-to-Right",
    lre = "Left-to-Right Embedding",
    lro = "Left-to-Right Override",
    r   = "Right-to-Left",
    al  = "Right-to-Left Arabic",
    rle = "Right-to-Left Embedding",
    rlo = "Right-to-Left Override",
    pdf = "Pop Directional Format",
    en  = "European Number",
    es  = "European Number Separator",
    et  = "European Number Terminator",
    an  = "Arabic Number",
    cs  = "Common Number Separator",
    nsm = "Non-Spacing Mark",
    bn  = "Boundary Neutral",
    b   = "Paragraph Separator",
    s   = "Segment Separator",
    ws  = "Whitespace",
    on  = "Other Neutrals",
}

local _empty_table_ = { __index = function(t,k) return "" end }

function table.set_empty_metatable(t)
    setmetatable(t,_empty_table_)
end

table.set_empty_metatable(data)

--[[ldx--
<p>At this point we assume that the big data table is loaded. From this
table we derive a few more.</p>
--ldx]]--

if not characters.fallbacks then

    characters.fallbacks   = { }
    characters.directions  = { }

    local fallbacks  = characters.fallbacks
    local directions = characters.directions

    for k,v in next, data do
        local specials = v.specials
        if specials and specials[1] == "compat" and specials[2] == 0x0020 and specials[3] then
            local s = specials[3]
            fallbacks[k] = s
            fallbacks[s] = k
        end
        directions[k] = v.direction
    end

end

storage.register("characters.fallbacks", characters.fallbacks, "characters.fallbacks")
storage.register("characters.directions", characters.directions, "characters.directions")

--[[ldx--
<p>The <type>context</type> namespace is used to store methods and data
which is rather specific to <l n='context'/>.</p>
--ldx]]--

--[[ldx--
<p>Instead of using a <l n='tex'/> file to define the named glyphs, we
use the table. After all, we have this information available anyway.</p>
--ldx]]--

function characters.makeactive(n,name) -- let ?
    texsprint(ctxcatcodes,format("\\catcode%s=13\\unexpanded\\def %s{\\%s}",n,utfchar(n),name))
end

function tex.uprint(n)
    texsprint(ctxcatcodes,utfchar(n))
end

local template_a = "\\startextendcatcodetable{%s}\\chardef\\l=11\\chardef\\a=13\\let\\c\\catcode%s\\let\\a\\undefined\\let\\l\\undefined\\let\\c\\undefined\\stopextendcatcodetable"
local template_b = "\\chardef\\l=11\\chardef\\a=13\\let\\c\\catcode%s\\let\\a\\undefined\\let\\l\\undefined\\let\\c\\undefined"

-- we need a function for setting the codes ....

function characters.define(tobelettered, tobeactivated) -- catcodetables
    local is_character, is_command, is_letter = characters.is_character, characters.is_command, characters.is_letter
    local lettered, activated = { }, { }
    for u, chr in next, data do
        -- we can use a macro instead of direct settings
        local fallback = chr.fallback
        if fallback then
        --  texprint(format("{\\catcode %s=13\\unexpanded\\gdef %s{\\checkedchar{%s}{%s}}}",u,utfchar(u),u,fallback))
            texsprint("{\\catcode",u,"=13\\unexpanded\\gdef ",utfchar(u),"{\\checkedchar{",u,"}{",fallback,"}}}") -- no texprint
            activated[#activated+1] = "\\c"..u.."\\a"
        else
            local contextname = chr.contextname
            local category = chr.category
            if contextname then
                if is_character[category] then
                 -- by this time, we're still in normal catcode mode
                 -- subtle: not "\\",contextname but "\\"..contextname
                    if chr.unicodeslot < 128 then
                    --  texprint(ctxcatcodes, "\\chardef\\"..contextname,"=",u)
                        texprint(ctxcatcodes,format("\\chardef\\%s=%s",contextname,u))
                    else
                    --  texprint(ctxcatcodes, "\\let\\"..contextname,"=",utfchar(u))
                        texprint(ctxcatcodes,format("\\let\\%s=%s",contextname,utfchar(u)))
                        if is_letter[category] then
                            lettered[#lettered+1] = "\\c"..u.."\\l"
                        end
                    end
                elseif is_command[category] then
                    -- this might change: contextcommand ipv contextname
                --  texprint(format("{\\catcode %s=13\\unexpanded\\gdef %s{\\%s}}",u,utfchar(u),contextname))
                    texsprint("{\\catcode",u,"=13\\unexpanded\\gdef ",utfchar(u),"{\\"..contextname,"}}") -- no texprint
                    activated[#activated+1] = "\\c"..u.."\\a"
                end
            elseif is_letter[category] then
                if u >= 128 and u <= 65536 then -- catch private mess
                    lettered[#lettered+1] = "\\c"..u.."\\l"
                end
            end
        end
        if chr.range then
            lettered[#lettered+1] = format('\\dofastrecurse{"%05X}{"%05X}{1}{\\c\\fastrecursecounter\\l}',u,chr.range)
        end
    end
 -- if false then
    lettered[#lettered+1] = "\\c"..0x200C.."\\l" -- non-joiner
    lettered[#lettered+1] = "\\c"..0x200D.."\\l" -- joiner
 -- fi
    if tobelettered then
        lettered = concat(lettered)
        if true then
            texsprint(ctxcatcodes,format(template_b,lettered))
        else
            for l=1,#tobelettered do
                texsprint(ctxcatcodes,format(template_a,tobelettered[l],lettered))
            end
        end
    end
    if tobeactivated then
        activated = concat(activated)
        for a=1,#tobeactivated do
            texsprint(ctxcatcodes,format(template_a,tobeactivated[a],activated))
        end
    end
end

function characters.charcode(box)
    local b = tex.box[box]
    local l = b.list
    texsprint((l and l.id == node.id('glyph') and l.char) or 0)
end

--[[ldx--
<p>Setting the lccodes is also done in a loop over the data table.</p>
--ldx]]--

-- we need a function ...

function characters.setcodes()
    for code, chr in next, data do
        local cc = chr.category
        if cc == 'll' or cc == 'lu' or cc == 'lt' then
            local lc, uc = chr.lccode, chr.uccode
            if not lc then chr.lccode, lc = code, code end
            if not uc then chr.uccode, uc = code, code end
            texsprint(ctxcatcodes,format("\\setcclcuc{%i}{%i}{%i}",code,lc,uc))
        end
        if cc == "lu" then
            texprint(ctxcatcodes,"\\sfcode ",code,"999 ")
        end
        if cc == "lo" and chr.range then
            texsprint(ctxcatcodes,format('\\dofastrecurse{"%05X}{"%05X}{1}{\\setcclcucself\\fastrecursecounter}',code,chr.range))
        end
    end
end

--[[ldx--
<p>Next comes a whole series of helper methods. These are (will be) part
of the official <l n='api'/>.</p>
--ldx]]--

--[[ldx--
<p>This converts a string (if given) into a number.</p>
--ldx]]--

function characters.number(n)
    if type(n) == "string" then return tonumber(n,16) else return n end
end

--[[ldx--
<p>Checking for valid characters.</p>
--ldx]]--

function characters.is_valid(s)
    return s or ""
end

function characters.checked(s, default)
    return s or default
end

characters.valid = characters.is_valid

--[[ldx--
<p></p>
--ldx]]--
-- set a table entry; index is number (can be different from unicodeslot)

function characters.set(n, c)
    data[characters.number(n)] = c
end

--[[ldx--
<p>Get a table entry happens by number. Keep in mind that the unicodeslot
can be different (not likely).</p>
--ldx]]--

function characters.get(n)
    return data[characters.number(n)]
end

--[[ldx--
<p>A couple of convenience methods. Beware, these are not that fast due
to the checking.</p>
--ldx]]--

function characters.hexindex(n)
    return format("%04X", characters.valid(data[characters.number(n)].unicodeslot))
end

function characters.contextname(n)
    return characters.valid(data[characters.number(n)].contextname)
end

function characters.adobename(n)
    return characters.valid(data[characters.number(n)].adobename)
end

function characters.description(n)
    return characters.valid(data[characters.number(n)].description)
end

function characters.category(n)
    return characters.valid(data[characters.number(n)].category)
end

--[[ldx--
<p>Requesting lower and uppercase codes:</p>
--ldx]]--

function characters.uccode(n) return data[n].uccode or n end
function characters.lccode(n) return data[n].lccode or n end

function characters.flush(n)
    local c = data[n]
    if c and c.contextname then
        texsprint(texcatcodes, "\\"..c.contextname)
    else
        texsprint(utfchar(n))
    end
end

function characters.shape(n)
    local shcode = data[n].shcode
    if not shcode then
        return n, nil
    elseif type(shcode) == "table" then
        return shcode[1], shcode[#shcode]
    else
        return shcode, nil
    end
end

--[[ldx--
<p>Categories play an important role, so here are some checkers.</p>
--ldx]]--

function characters.is_of_category(token,category)
    if type(token) == "string" then
        return data[utfbyte(token)].category == category
    else
        return data[token].category == category
    end
end

function characters.i_is_of_category(i,category) -- by index (number)
    local cd = data[i]
    return cd and cd.category == category
end

function characters.n_is_of_category(n,category) -- by name (string)
    local cd = data[utfbyte(n)]
    return cd and cd.category == category
end

-- xml support (moved)

function characters.remapentity(chr,slot)
    texsprint(format("{\\catcode%s=13\\xdef%s{\\string%s}}",slot,utfchar(slot),chr))
end

characters.active_offset = 0x10000 -- there will be remapped in that byte range

-- xml.entities = xml.entities or { }
--
-- storage.register("xml/entities",xml.entities,"xml.entities") -- this will move to lxml
--
-- function characters.setmkiventities()
--     local entities = xml.entities
--     entities.lt  = "<"
--     entities.amp = "&"
--     entities.gt  = ">"
-- end
--
-- function characters.setmkiientities()
--     local entities = xml.entities
--     entities.lt  = utfchar(characters.active_offset + utfbyte("<"))
--     entities.amp = utfchar(characters.active_offset + utfbyte("&"))
--     entities.gt  = utfchar(characters.active_offset + utfbyte(">"))
-- end

-- some day we will make a table

function characters.lower(str)
    local new = { }
    for u in utfvalues(str) do
        new[#new+1] = utfchar(data[u].lccode or u)
    end
    return concat(new)
end

function characters.upper(str)
    local new = { }
    for u in utfvalues(str) do
        new[#new+1] = utfchar(data[u].uccode or u)
    end
    return concat(new)
end

-- -- some day we might go this route, but it does not really save that much
-- -- so not now (we can generate a lot using mtx-unicode that operates on the
-- -- database)
--
-- -- category cjkwd direction linebreak
--
-- -- adobename comment contextcommand contextname description fallback lccode
-- -- mathclass mathfiller mathname mathspec mathstretch mathsymbol mirror
-- -- range shcode specials uccode uccodes unicodeslot
--
-- local data = {
--     ['one']={
--         common = {
--             category="cc",
--             direction="bn",
--             linebreak="cm",
--         },
--         vector = {
--             [0x0000] = {
--                 description="NULL",
--                 group='one',
--                 unicodeslot=0x0000,
--             },
--             {
--                 description="START OF HEADING",
--                 group='one',
--                 unicodeslot=0x0001,
--             },
--         }
--     }
-- }
--
-- local chardata, groupdata = { }, { }
--
-- for group, gdata in next, data do
--     local common, vector = { __index = gdata.common }, gdata.vector
--     for character, cdata in next, vector do
--         chardata[character] = cdata
--         setmetatable(cdata,common)
--     end
--     groupdata[group] = gdata
-- end

--~ characters.data, characters.groups = chardata, groupdata