From: Jarkko Hietaniemi <jhi@iki.fi>
Date: Fri, 26 Oct 2001 14:12:04 +0000 (+0000)
Subject: Unicode: add the case folding table.
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=c4051cc5dfb167f01d02a988561fb93023e83cac;p=p5sagit%2Fp5-mst-13.2.git

Unicode: add the case folding table.

p4raw-id: //depot/perl@12689
---

diff --git a/MANIFEST b/MANIFEST
index 2e3e9af..99e6773 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -1606,6 +1606,7 @@ lib/unicore/Scripts.pl		Unicode character database
 lib/unicore/Scripts.txt		Unicode character database
 lib/unicore/SpecCase.txt	Unicode character database
 lib/unicore/To/Digit.pl		Unicode character database
+lib/unicore/To/Fold.pl		Unicode character database
 lib/unicore/To/Lower.pl		Unicode character database
 lib/unicore/To/Title.pl		Unicode character database
 lib/unicore/To/Upper.pl		Unicode character database
diff --git a/lib/unicore/To/Fold.pl b/lib/unicore/To/Fold.pl
new file mode 100644
index 0000000..5a24150
--- /dev/null
+++ b/lib/unicore/To/Fold.pl
@@ -0,0 +1,830 @@
+# !!!!!!!   DO NOT EDIT THIS FILE   !!!!!!! 
+# This file is built by mktables from e.g. Unicode.txt.
+# Any changes made here will be lost!
+
+%utf8::ToSpecFold = (
+'223' => "\x{0073}\x{0073}",
+'304' => "\x{0069}",
+'305' => "\x{0069}",
+'329' => "\x{02BC}\x{006E}",
+'496' => "\x{006A}\x{030C}",
+'912' => "\x{03B9}\x{0308}\x{0301}",
+'944' => "\x{03C5}\x{0308}\x{0301}",
+'1415' => "\x{0565}\x{0582}",
+'7830' => "\x{0068}\x{0331}",
+'7831' => "\x{0074}\x{0308}",
+'7832' => "\x{0077}\x{030A}",
+'7833' => "\x{0079}\x{030A}",
+'7834' => "\x{0061}\x{02BE}",
+'8016' => "\x{03C5}\x{0313}",
+'8018' => "\x{03C5}\x{0313}\x{0300}",
+'8020' => "\x{03C5}\x{0313}\x{0301}",
+'8022' => "\x{03C5}\x{0313}\x{0342}",
+'8064' => "\x{1F00}\x{03B9}",
+'8065' => "\x{1F01}\x{03B9}",
+'8066' => "\x{1F02}\x{03B9}",
+'8067' => "\x{1F03}\x{03B9}",
+'8068' => "\x{1F04}\x{03B9}",
+'8069' => "\x{1F05}\x{03B9}",
+'8070' => "\x{1F06}\x{03B9}",
+'8071' => "\x{1F07}\x{03B9}",
+'8072' => "\x{1F00}\x{03B9}",
+'8073' => "\x{1F01}\x{03B9}",
+'8074' => "\x{1F02}\x{03B9}",
+'8075' => "\x{1F03}\x{03B9}",
+'8076' => "\x{1F04}\x{03B9}",
+'8077' => "\x{1F05}\x{03B9}",
+'8078' => "\x{1F06}\x{03B9}",
+'8079' => "\x{1F07}\x{03B9}",
+'8080' => "\x{1F20}\x{03B9}",
+'8081' => "\x{1F21}\x{03B9}",
+'8082' => "\x{1F22}\x{03B9}",
+'8083' => "\x{1F23}\x{03B9}",
+'8084' => "\x{1F24}\x{03B9}",
+'8085' => "\x{1F25}\x{03B9}",
+'8086' => "\x{1F26}\x{03B9}",
+'8087' => "\x{1F27}\x{03B9}",
+'8088' => "\x{1F20}\x{03B9}",
+'8089' => "\x{1F21}\x{03B9}",
+'8090' => "\x{1F22}\x{03B9}",
+'8091' => "\x{1F23}\x{03B9}",
+'8092' => "\x{1F24}\x{03B9}",
+'8093' => "\x{1F25}\x{03B9}",
+'8094' => "\x{1F26}\x{03B9}",
+'8095' => "\x{1F27}\x{03B9}",
+'8096' => "\x{1F60}\x{03B9}",
+'8097' => "\x{1F61}\x{03B9}",
+'8098' => "\x{1F62}\x{03B9}",
+'8099' => "\x{1F63}\x{03B9}",
+'8100' => "\x{1F64}\x{03B9}",
+'8101' => "\x{1F65}\x{03B9}",
+'8102' => "\x{1F66}\x{03B9}",
+'8103' => "\x{1F67}\x{03B9}",
+'8104' => "\x{1F60}\x{03B9}",
+'8105' => "\x{1F61}\x{03B9}",
+'8106' => "\x{1F62}\x{03B9}",
+'8107' => "\x{1F63}\x{03B9}",
+'8108' => "\x{1F64}\x{03B9}",
+'8109' => "\x{1F65}\x{03B9}",
+'8110' => "\x{1F66}\x{03B9}",
+'8111' => "\x{1F67}\x{03B9}",
+'8114' => "\x{1F70}\x{03B9}",
+'8115' => "\x{03B1}\x{03B9}",
+'8116' => "\x{03AC}\x{03B9}",
+'8118' => "\x{03B1}\x{0342}",
+'8119' => "\x{03B1}\x{0342}\x{03B9}",
+'8124' => "\x{03B1}\x{03B9}",
+'8130' => "\x{1F74}\x{03B9}",
+'8131' => "\x{03B7}\x{03B9}",
+'8132' => "\x{03AE}\x{03B9}",
+'8134' => "\x{03B7}\x{0342}",
+'8135' => "\x{03B7}\x{0342}\x{03B9}",
+'8140' => "\x{03B7}\x{03B9}",
+'8146' => "\x{03B9}\x{0308}\x{0300}",
+'8147' => "\x{03B9}\x{0308}\x{0301}",
+'8150' => "\x{03B9}\x{0342}",
+'8151' => "\x{03B9}\x{0308}\x{0342}",
+'8162' => "\x{03C5}\x{0308}\x{0300}",
+'8163' => "\x{03C5}\x{0308}\x{0301}",
+'8164' => "\x{03C1}\x{0313}",
+'8166' => "\x{03C5}\x{0342}",
+'8167' => "\x{03C5}\x{0308}\x{0342}",
+'8178' => "\x{1F7C}\x{03B9}",
+'8179' => "\x{03C9}\x{03B9}",
+'8180' => "\x{03CE}\x{03B9}",
+'8182' => "\x{03C9}\x{0342}",
+'8183' => "\x{03C9}\x{0342}\x{03B9}",
+'8188' => "\x{03C9}\x{03B9}",
+'64256' => "\x{0066}\x{0066}",
+'64257' => "\x{0066}\x{0069}",
+'64258' => "\x{0066}\x{006C}",
+'64259' => "\x{0066}\x{0066}\x{0069}",
+'64260' => "\x{0066}\x{0066}\x{006C}",
+'64261' => "\x{0073}\x{0074}",
+'64262' => "\x{0073}\x{0074}",
+'64275' => "\x{0574}\x{0576}",
+'64276' => "\x{0574}\x{0565}",
+'64277' => "\x{0574}\x{056B}",
+'64278' => "\x{057E}\x{0576}",
+'64279' => "\x{0574}\x{056D}",
+);
+
+return <<'END';
+0041		0061
+0042		0062
+0043		0063
+0044		0064
+0045		0065
+0046		0066
+0047		0067
+0048		0068
+0049		0069
+004A		006A
+004B		006B
+004C		006C
+004D		006D
+004E		006E
+004F		006F
+0050		0070
+0051		0071
+0052		0072
+0053		0073
+0054		0074
+0055		0075
+0056		0076
+0057		0077
+0058		0078
+0059		0079
+005A		007A
+00B5		03BC
+00C0		00E0
+00C1		00E1
+00C2		00E2
+00C3		00E3
+00C4		00E4
+00C5		00E5
+00C6		00E6
+00C7		00E7
+00C8		00E8
+00C9		00E9
+00CA		00EA
+00CB		00EB
+00CC		00EC
+00CD		00ED
+00CE		00EE
+00CF		00EF
+00D0		00F0
+00D1		00F1
+00D2		00F2
+00D3		00F3
+00D4		00F4
+00D5		00F5
+00D6		00F6
+00D8		00F8
+00D9		00F9
+00DA		00FA
+00DB		00FB
+00DC		00FC
+00DD		00FD
+00DE		00FE
+0100		0101
+0102		0103
+0104		0105
+0106		0107
+0108		0109
+010A		010B
+010C		010D
+010E		010F
+0110		0111
+0112		0113
+0114		0115
+0116		0117
+0118		0119
+011A		011B
+011C		011D
+011E		011F
+0120		0121
+0122		0123
+0124		0125
+0126		0127
+0128		0129
+012A		012B
+012C		012D
+012E		012F
+0132		0133
+0134		0135
+0136		0137
+0139		013A
+013B		013C
+013D		013E
+013F		0140
+0141		0142
+0143		0144
+0145		0146
+0147		0148
+014A		014B
+014C		014D
+014E		014F
+0150		0151
+0152		0153
+0154		0155
+0156		0157
+0158		0159
+015A		015B
+015C		015D
+015E		015F
+0160		0161
+0162		0163
+0164		0165
+0166		0167
+0168		0169
+016A		016B
+016C		016D
+016E		016F
+0170		0171
+0172		0173
+0174		0175
+0176		0177
+0178		00FF
+0179		017A
+017B		017C
+017D		017E
+017F		0073
+0181		0253
+0182		0183
+0184		0185
+0186		0254
+0187		0188
+0189		0256
+018A		0257
+018B		018C
+018E		01DD
+018F		0259
+0190		025B
+0191		0192
+0193		0260
+0194		0263
+0196		0269
+0197		0268
+0198		0199
+019C		026F
+019D		0272
+019F		0275
+01A0		01A1
+01A2		01A3
+01A4		01A5
+01A6		0280
+01A7		01A8
+01A9		0283
+01AC		01AD
+01AE		0288
+01AF		01B0
+01B1		028A
+01B2		028B
+01B3		01B4
+01B5		01B6
+01B7		0292
+01B8		01B9
+01BC		01BD
+01C4	01C5	01C6
+01C7	01C8	01C9
+01CA	01CB	01CC
+01CD		01CE
+01CF		01D0
+01D1		01D2
+01D3		01D4
+01D5		01D6
+01D7		01D8
+01D9		01DA
+01DB		01DC
+01DE		01DF
+01E0		01E1
+01E2		01E3
+01E4		01E5
+01E6		01E7
+01E8		01E9
+01EA		01EB
+01EC		01ED
+01EE		01EF
+01F1	01F2	01F3
+01F4		01F5
+01F6		0195
+01F7		01BF
+01F8		01F9
+01FA		01FB
+01FC		01FD
+01FE		01FF
+0200		0201
+0202		0203
+0204		0205
+0206		0207
+0208		0209
+020A		020B
+020C		020D
+020E		020F
+0210		0211
+0212		0213
+0214		0215
+0216		0217
+0218		0219
+021A		021B
+021C		021D
+021E		021F
+0222		0223
+0224		0225
+0226		0227
+0228		0229
+022A		022B
+022C		022D
+022E		022F
+0230		0231
+0232		0233
+0345		03B9
+0386		03AC
+0388		03AD
+0389		03AE
+038A		03AF
+038C		03CC
+038E		03CD
+038F		03CE
+0391		03B1
+0392		03B2
+0393		03B3
+0394		03B4
+0395		03B5
+0396		03B6
+0397		03B7
+0398		03B8
+0399		03B9
+039A		03BA
+039B		03BB
+039C		03BC
+039D		03BD
+039E		03BE
+039F		03BF
+03A0		03C0
+03A1		03C1
+03A3		03C3
+03A4		03C4
+03A5		03C5
+03A6		03C6
+03A7		03C7
+03A8		03C8
+03A9		03C9
+03AA		03CA
+03AB		03CB
+03C2		03C3
+03D0		03B2
+03D1		03B8
+03D5		03C6
+03D6		03C0
+03DA		03DB
+03DC		03DD
+03DE		03DF
+03E0		03E1
+03E2		03E3
+03E4		03E5
+03E6		03E7
+03E8		03E9
+03EA		03EB
+03EC		03ED
+03EE		03EF
+03F0		03BA
+03F1		03C1
+03F2		03C3
+03F4		03B8
+03F5		03B5
+0400		0450
+0401		0451
+0402		0452
+0403		0453
+0404		0454
+0405		0455
+0406		0456
+0407		0457
+0408		0458
+0409		0459
+040A		045A
+040B		045B
+040C		045C
+040D		045D
+040E		045E
+040F		045F
+0410		0430
+0411		0431
+0412		0432
+0413		0433
+0414		0434
+0415		0435
+0416		0436
+0417		0437
+0418		0438
+0419		0439
+041A		043A
+041B		043B
+041C		043C
+041D		043D
+041E		043E
+041F		043F
+0420		0440
+0421		0441
+0422		0442
+0423		0443
+0424		0444
+0425		0445
+0426		0446
+0427		0447
+0428		0448
+0429		0449
+042A		044A
+042B		044B
+042C		044C
+042D		044D
+042E		044E
+042F		044F
+0460		0461
+0462		0463
+0464		0465
+0466		0467
+0468		0469
+046A		046B
+046C		046D
+046E		046F
+0470		0471
+0472		0473
+0474		0475
+0476		0477
+0478		0479
+047A		047B
+047C		047D
+047E		047F
+0480		0481
+048C		048D
+048E		048F
+0490		0491
+0492		0493
+0494		0495
+0496		0497
+0498		0499
+049A		049B
+049C		049D
+049E		049F
+04A0		04A1
+04A2		04A3
+04A4		04A5
+04A6		04A7
+04A8		04A9
+04AA		04AB
+04AC		04AD
+04AE		04AF
+04B0		04B1
+04B2		04B3
+04B4		04B5
+04B6		04B7
+04B8		04B9
+04BA		04BB
+04BC		04BD
+04BE		04BF
+04C1		04C2
+04C3		04C4
+04C7		04C8
+04CB		04CC
+04D0		04D1
+04D2		04D3
+04D4		04D5
+04D6		04D7
+04D8		04D9
+04DA		04DB
+04DC		04DD
+04DE		04DF
+04E0		04E1
+04E2		04E3
+04E4		04E5
+04E6		04E7
+04E8		04E9
+04EA		04EB
+04EC		04ED
+04EE		04EF
+04F0		04F1
+04F2		04F3
+04F4		04F5
+04F8		04F9
+0531		0561
+0532		0562
+0533		0563
+0534		0564
+0535		0565
+0536		0566
+0537		0567
+0538		0568
+0539		0569
+053A		056A
+053B		056B
+053C		056C
+053D		056D
+053E		056E
+053F		056F
+0540		0570
+0541		0571
+0542		0572
+0543		0573
+0544		0574
+0545		0575
+0546		0576
+0547		0577
+0548		0578
+0549		0579
+054A		057A
+054B		057B
+054C		057C
+054D		057D
+054E		057E
+054F		057F
+0550		0580
+0551		0581
+0552		0582
+0553		0583
+0554		0584
+0555		0585
+0556		0586
+1E00		1E01
+1E02		1E03
+1E04		1E05
+1E06		1E07
+1E08		1E09
+1E0A		1E0B
+1E0C		1E0D
+1E0E		1E0F
+1E10		1E11
+1E12		1E13
+1E14		1E15
+1E16		1E17
+1E18		1E19
+1E1A		1E1B
+1E1C		1E1D
+1E1E		1E1F
+1E20		1E21
+1E22		1E23
+1E24		1E25
+1E26		1E27
+1E28		1E29
+1E2A		1E2B
+1E2C		1E2D
+1E2E		1E2F
+1E30		1E31
+1E32		1E33
+1E34		1E35
+1E36		1E37
+1E38		1E39
+1E3A		1E3B
+1E3C		1E3D
+1E3E		1E3F
+1E40		1E41
+1E42		1E43
+1E44		1E45
+1E46		1E47
+1E48		1E49
+1E4A		1E4B
+1E4C		1E4D
+1E4E		1E4F
+1E50		1E51
+1E52		1E53
+1E54		1E55
+1E56		1E57
+1E58		1E59
+1E5A		1E5B
+1E5C		1E5D
+1E5E		1E5F
+1E60		1E61
+1E62		1E63
+1E64		1E65
+1E66		1E67
+1E68		1E69
+1E6A		1E6B
+1E6C		1E6D
+1E6E		1E6F
+1E70		1E71
+1E72		1E73
+1E74		1E75
+1E76		1E77
+1E78		1E79
+1E7A		1E7B
+1E7C		1E7D
+1E7E		1E7F
+1E80		1E81
+1E82		1E83
+1E84		1E85
+1E86		1E87
+1E88		1E89
+1E8A		1E8B
+1E8C		1E8D
+1E8E		1E8F
+1E90		1E91
+1E92		1E93
+1E94		1E95
+1E9B		1E61
+1EA0		1EA1
+1EA2		1EA3
+1EA4		1EA5
+1EA6		1EA7
+1EA8		1EA9
+1EAA		1EAB
+1EAC		1EAD
+1EAE		1EAF
+1EB0		1EB1
+1EB2		1EB3
+1EB4		1EB5
+1EB6		1EB7
+1EB8		1EB9
+1EBA		1EBB
+1EBC		1EBD
+1EBE		1EBF
+1EC0		1EC1
+1EC2		1EC3
+1EC4		1EC5
+1EC6		1EC7
+1EC8		1EC9
+1ECA		1ECB
+1ECC		1ECD
+1ECE		1ECF
+1ED0		1ED1
+1ED2		1ED3
+1ED4		1ED5
+1ED6		1ED7
+1ED8		1ED9
+1EDA		1EDB
+1EDC		1EDD
+1EDE		1EDF
+1EE0		1EE1
+1EE2		1EE3
+1EE4		1EE5
+1EE6		1EE7
+1EE8		1EE9
+1EEA		1EEB
+1EEC		1EED
+1EEE		1EEF
+1EF0		1EF1
+1EF2		1EF3
+1EF4		1EF5
+1EF6		1EF7
+1EF8		1EF9
+1F08		1F00
+1F09		1F01
+1F0A		1F02
+1F0B		1F03
+1F0C		1F04
+1F0D		1F05
+1F0E		1F06
+1F0F		1F07
+1F18		1F10
+1F19		1F11
+1F1A		1F12
+1F1B		1F13
+1F1C		1F14
+1F1D		1F15
+1F28		1F20
+1F29		1F21
+1F2A		1F22
+1F2B		1F23
+1F2C		1F24
+1F2D		1F25
+1F2E		1F26
+1F2F		1F27
+1F38		1F30
+1F39		1F31
+1F3A		1F32
+1F3B		1F33
+1F3C		1F34
+1F3D		1F35
+1F3E		1F36
+1F3F		1F37
+1F48		1F40
+1F49		1F41
+1F4A		1F42
+1F4B		1F43
+1F4C		1F44
+1F4D		1F45
+1F59		1F51
+1F5B		1F53
+1F5D		1F55
+1F5F		1F57
+1F68		1F60
+1F69		1F61
+1F6A		1F62
+1F6B		1F63
+1F6C		1F64
+1F6D		1F65
+1F6E		1F66
+1F6F		1F67
+1FB8		1FB0
+1FB9		1FB1
+1FBA		1F70
+1FBB		1F71
+1FBE		03B9
+1FC8		1F72
+1FC9		1F73
+1FCA		1F74
+1FCB		1F75
+1FD8		1FD0
+1FD9		1FD1
+1FDA		1F76
+1FDB		1F77
+1FE8		1FE0
+1FE9		1FE1
+1FEA		1F7A
+1FEB		1F7B
+1FEC		1FE5
+1FF8		1F78
+1FF9		1F79
+1FFA		1F7C
+1FFB		1F7D
+2126		03C9
+212A		006B
+212B		00E5
+2160		2170
+2161		2171
+2162		2172
+2163		2173
+2164		2174
+2165		2175
+2166		2176
+2167		2177
+2168		2178
+2169		2179
+216A		217A
+216B		217B
+216C		217C
+216D		217D
+216E		217E
+216F		217F
+24B6		24D0
+24B7		24D1
+24B8		24D2
+24B9		24D3
+24BA		24D4
+24BB		24D5
+24BC		24D6
+24BD		24D7
+24BE		24D8
+24BF		24D9
+24C0		24DA
+24C1		24DB
+24C2		24DC
+24C3		24DD
+24C4		24DE
+24C5		24DF
+24C6		24E0
+24C7		24E1
+24C8		24E2
+24C9		24E3
+24CA		24E4
+24CB		24E5
+24CC		24E6
+24CD		24E7
+24CE		24E8
+24CF		24E9
+FF21		FF41
+FF22		FF42
+FF23		FF43
+FF24		FF44
+FF25		FF45
+FF26		FF46
+FF27		FF47
+FF28		FF48
+FF29		FF49
+FF2A		FF4A
+FF2B		FF4B
+FF2C		FF4C
+FF2D		FF4D
+FF2E		FF4E
+FF2F		FF4F
+FF30		FF50
+FF31		FF51
+FF32		FF52
+FF33		FF53
+FF34		FF54
+FF35		FF55
+FF36		FF56
+FF37		FF57
+FF38		FF58
+FF39		FF59
+FF3A		FF5A
+10400		10428
+10401		10429
+10402		1042A
+10403		1042B
+10404		1042C
+10405		1042D
+10406		1042E
+10407		1042F
+10408		10430
+10409		10431
+1040A		10432
+1040B		10433
+1040C		10434
+1040D		10435
+1040E		10436
+1040F		10437
+10410		10438
+10411		10439
+10412		1043A
+10413		1043B
+10414		1043C
+10415		1043D
+10416		1043E
+10417		1043F
+10418		10440
+10419		10441
+1041A		10442
+1041B		10443
+1041C		10444
+1041D		10445
+1041E		10446
+1041F		10447
+10420		10448
+10421		10449
+10422		1044A
+10423		1044B
+10424		1044C
+10425		1044D
+END
diff --git a/lib/unicore/mktables b/lib/unicore/mktables
index 7d8912d..ea04974 100644
--- a/lib/unicore/mktables
+++ b/lib/unicore/mktables
@@ -748,7 +748,7 @@ if (open(my $SpecCase, "SpecCase.txt")) {
 # Prepend them to the To/{Upper,Lower,Title}.pl.
 
 for my $case (qw(Lower Title Upper)) {
-    my $NormalCase = do "To/$case.pl";
+    my $NormalCase = do "To/$case.pl" || die "$0: To/$case.pl: $!\n";
     if (open(my $Case, ">To/$case.pl")) {
 	header($Case);
 	print $Case <<EOT;
@@ -773,5 +773,59 @@ EOT
     }
 }
 
+#
+# Read in the case foldings.
+#
+# We will do full case folding, C + F + I (see CaseFold.txt).
+#
+
+if (open(my $CaseFold, "CaseFold.txt")) {
+    my @Fold;
+    my %Fold;
+
+    while (<$CaseFold>) {
+	next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/;
+
+	my ($code, $status, $fold) = ($1, $2, $3);
+
+	if ($status eq 'C') { # Common: one-to-one folding
+	    append(\@Fold, $code, $fold);
+	} else { # F: full, or I: dotted uppercase I -> dotless lowercase I
+	    $Fold{hex($code)} = $fold;
+	}
+    }
+
+    flush(\@Fold, "To/Fold.pl");
+
+    #
+    # Prepend the special foldings to the common foldings.
+    #
+
+    my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n";
+    if (open(my $Fold, ">To/Fold.pl")) {
+	header($Fold);
+	print $Fold <<EOT;
+
+%utf8::ToSpecFold = (
+EOT
+        for my $code (sort { $a <=> $b } keys %Fold) {
+	    my $foldstr =
+		join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code};
+	    print $Fold qq['$code' => "$foldstr",\n];
+	}
+	print $Fold <<EOT;
+);
+
+EOT
+	begin($Fold);
+	print $Fold $CommonFold;
+	end($Fold);
+    } else {
+	die "$0: To/Fold.pl: $!\n";
+    }
+} else {
+    die "$0: CaseFold.txt: $!\n";
+}
+
 # That's all, folks!