summaryrefslogtreecommitdiff
path: root/lualibs-unicode.lua
blob: 02dd1a00304a32b0dbef8b04f5622b4281327a8e (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
if not modules then modules = { } end modules ['l-unicode'] = {
    version   = 1.001,
    comment   = "companion to luat-lib.mkiv",
    author    = "Hans Hagen, PRAGMA-ADE, Hasselt NL",
    copyright = "PRAGMA ADE / ConTeXt Development Team",
    license   = "see context related readme files"
}

-- in lua 5.3:

-- utf8.char(···)         : concatinated
-- utf8.charpatt          : "[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
-- utf8.codes(s)          : for p, c in utf8.codes(s) do body end
-- utf8.codepoint(s [, i [, j]])
-- utf8.len(s [, i])
-- utf8.offset(s, n [, i])

-- todo: utf.sub replacement (used in syst-aux)
-- we put these in the utf namespace:

utf = utf or (unicode and unicode.utf8) or { }

utf.characters = utf.characters or string.utfcharacters
utf.values     = utf.values     or string.utfvalues

-- string.utfvalues
-- string.utfcharacters
-- string.characters
-- string.characterpairs
-- string.bytes
-- string.bytepairs

local type = type
local char, byte, format, sub, gmatch = string.char, string.byte, string.format, string.sub, string.gmatch
local concat = table.concat
local P, C, R, Cs, Ct, Cmt, Cc, Carg, Cp = lpeg.P, lpeg.C, lpeg.R, lpeg.Cs, lpeg.Ct, lpeg.Cmt, lpeg.Cc, lpeg.Carg, lpeg.Cp

local lpegmatch       = lpeg.match
local patterns        = lpeg.patterns
local tabletopattern  = lpeg.utfchartabletopattern

local bytepairs       = string.bytepairs

local finder          = lpeg.finder
local replacer        = lpeg.replacer

local utfvalues       = utf.values
local utfgmatch       = utf.gmatch -- not always present

local p_utftype       = patterns.utftype
local p_utfstricttype = patterns.utfstricttype
local p_utfoffset     = patterns.utfoffset
local p_utf8char      = patterns.utf8character
local p_utf8byte      = patterns.utf8byte
local p_utfbom        = patterns.utfbom
local p_newline       = patterns.newline
local p_whitespace    = patterns.whitespace

if not unicode then

    unicode = { utf = utf } -- for a while

end

if not utf.char then

    local floor, char = math.floor, string.char

    function utf.char(n)
        if n < 0x80 then
            -- 0aaaaaaa : 0x80
            return char(n)
        elseif n < 0x800 then
            -- 110bbbaa : 0xC0 : n >> 6
            -- 10aaaaaa : 0x80 : n & 0x3F
            return char(
                0xC0 + floor(n/0x40),
                0x80 + (n % 0x40)
            )
        elseif n < 0x10000 then
            -- 1110bbbb : 0xE0 :  n >> 12
            -- 10bbbbaa : 0x80 : (n >>  6) & 0x3F
            -- 10aaaaaa : 0x80 :  n        & 0x3F
            return char(
                0xE0 + floor(n/0x1000),
                0x80 + (floor(n/0x40) % 0x40),
                0x80 + (n % 0x40)
            )
        elseif n < 0x200000 then
            -- 11110ccc : 0xF0 :  n >> 18
            -- 10ccbbbb : 0x80 : (n >> 12) & 0x3F
            -- 10bbbbaa : 0x80 : (n >>  6) & 0x3F
            -- 10aaaaaa : 0x80 :  n        & 0x3F
            -- dddd     : ccccc - 1
            return char(
                0xF0 +  floor(n/0x40000),
                0x80 + (floor(n/0x1000) % 0x40),
                0x80 + (floor(n/0x40) % 0x40),
                0x80 + (n % 0x40)
            )
        else
            return ""
        end
    end

end

if not utf.byte then

    local utf8byte = patterns.utf8byte

    function utf.byte(c)
        return lpegmatch(utf8byte,c)
    end

end

local utfchar, utfbyte = utf.char, utf.byte

-- As we want to get rid of the (unmaintained) utf library we implement our own
-- variants (in due time an independent module):

function utf.filetype(data)
    return data and lpegmatch(p_utftype,data) or "unknown"
end

local toentities = Cs (
    (
        patterns.utf8one
            + (
                patterns.utf8two
              + patterns.utf8three
              + patterns.utf8four
            ) / function(s) local b = utfbyte(s) if b < 127 then return s else return format("&#%X;",b) end end
    )^0
)

patterns.toentities = toentities

function utf.toentities(str)
    return lpegmatch(toentities,str)
end

-- local utfchr = { } -- 60K -> 2.638 M extra mem but currently not called that often (on latin)
--
-- setmetatable(utfchr, { __index = function(t,k) local v = utfchar(k) t[k] = v return v end } )
--
-- collectgarbage("collect")
-- local u = collectgarbage("count")*1024
-- local t = os.clock()
-- for i=1,1000 do
--     for i=1,600 do
--         local a = utfchr[i]
--     end
-- end
-- print(os.clock()-t,collectgarbage("count")*1024-u)

-- collectgarbage("collect")
-- local t = os.clock()
-- for i=1,1000 do
--     for i=1,600 do
--         local a = utfchar(i)
--     end
-- end
-- print(os.clock()-t,collectgarbage("count")*1024-u)

-- local byte = string.byte
-- local utfchar = utf.char

local one  = P(1)
local two  = C(1) * C(1)
local four = C(R(utfchar(0xD8),utfchar(0xFF))) * C(1) * C(1) * C(1)

-- actually one of them is already utf ... sort of useless this one

-- function utf.char(n)
--     if n < 0x80 then
--         return char(n)
--     elseif n < 0x800 then
--         return char(
--             0xC0 + floor(n/0x40),
--             0x80 + (n % 0x40)
--         )
--     elseif n < 0x10000 then
--         return char(
--             0xE0 + floor(n/0x1000),
--             0x80 + (floor(n/0x40) % 0x40),
--             0x80 + (n % 0x40)
--         )
--     elseif n < 0x40000 then
--         return char(
--             0xF0 + floor(n/0x40000),
--             0x80 + floor(n/0x1000),
--             0x80 + (floor(n/0x40) % 0x40),
--             0x80 + (n % 0x40)
--         )
--     else
--      -- return char(
--      --     0xF1 + floor(n/0x1000000),
--      --     0x80 + floor(n/0x40000),
--      --     0x80 + floor(n/0x1000),
--      --     0x80 + (floor(n/0x40) % 0x40),
--      --     0x80 + (n % 0x40)
--      -- )
--         return "?"
--     end
-- end
--
-- merge into:

local pattern = P("\254\255") * Cs( (
                    four  / function(a,b,c,d)
                                local ab = 0xFF * byte(a) + byte(b)
                                local cd = 0xFF * byte(c) + byte(d)
                                return utfchar((ab-0xD800)*0x400 + (cd-0xDC00) + 0x10000)
                            end
                  + two   / function(a,b)
                                return utfchar(byte(a)*256 + byte(b))
                            end
                  + one
                )^1 )
              + P("\255\254") * Cs( (
                    four  / function(b,a,d,c)
                                local ab = 0xFF * byte(a) + byte(b)
                                local cd = 0xFF * byte(c) + byte(d)
                                return utfchar((ab-0xD800)*0x400 + (cd-0xDC00) + 0x10000)
                            end
                  + two   / function(b,a)
                                return utfchar(byte(a)*256 + byte(b))
                            end
                  + one
                )^1 )

function string.toutf(s) -- in string namespace
    return lpegmatch(pattern,s) or s -- todo: utf32
end

local validatedutf = Cs (
    (
        patterns.utf8one
      + patterns.utf8two
      + patterns.utf8three
      + patterns.utf8four
      + P(1) / "�"
    )^0
)

patterns.validatedutf = validatedutf

function utf.is_valid(str)
    return type(str) == "string" and lpegmatch(validatedutf,str) or false
end

if not utf.len then

    -- -- alternative 1: 0.77
    --
    -- local utfcharcounter = utfbom^-1 * Cs((p_utf8char/'!')^0)
    --
    -- function utf.len(str)
    --     return #lpegmatch(utfcharcounter,str or "")
    -- end
    --
    -- -- alternative 2: 1.70
    --
    -- local n = 0
    --
    -- local utfcharcounter = utfbom^-1 * (p_utf8char/function() n = n + 1 end)^0 -- slow
    --
    -- function utf.length(str)
    --     n = 0
    --     lpegmatch(utfcharcounter,str or "")
    --     return n
    -- end
    --
    -- -- alternative 3: 0.24 (native unicode.utf8.len: 0.047)

    -- local n = 0
    --
    -- -- local utfcharcounter = lpeg.patterns.utfbom^-1 * P ( ( Cp() * (
    -- --     patterns.utf8one  ^1 * Cc(1)
    -- --   + patterns.utf8two  ^1 * Cc(2)
    -- --   + patterns.utf8three^1 * Cc(3)
    -- --   + patterns.utf8four ^1 * Cc(4) ) * Cp() / function(f,d,t) n = n + (t - f)/d end
    -- --  )^0 ) -- just as many captures as below
    --
    -- -- local utfcharcounter = lpeg.patterns.utfbom^-1 * P ( (
    -- --     (Cmt(patterns.utf8one  ^1,function(_,_,s) n = n + #s   return true end))
    -- --   + (Cmt(patterns.utf8two  ^1,function(_,_,s) n = n + #s/2 return true end))
    -- --   + (Cmt(patterns.utf8three^1,function(_,_,s) n = n + #s/3 return true end))
    -- --   + (Cmt(patterns.utf8four ^1,function(_,_,s) n = n + #s/4 return true end))
    -- -- )^0 ) -- not interesting as it creates strings but sometimes faster
    --
    -- -- The best so far:
    --
    -- local utfcharcounter = utfbom^-1 * P ( (
    --     Cp() * (patterns.utf8one  )^1 * Cp() / function(f,t) n = n +  t - f    end
    --   + Cp() * (patterns.utf8two  )^1 * Cp() / function(f,t) n = n + (t - f)/2 end
    --   + Cp() * (patterns.utf8three)^1 * Cp() / function(f,t) n = n + (t - f)/3 end
    --   + Cp() * (patterns.utf8four )^1 * Cp() / function(f,t) n = n + (t - f)/4 end
    -- )^0 )

    -- function utf.len(str)
    --     n = 0
    --     lpegmatch(utfcharcounter,str or "")
    --     return n
    -- end

    local n, f = 0, 1

    local utfcharcounter = patterns.utfbom^-1 * Cmt (
        Cc(1) * patterns.utf8one  ^1
      + Cc(2) * patterns.utf8two  ^1
      + Cc(3) * patterns.utf8three^1
      + Cc(4) * patterns.utf8four ^1,
        function(_,t,d) -- due to Cc no string captures, so faster
            n = n + (t - f)/d
            f = t
            return true
        end
    )^0

    function utf.len(str)
        n, f = 0, 1
        lpegmatch(utfcharcounter,str or "")
        return n
    end

    -- -- these are quite a bit slower:

    -- utfcharcounter = utfbom^-1 * (Cmt(P(1) * R("\128\191")^0, function() n = n + 1 return true end))^0 -- 50+ times slower
    -- utfcharcounter = utfbom^-1 * (Cmt(P(1), function() n = n + 1 return true end) * R("\128\191")^0)^0 -- 50- times slower

end

utf.length = utf.len

if not utf.sub then

    -- inefficient as lpeg just copies ^n

    -- local function sub(str,start,stop)
    --     local pattern = p_utf8char^-(start-1) * C(p_utf8char^-(stop-start+1))
    --     inspect(pattern)
    --     return lpegmatch(pattern,str) or ""
    -- end

    -- local b, e, n, first, last = 0, 0, 0, 0, 0
    --
    -- local function slide(s,p)
    --     n = n + 1
    --     if n == first then
    --         b = p
    --         if not last then
    --             return nil
    --         end
    --     end
    --     if n == last then
    --         e = p
    --         return nil
    --     else
    --         return p
    --     end
    -- end
    --
    -- local pattern = Cmt(p_utf8char,slide)^0
    --
    -- function utf.sub(str,start,stop) -- todo: from the end
    --     if not start then
    --         return str
    --     end
    --     b, e, n, first, last = 0, 0, 0, start, stop
    --     lpegmatch(pattern,str)
    --     if not stop then
    --         return sub(str,b)
    --     else
    --         return sub(str,b,e-1)
    --     end
    -- end

    -- print(utf.sub("Hans Hagen is my name"))
    -- print(utf.sub("Hans Hagen is my name",5))
    -- print(utf.sub("Hans Hagen is my name",5,10))

    local utflength = utf.length

    -- also negative indices, upto 10 times slower than a c variant

    local b, e, n, first, last = 0, 0, 0, 0, 0

    local function slide_zero(s,p)
        n = n + 1
        if n >= last then
            e = p - 1
        else
            return p
        end
    end

    local function slide_one(s,p)
        n = n + 1
        if n == first then
            b = p
        end
        if n >= last then
            e = p - 1
        else
            return p
        end
    end

    local function slide_two(s,p)
        n = n + 1
        if n == first then
            b = p
        else
            return true
        end
    end

    local pattern_zero = Cmt(p_utf8char,slide_zero)^0
    local pattern_one  = Cmt(p_utf8char,slide_one )^0
    local pattern_two  = Cmt(p_utf8char,slide_two )^0

    function utf.sub(str,start,stop)
        if not start then
            return str
        end
        if start == 0 then
            start = 1
        end
        if not stop then
            if start < 0 then
                local l = utflength(str) -- we can inline this function if needed
                start = l + start
            else
                start = start - 1
            end
            b, n, first = 0, 0, start
            lpegmatch(pattern_two,str)
            if n >= first then
                return sub(str,b)
            else
                return ""
            end
        end
        if start < 0 or stop < 0 then
            local l = utf.length(str)
            if start < 0 then
                start = l + start
                if start <= 0 then
                    start = 1
                else
                    start = start + 1
                end
            end
            if stop < 0 then
                stop = l + stop
                if stop == 0 then
                    stop = 1
                else
                    stop = stop + 1
                end
            end
        end
        if start > stop then
            return ""
        elseif start > 1 then
            b, e, n, first, last = 0, 0, 0, start - 1, stop
            lpegmatch(pattern_one,str)
            if n >= first and e == 0 then
                e = #str
            end
            return sub(str,b,e)
        else
            b, e, n, last = 1, 0, 0, stop
            lpegmatch(pattern_zero,str)
            if e == 0 then
                e = #str
            end
            return sub(str,b,e)
        end
    end

    -- local n = 100000
    -- local str = string.rep("123456àáâãäå",100)
    --
    -- for i=-15,15,1 do
    --     for j=-15,15,1 do
    --         if utf.xsub(str,i,j) ~= utf.sub(str,i,j) then
    --             print("error",i,j,"l>"..utf.xsub(str,i,j),"s>"..utf.sub(str,i,j))
    --         end
    --     end
    --     if utf.xsub(str,i) ~= utf.sub(str,i) then
    --         print("error",i,"l>"..utf.xsub(str,i),"s>"..utf.sub(str,i))
    --     end
    -- end

    -- print(" 1, 7",utf.xsub(str, 1, 7),utf.sub(str, 1, 7))
    -- print(" 0, 7",utf.xsub(str, 0, 7),utf.sub(str, 0, 7))
    -- print(" 0, 9",utf.xsub(str, 0, 9),utf.sub(str, 0, 9))
    -- print(" 4   ",utf.xsub(str, 4   ),utf.sub(str, 4   ))
    -- print(" 0   ",utf.xsub(str, 0   ),utf.sub(str, 0   ))
    -- print(" 0, 0",utf.xsub(str, 0, 0),utf.sub(str, 0, 0))
    -- print(" 4, 4",utf.xsub(str, 4, 4),utf.sub(str, 4, 4))
    -- print(" 4, 0",utf.xsub(str, 4, 0),utf.sub(str, 4, 0))
    -- print("-3, 0",utf.xsub(str,-3, 0),utf.sub(str,-3, 0))
    -- print(" 0,-3",utf.xsub(str, 0,-3),utf.sub(str, 0,-3))
    -- print(" 5,-3",utf.xsub(str,-5,-3),utf.sub(str,-5,-3))
    -- print("-3   ",utf.xsub(str,-3   ),utf.sub(str,-3   ))

end

-- a replacement for simple gsubs:

-- function utf.remapper(mapping)
--     local pattern = Cs((p_utf8char/mapping)^0)
--     return function(str)
--         if not str or str == "" then
--             return ""
--         else
--             return lpegmatch(pattern,str)
--         end
--     end, pattern
-- end

function utf.remapper(mapping,option) -- static also returns a pattern
    local variant = type(mapping)
    if variant == "table" then
        if option == "dynamic" then
            local pattern = false
            table.setmetatablenewindex(mapping,function(t,k,v) rawset(t,k,v) pattern = false end)
            return function(str)
                if not str or str == "" then
                    return ""
                else
                    if not pattern then
                        pattern = Cs((tabletopattern(mapping)/mapping + p_utf8char)^0)
                    end
                    return lpegmatch(pattern,str)
                end
            end
        elseif option == "pattern" then
            return Cs((tabletopattern(mapping)/mapping + p_utf8char)^0)
     -- elseif option == "static" then
        else
            local pattern = Cs((tabletopattern(mapping)/mapping + p_utf8char)^0)
            return function(str)
                if not str or str == "" then
                    return ""
                else
                    return lpegmatch(pattern,str)
                end
            end, pattern
        end
    elseif variant == "function" then
        if option == "pattern" then
            return Cs((p_utf8char/mapping + p_utf8char)^0)
        else
            local pattern = Cs((p_utf8char/mapping + p_utf8char)^0)
            return function(str)
                if not str or str == "" then
                    return ""
                else
                    return lpegmatch(pattern,str)
                end
            end, pattern
        end
    else
        -- is actually an error
        return function(str)
            return str or ""
        end
    end
end

-- local remap = utf.remapper { a = 'd', b = "c", c = "b", d = "a" }
-- print(remap("abcd 1234 abcd"))

function utf.replacer(t) -- no precheck, always string builder
    local r = replacer(t,false,false,true)
    return function(str)
        return lpegmatch(r,str)
    end
end

function utf.subtituter(t) -- with precheck and no building if no match
    local f = finder  (t)
    local r = replacer(t,false,false,true)
    return function(str)
        local i = lpegmatch(f,str)
        if not i then
            return str
        elseif i > #str then
            return str
        else
         -- return sub(str,1,i-2) .. lpegmatch(r,str,i-1) -- slower
            return lpegmatch(r,str)
        end
    end
end

-- inspect(utf.split("a b c d"))
-- inspect(utf.split("a b c d",true))

local utflinesplitter     = p_utfbom^-1 * lpeg.tsplitat(p_newline)
local utfcharsplitter_ows = p_utfbom^-1 * Ct(C(p_utf8char)^0)
local utfcharsplitter_iws = p_utfbom^-1 * Ct((p_whitespace^1 + C(p_utf8char))^0)
local utfcharsplitter_raw = Ct(C(p_utf8char)^0)

patterns.utflinesplitter  = utflinesplitter

function utf.splitlines(str)
    return lpegmatch(utflinesplitter,str or "")
end

function utf.split(str,ignorewhitespace) -- new
    if ignorewhitespace then
        return lpegmatch(utfcharsplitter_iws,str or "")
    else
        return lpegmatch(utfcharsplitter_ows,str or "")
    end
end

function utf.totable(str) -- keeps bom
    return lpegmatch(utfcharsplitter_raw,str)
end

-- 0  EF BB BF      UTF-8
-- 1  FF FE         UTF-16-little-endian
-- 2  FE FF         UTF-16-big-endian
-- 3  FF FE 00 00   UTF-32-little-endian
-- 4  00 00 FE FF   UTF-32-big-endian
--
-- \000 fails in <= 5.0 but is valid in >=5.1 where %z is depricated

-- utf.name = {
--     [0] = 'utf-8',
--     [1] = 'utf-16-le',
--     [2] = 'utf-16-be',
--     [3] = 'utf-32-le',
--     [4] = 'utf-32-be'
-- }
--
-- function utf.magic(f)
--     local str = f:read(4)
--     if not str then
--         f:seek('set')
--         return 0
--  -- elseif find(str,"^%z%z\254\255") then            -- depricated
--  -- elseif find(str,"^\000\000\254\255") then        -- not permitted and bugged
--     elseif find(str,"\000\000\254\255",1,true) then  -- seems to work okay (TH)
--         return 4
--  -- elseif find(str,"^\255\254%z%z") then            -- depricated
--  -- elseif find(str,"^\255\254\000\000") then        -- not permitted and bugged
--     elseif find(str,"\255\254\000\000",1,true) then  -- seems to work okay (TH)
--         return 3
--     elseif find(str,"^\254\255") then
--         f:seek('set',2)
--         return 2
--     elseif find(str,"^\255\254") then
--         f:seek('set',2)
--         return 1
--     elseif find(str,"^\239\187\191") then
--         f:seek('set',3)
--         return 0
--     else
--         f:seek('set')
--         return 0
--     end
-- end

function utf.magic(f) -- not used
    local str = f:read(4) or ""
    local off = lpegmatch(p_utfoffset,str)
    if off < 4 then
        f:seek('set',off)
    end
    return lpegmatch(p_utftype,str)
end

local utf16_to_utf8_be, utf16_to_utf8_le
local utf32_to_utf8_be, utf32_to_utf8_le

local utf_16_be_getbom = patterns.utfbom_16_be^-1
local utf_16_le_getbom = patterns.utfbom_16_le^-1
local utf_32_be_getbom = patterns.utfbom_32_be^-1
local utf_32_le_getbom = patterns.utfbom_32_le^-1

local utf_16_be_linesplitter = utf_16_be_getbom * lpeg.tsplitat(patterns.utf_16_be_nl)
local utf_16_le_linesplitter = utf_16_le_getbom * lpeg.tsplitat(patterns.utf_16_le_nl)
local utf_32_be_linesplitter = utf_32_be_getbom * lpeg.tsplitat(patterns.utf_32_be_nl)
local utf_32_le_linesplitter = utf_32_le_getbom * lpeg.tsplitat(patterns.utf_32_le_nl)

-- we have three possibilities: bytepairs (using tables), gmatch (using tables), gsub and
-- lpeg. Bytepairs are the fastert but as soon as we need to remove bombs and so the gain
-- is less due to more testing. Also, we seldom have to convert utf16 so we don't care to
-- much about a few  milliseconds more runtime. The lpeg variant is upto 20% slower but
-- still pretty fast.
--
-- for historic resone we keep the bytepairs variants around .. beware they don't grab the
-- bom like the lpegs do so they're not dropins in the functions that follow
--
-- utf16_to_utf8_be = function(s)
--     if not s then
--         return nil
--     elseif s == "" then
--         return ""
--     end
--     local result, r, more = { }, 0, 0
--     for left, right in bytepairs(s) do
--         if right then
--             local now = 256*left + right
--             if more > 0 then
--                 now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
--                 more = 0
--                 r = r + 1
--                 result[r] = utfchar(now)
--             elseif now >= 0xD800 and now <= 0xDBFF then
--                 more = now
--             else
--                 r = r + 1
--                 result[r] = utfchar(now)
--             end
--         end
--     end
--     return concat(result)
-- end
--
-- local utf16_to_utf8_be_t = function(t)
--     if not t then
--         return nil
--     elseif type(t) == "string" then
--         t = lpegmatch(utf_16_be_linesplitter,t)
--     end
--     local result = { } -- we reuse result
--     for i=1,#t do
--         local s = t[i]
--         if s ~= "" then
--             local r, more = 0, 0
--             for left, right in bytepairs(s) do
--                 if right then
--                     local now = 256*left + right
--                     if more > 0 then
--                         now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
--                         more = 0
--                         r = r + 1
--                         result[r] = utfchar(now)
--                     elseif now >= 0xD800 and now <= 0xDBFF then
--                         more = now
--                     else
--                         r = r + 1
--                         result[r] = utfchar(now)
--                     end
--                 end
--             end
--             t[i] = concat(result,"",1,r) -- we reused tmp, hence t
--         end
--     end
--     return t
-- end
--
-- utf16_to_utf8_le = function(s)
--     if not s then
--         return nil
--     elseif s == "" then
--         return ""
--     end
--     local result, r, more = { }, 0, 0
--     for left, right in bytepairs(s) do
--         if right then
--             local now = 256*right + left
--             if more > 0 then
--                 now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
--                 more = 0
--                 r = r + 1
--                 result[r] = utfchar(now)
--             elseif now >= 0xD800 and now <= 0xDBFF then
--                 more = now
--             else
--                 r = r + 1
--                 result[r] = utfchar(now)
--             end
--         end
--     end
--     return concat(result)
-- end
--
-- local utf16_to_utf8_le_t = function(t)
--     if not t then
--         return nil
--     elseif type(t) == "string" then
--         t = lpegmatch(utf_16_le_linesplitter,t)
--     end
--     local result = { } -- we reuse result
--     for i=1,#t do
--         local s = t[i]
--         if s ~= "" then
--             local r, more = 0, 0
--             for left, right in bytepairs(s) do
--                 if right then
--                     local now = 256*right + left
--                     if more > 0 then
--                         now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
--                         more = 0
--                         r = r + 1
--                         result[r] = utfchar(now)
--                     elseif now >= 0xD800 and now <= 0xDBFF then
--                         more = now
--                     else
--                         r = r + 1
--                         result[r] = utfchar(now)
--                     end
--                 end
--             end
--             t[i] = concat(result,"",1,r) -- we reused tmp, hence t
--         end
--     end
--     return t
-- end
--
-- local utf32_to_utf8_be_t = function(t)
--     if not t then
--         return nil
--     elseif type(t) == "string" then
--         t = lpegmatch(utflinesplitter,t)
--     end
--     local result = { } -- we reuse result
--     for i=1,#t do
--         local r, more = 0, -1
--         for a,b in bytepairs(t[i]) do
--             if a and b then
--                 if more < 0 then
--                     more = 256*256*256*a + 256*256*b
--                 else
--                     r = r + 1
--                     result[t] = utfchar(more + 256*a + b)
--                     more = -1
--                 end
--             else
--                 break
--             end
--         end
--         t[i] = concat(result,"",1,r)
--     end
--     return t
-- end
--
-- local utf32_to_utf8_le_t = function(t)
--     if not t then
--         return nil
--     elseif type(t) == "string" then
--         t = lpegmatch(utflinesplitter,t)
--     end
--     local result = { } -- we reuse result
--     for i=1,#t do
--         local r, more = 0, -1
--         for a,b in bytepairs(t[i]) do
--             if a and b then
--                 if more < 0 then
--                     more = 256*b + a
--                 else
--                     r = r + 1
--                     result[t] = utfchar(more + 256*256*256*b + 256*256*a)
--                     more = -1
--                 end
--             else
--                 break
--             end
--         end
--         t[i] = concat(result,"",1,r)
--     end
--     return t
-- end

local more = 0

local p_utf16_to_utf8_be = C(1) * C(1) /function(left,right)
    local now = 256*byte(left) + byte(right)
    if more > 0 then
        now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
        more = 0
        return utfchar(now)
    elseif now >= 0xD800 and now <= 0xDBFF then
        more = now
        return "" -- else the c's end up in the stream
    else
        return utfchar(now)
    end
end

local p_utf16_to_utf8_le = C(1) * C(1) /function(right,left)
    local now = 256*byte(left) + byte(right)
    if more > 0 then
        now = (more-0xD800)*0x400 + (now-0xDC00) + 0x10000 -- the 0x10000 smells wrong
        more = 0
        return utfchar(now)
    elseif now >= 0xD800 and now <= 0xDBFF then
        more = now
        return "" -- else the c's end up in the stream
    else
        return utfchar(now)
    end
end
local p_utf32_to_utf8_be = C(1) * C(1) * C(1) * C(1) /function(a,b,c,d)
    return utfchar(256*256*256*byte(a) + 256*256*byte(b) + 256*byte(c) + byte(d))
end

local p_utf32_to_utf8_le = C(1) * C(1) * C(1) * C(1) /function(a,b,c,d)
    return utfchar(256*256*256*byte(d) + 256*256*byte(c) + 256*byte(b) + byte(a))
end

p_utf16_to_utf8_be = P(true) / function() more = 0 end * utf_16_be_getbom * Cs(p_utf16_to_utf8_be^0)
p_utf16_to_utf8_le = P(true) / function() more = 0 end * utf_16_le_getbom * Cs(p_utf16_to_utf8_le^0)
p_utf32_to_utf8_be = P(true) / function() more = 0 end * utf_32_be_getbom * Cs(p_utf32_to_utf8_be^0)
p_utf32_to_utf8_le = P(true) / function() more = 0 end * utf_32_le_getbom * Cs(p_utf32_to_utf8_le^0)

patterns.utf16_to_utf8_be = p_utf16_to_utf8_be
patterns.utf16_to_utf8_le = p_utf16_to_utf8_le
patterns.utf32_to_utf8_be = p_utf32_to_utf8_be
patterns.utf32_to_utf8_le = p_utf32_to_utf8_le

utf16_to_utf8_be = function(s)
    if s and s ~= "" then
        return lpegmatch(p_utf16_to_utf8_be,s)
    else
        return s
    end
end

local utf16_to_utf8_be_t = function(t)
    if not t then
        return nil
    elseif type(t) == "string" then
        t = lpegmatch(utf_16_be_linesplitter,t)
    end
    for i=1,#t do
        local s = t[i]
        if s ~= "" then
            t[i] = lpegmatch(p_utf16_to_utf8_be,s)
        end
    end
    return t
end

utf16_to_utf8_le = function(s)
    if s and s ~= "" then
        return lpegmatch(p_utf16_to_utf8_le,s)
    else
        return s
    end
end

local utf16_to_utf8_le_t = function(t)
    if not t then
        return nil
    elseif type(t) == "string" then
        t = lpegmatch(utf_16_le_linesplitter,t)
    end
    for i=1,#t do
        local s = t[i]
        if s ~= "" then
            t[i] = lpegmatch(p_utf16_to_utf8_le,s)
        end
    end
    return t
end

utf32_to_utf8_be = function(s)
    if s and s ~= "" then
        return lpegmatch(p_utf32_to_utf8_be,s)
    else
        return s
    end
end

local utf32_to_utf8_be_t = function(t)
    if not t then
        return nil
    elseif type(t) == "string" then
        t = lpegmatch(utf_32_be_linesplitter,t)
    end
    for i=1,#t do
        local s = t[i]
        if s ~= "" then
            t[i] = lpegmatch(p_utf32_to_utf8_be,s)
        end
    end
    return t
end

utf32_to_utf8_le = function(s)
    if s and s ~= "" then
        return lpegmatch(p_utf32_to_utf8_le,s)
    else
        return s
    end
end

local utf32_to_utf8_le_t = function(t)
    if not t then
        return nil
    elseif type(t) == "string" then
        t = lpegmatch(utf_32_le_linesplitter,t)
    end
    for i=1,#t do
        local s = t[i]
        if s ~= "" then
            t[i] = lpegmatch(p_utf32_to_utf8_le,s)
        end
    end
    return t
end

utf.utf16_to_utf8_le_t = utf16_to_utf8_le_t
utf.utf16_to_utf8_be_t = utf16_to_utf8_be_t
utf.utf32_to_utf8_le_t = utf32_to_utf8_le_t
utf.utf32_to_utf8_be_t = utf32_to_utf8_be_t

utf.utf16_to_utf8_le   = utf16_to_utf8_le
utf.utf16_to_utf8_be   = utf16_to_utf8_be
utf.utf32_to_utf8_le   = utf32_to_utf8_le
utf.utf32_to_utf8_be   = utf32_to_utf8_be

function utf.utf8_to_utf8_t(t)
    return type(t) == "string" and lpegmatch(utflinesplitter,t) or t
end

function utf.utf16_to_utf8_t(t,endian)
    return endian and utf16_to_utf8_be_t(t) or utf16_to_utf8_le_t(t) or t
end

function utf.utf32_to_utf8_t(t,endian)
    return endian and utf32_to_utf8_be_t(t) or utf32_to_utf8_le_t(t) or t
end

local function little(b)
    if b < 0x10000 then
        return char(b%256,b/256)
    else
        b = b - 0x10000
        local b1, b2 = b/1024 + 0xD800, b%1024 + 0xDC00
        return char(b1%256,b1/256,b2%256,b2/256)
    end
end

local function big(b)
    if b < 0x10000 then
        return char(b/256,b%256)
    else
        b = b - 0x10000
        local b1, b2 = b/1024 + 0xD800, b%1024 + 0xDC00
        return char(b1/256,b1%256,b2/256,b2%256)
    end
end

local l_remap = Cs((p_utf8byte/little+P(1)/"")^0)
local b_remap = Cs((p_utf8byte/big   +P(1)/"")^0)

local function utf8_to_utf16_be(str,nobom)
    if nobom then
        return lpegmatch(b_remap,str)
    else
        return char(254,255) .. lpegmatch(b_remap,str)
    end
end

local function utf8_to_utf16_le(str,nobom)
    if nobom then
        return lpegmatch(l_remap,str)
    else
        return char(255,254) .. lpegmatch(l_remap,str)
    end
end

utf.utf8_to_utf16_be = utf8_to_utf16_be
utf.utf8_to_utf16_le = utf8_to_utf16_le

function utf.utf8_to_utf16(str,littleendian,nobom)
    if littleendian then
        return utf8_to_utf16_le(str,nobom)
    else
        return utf8_to_utf16_be(str,nobom)
    end
end

-- function utf.tocodes(str,separator) -- can be sped up with an lpeg
--     local t, n = { }, 0
--     for u in utfvalues(str) do
--         n = n + 1
--         t[n] = format("0x%04X",u)
--     end
--     return concat(t,separator or " ")
-- end

local pattern = Cs (
    (p_utf8byte           / function(unicode          ) return format(  "0x%04X",          unicode) end) *
    (p_utf8byte * Carg(1) / function(unicode,separator) return format("%s0x%04X",separator,unicode) end)^0
)

function utf.tocodes(str,separator)
    return lpegmatch(pattern,str,1,separator or " ")
end

function utf.ustring(s)
    return format("U+%05X",type(s) == "number" and s or utfbyte(s))
end

function utf.xstring(s)
    return format("0x%05X",type(s) == "number" and s or utfbyte(s))
end

function utf.toeight(str)
    if not str or str == "" then
        return nil
    end
    local utftype = lpegmatch(p_utfstricttype,str)
    if utftype == "utf-8" then
        return sub(str,4)               -- remove the bom
    elseif utftype == "utf-16-be" then
        return utf16_to_utf8_be(str)    -- bom gets removed
    elseif utftype == "utf-16-le" then
        return utf16_to_utf8_le(str)    -- bom gets removed
    else
        return str
    end
end

--

local p_nany = p_utf8char / ""

if utfgmatch then

    function utf.count(str,what)
        if type(what) == "string" then
            local n = 0
            for _ in utfgmatch(str,what) do
                n = n + 1
            end
            return n
        else -- 4 times slower but still faster than / function
            return #lpegmatch(Cs((P(what)/" " + p_nany)^0),str)
        end
    end

else

    local cache = { }

    function utf.count(str,what)
        if type(what) == "string" then
            local p = cache[what]
            if not p then
                p = Cs((P(what)/" " + p_nany)^0)
                cache[p] = p
            end
            return #lpegmatch(p,str)
        else -- 4 times slower but still faster than / function
            return #lpegmatch(Cs((P(what)/" " + p_nany)^0),str)
        end
    end

end

-- maybe also register as string.utf*


if not utf.characters then

    -- New: this gmatch hack is taken from the Lua 5.2 book. It's about two times slower
    -- than the built-in string.utfcharacters.

    function utf.characters(str)
        return gmatch(str,".[\128-\191]*")
    end

    string.utfcharacters = utf.characters

end

if not utf.values then

    -- So, a logical next step is to check for the values variant. It over five times
    -- slower than the built-in string.utfvalues. I optimized it a bit for n=0,1.

    ----- wrap, yield, gmatch = coroutine.wrap, coroutine.yield, string.gmatch
    local find =  string.find

    local dummy = function()
        -- we share this one
    end

    -- function utf.values(str)
    --     local n = #str
    --     if n == 0 then
    --         return wrap(dummy)
    --     elseif n == 1 then
    --         return wrap(function() yield(utfbyte(str)) end)
    --     else
    --         return wrap(function() for s in gmatch(str,".[\128-\191]*") do
    --             yield(utfbyte(s))
    --         end end)
    --     end
    -- end
    --
    -- faster:

    function utf.values(str)
        local n = #str
        if n == 0 then
            return dummy
        elseif n == 1 then
            return function() return utfbyte(str) end
        else
            local p = 1
         -- local n = #str
            return function()
             -- if p <= n then -- slower than the last find
                    local b, e = find(str,".[\128-\191]*",p)
                    if b then
                        p = e + 1
                        return utfbyte(sub(str,b,e))
                    end
             -- end
            end
        end
    end

    -- slower:
    --
    -- local pattern = C(patterns.utf8character) * Cp()
    -- ----- pattern = patterns.utf8character/utfbyte * Cp()
    -- ----- pattern = patterns.utf8byte * Cp()
    --
    -- function utf.values(str) -- one of the cases where a find is faster than an lpeg
    --     local n = #str
    --     if n == 0 then
    --         return dummy
    --     elseif n == 1 then
    --         return function() return utfbyte(str) end
    --     else
    --         local p = 1
    --         return function()
    --             local s, e = lpegmatch(pattern,str,p)
    --             if e then
    --                 p = e
    --                 return utfbyte(s)
    --              -- return s
    --             end
    --         end
    --     end
    -- end

    string.utfvalues = utf.values

end

function utf.chrlen(u) -- u is number
    return
        (u < 0x80 and 1) or
        (u < 0xE0 and 2) or
        (u < 0xF0 and 3) or
        (u < 0xF8 and 4) or
        (u < 0xFC and 5) or
        (u < 0xFE and 6) or 0
end