Commit | Line | Data |
---|---|---|
d73e5302 | 1 | #!/usr/bin/perl -w |
d73e5302 | 2 | use strict; |
cf25bb62 JH |
3 | use Carp; |
4 | ## | |
5 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) | |
6 | ## from the Unicode database files (lib/unicore/*.txt). | |
7 | ## | |
d2d499f5 | 8 | |
d73e5302 JH |
9 | mkdir("In", 0755); |
10 | mkdir("Is", 0755); | |
11 | mkdir("To", 0755); | |
12 | ||
cf25bb62 JH |
13 | ## |
14 | ## Process any args. | |
15 | ## | |
16 | my $Verbose = 0; | |
17 | ||
18 | while (@ARGV) | |
19 | { | |
20 | my $arg = shift @ARGV; | |
21 | if ($arg eq '-v') { | |
22 | $Verbose = 1; | |
23 | } elsif ($arg eq '-q') { | |
24 | $Verbose = 0; | |
d73e5302 | 25 | } else { |
cf25bb62 | 26 | die "usage: $0 [-v|-q]"; |
d73e5302 JH |
27 | } |
28 | } | |
29 | ||
cf25bb62 | 30 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 3.1.1. |
d73e5302 | 31 | |
cf25bb62 JH |
32 | my $now = localtime; |
33 | my $HEADER=<<"EOF"; | |
d73e5302 JH |
34 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! |
35 | # This file is built by $0 from e.g. Unicode.txt. | |
36 | # Any changes made here will be lost! | |
cf25bb62 JH |
37 | |
38 | EOF | |
39 | ||
40 | ## | |
41 | ## The main datastructure (a "Table") represents a set of code points that | |
42 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, | |
43 | ## etc.). They are kept as ranges of code points (starting and ending of | |
44 | ## each range). | |
45 | ## | |
46 | ## For example, a range ASCII LETTERS would be represented as: | |
47 | ## [ [ 0x41 => 0x5A, 'UPPER' ], | |
48 | ## [ 0x61 => 0x7A, 'LOWER, ] ] | |
49 | ## | |
50 | sub RANGE_START() { 0 } ## index into range element | |
51 | sub RANGE_END() { 1 } ## index into range element | |
52 | sub RANGE_NAME() { 2 } ## index into range element | |
53 | ||
54 | my %TableInfo; | |
55 | my %FuzzyNames; | |
56 | my %AliasInfo; | |
57 | ||
58 | ## | |
59 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. | |
60 | ## | |
61 | ## Called like: | |
62 | ## New_Prop(In => 'Greek', $Table, AllowFuzzy => 1); | |
63 | ## | |
64 | ## Normally, these parameters are set when the Table is created (when the | |
65 | ## Table->New constructor is called), but there are times when it needs to | |
66 | ## be done after-the-fact...) | |
67 | ## | |
68 | sub New_Prop($$$@) | |
69 | { | |
70 | my $Type = shift; ## "Is" or "In"; | |
71 | my $Name = shift; | |
72 | my $Table = shift; | |
73 | ||
74 | ## remaining args are optional key/val | |
75 | my %Args = @_; | |
76 | ||
77 | my $AllowFuzzy = delete $Args{AllowFuzzy}; | |
78 | ||
79 | ## sanity check a few args | |
80 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { | |
81 | confess "$0: bad args to New_Prop" | |
82 | } | |
83 | ||
84 | if (not $TableInfo{$Type}->{$Name}) | |
85 | { | |
86 | $TableInfo{$Type}->{$Name} = $Table; | |
87 | if ($AllowFuzzy) { | |
88 | $FuzzyNames{$Type}->{$Name} = $Name; | |
89 | } | |
90 | } | |
d73e5302 JH |
91 | } |
92 | ||
d73e5302 | 93 | |
cf25bb62 JH |
94 | ## |
95 | ## Creates a new Table object. | |
96 | ## | |
97 | ## Args are key/value pairs: | |
98 | ## In => Name -- Name of "In" property to be associated with | |
99 | ## Is => Name -- Name of "Is" property to be associated with | |
100 | ## AllowFuzzy => Boolean -- True if name can be accessed "fuzzily" | |
101 | ## | |
102 | ## No args are required. | |
103 | ## | |
104 | sub Table::New | |
105 | { | |
106 | my $class = shift; | |
107 | my %Args = @_; | |
108 | ||
109 | my $Table = bless [], $class; | |
110 | ||
111 | my $AllowFuzzy = delete $Args{AllowFuzzy}; | |
112 | ||
113 | for my $Type ('Is', 'In') | |
114 | { | |
115 | if (my $Name = delete $Args{$Type}) { | |
116 | New_Prop($Type => $Name, $Table, AllowFuzzy => $AllowFuzzy); | |
117 | } | |
118 | } | |
119 | ||
120 | ## shouldn't have any left over | |
121 | if (%Args) { | |
122 | confess "$0: bad args to Table->New" | |
123 | } | |
124 | ||
125 | return $Table; | |
d73e5302 JH |
126 | } |
127 | ||
cf25bb62 JH |
128 | ## |
129 | ## Returns true if the Table has no code points | |
130 | ## | |
131 | sub Table::IsEmpty | |
132 | { | |
133 | my $Table = shift; #self | |
134 | return not @$Table; | |
d73e5302 JH |
135 | } |
136 | ||
cf25bb62 JH |
137 | ## |
138 | ## Returns true if the Table has code points | |
139 | ## | |
140 | sub Table::NotEmpty | |
141 | { | |
142 | my $Table = shift; #self | |
143 | return @$Table; | |
d73e5302 JH |
144 | } |
145 | ||
cf25bb62 JH |
146 | ## |
147 | ## Returns the maximum code point currently in the table. | |
148 | ## | |
149 | sub Table::Max | |
150 | { | |
151 | my $Table = shift; #self | |
152 | confess "oops" if $Table->IsEmpty; ## must have code points to have a max | |
153 | return $Table->[-1]->[RANGE_END]; | |
154 | } | |
d73e5302 | 155 | |
cf25bb62 JH |
156 | ## |
157 | ## Replaces the codepoints in the Table with those in the Table given | |
158 | ## as an arg. (NOTE: this is not a "deep copy"). | |
159 | ## | |
160 | sub Table::Replace($$) | |
161 | { | |
162 | my $Table = shift; #self | |
163 | my $New = shift; | |
d73e5302 | 164 | |
cf25bb62 JH |
165 | @$Table = @$New; |
166 | } | |
71d929cb | 167 | |
cf25bb62 JH |
168 | ## |
169 | ## Given a new code point, make the last range of the Table extend to | |
170 | ## include the new (and all intervening) code points. | |
171 | ## | |
172 | sub Table::Extend | |
173 | { | |
174 | my $Table = shift; #self | |
175 | my $codepoint = shift; | |
d73e5302 | 176 | |
cf25bb62 | 177 | my $PrevMax = $Table->Max; |
e904f995 | 178 | |
cf25bb62 | 179 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; |
e904f995 | 180 | |
cf25bb62 JH |
181 | $Table->[-1]->[RANGE_END] = $codepoint; |
182 | } | |
c3a8a2b8 | 183 | |
cf25bb62 JH |
184 | ## |
185 | ## Given a code point range start and end (and optional name), blindly | |
186 | ## append them to the list of ranges for the Table. | |
187 | ## | |
188 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
189 | ## | |
190 | sub Table::RawAppendRange | |
191 | { | |
192 | my $Table = shift; #self | |
193 | my $start = shift; | |
194 | my $end = shift; | |
195 | my $name = shift; | |
196 | $name = "" if not defined $name; ## warning: $name can be "0" | |
197 | ||
198 | push @$Table, [ $start, # RANGE_START | |
199 | $end, # RANGE_END | |
200 | $name ]; # RANGE_NAME | |
201 | } | |
c3a8a2b8 | 202 | |
cf25bb62 JH |
203 | ## |
204 | ## Given a code point (and optional name), add it to the Table. | |
205 | ## | |
206 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
207 | ## | |
208 | sub Table::Append | |
209 | { | |
210 | my $Table = shift; #self | |
211 | my $codepoint = shift; | |
212 | my $name = shift; | |
213 | $name = "" if not defined $name; ## warning: $name can be "0" | |
214 | ||
215 | ## | |
216 | ## If we've already got a range working, and this code point is the next | |
217 | ## one in line, and if the name is the same, just extend the current range. | |
218 | ## | |
219 | if ($Table->NotEmpty | |
220 | and | |
221 | $Table->Max == $codepoint - 1 | |
222 | and | |
223 | $Table->[-1]->[RANGE_NAME] eq $name) | |
224 | { | |
225 | $Table->Extend($codepoint); | |
c3a8a2b8 | 226 | } |
cf25bb62 JH |
227 | else |
228 | { | |
229 | $Table->RawAppendRange($codepoint, $codepoint, $name); | |
d73e5302 | 230 | } |
cf25bb62 | 231 | } |
d73e5302 | 232 | |
cf25bb62 JH |
233 | ## |
234 | ## Given a code point range starting value and ending value (and name), | |
235 | ## Add the range to teh Table. | |
236 | ## | |
237 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
238 | ## | |
239 | sub Table::AppendRange | |
240 | { | |
241 | my $Table = shift; #self | |
242 | my $start = shift; | |
243 | my $end = shift; | |
244 | my $name = shift; | |
245 | $name = "" if not defined $name; ## warning: $name can be "0" | |
246 | ||
247 | $Table->Append($start, $name); | |
248 | $Table->Extend($end) if $end > $start; | |
249 | } | |
d73e5302 | 250 | |
cf25bb62 JH |
251 | ## |
252 | ## Return a new Table that represents all code points not in the Table. | |
253 | ## | |
254 | sub Table::Invert | |
255 | { | |
256 | my $Table = shift; #self | |
257 | ||
258 | my $New = Table->New(); | |
259 | my $max = -1; | |
260 | for my $range (@$Table) | |
261 | { | |
262 | my $start = $range->[RANGE_START]; | |
263 | my $end = $range->[RANGE_END]; | |
264 | if ($start-1 >= $max+1) { | |
265 | $New->AppendRange($max+1, $start-1, ""); | |
266 | } | |
267 | $max = $end; | |
d73e5302 | 268 | } |
cf25bb62 JH |
269 | if ($max+1 < $LastUnicodeCodepoint) { |
270 | $New->AppendRange($max+1, $LastUnicodeCodepoint); | |
d73e5302 | 271 | } |
cf25bb62 JH |
272 | return $New; |
273 | } | |
d73e5302 | 274 | |
cf25bb62 JH |
275 | ## |
276 | ## Merges any number of other tables with $self, returning the new table. | |
277 | ## (existing tables are not modified) | |
278 | ## | |
a3a8c5f0 JH |
279 | ## |
280 | ## Args may be Tables, or individual code points (as integers). | |
281 | ## | |
cf25bb62 JH |
282 | ## Can be called as either a constructor or a method. |
283 | ## | |
284 | sub Table::Merge | |
285 | { | |
286 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class | |
287 | my @Tables = @_; | |
288 | ||
289 | ## Accumulate all records from all tables | |
290 | my @Records; | |
a3a8c5f0 JH |
291 | for my $Arg (@Tables) |
292 | { | |
293 | if (ref $Arg) { | |
294 | ## arg is a table -- get its ranges | |
295 | push @Records, @$Arg; | |
296 | } else { | |
297 | ## arg is a codepoint, make a range | |
298 | push @Records, [ $Arg, $Arg ] | |
299 | } | |
d73e5302 JH |
300 | } |
301 | ||
cf25bb62 JH |
302 | ## sort by range start, with longer ranges coming first. |
303 | my ($first, @Rest) = sort { | |
304 | ($a->[RANGE_START] <=> $b->[RANGE_START]) | |
305 | or | |
306 | ($b->[RANGE_END] <=> $b->[RANGE_END]) | |
307 | } @Records; | |
308 | ||
309 | my $New = Table->New(); | |
310 | ||
311 | ## Ensuring the first range is there makes the subsequent loop easier | |
312 | $New->AppendRange($first->[RANGE_START], | |
313 | $first->[RANGE_END]); | |
314 | ||
315 | ## Fold in records so long as they add new information. | |
316 | for my $set (@Rest) | |
317 | { | |
318 | my $start = $set->[RANGE_START]; | |
319 | my $end = $set->[RANGE_END]; | |
320 | if ($start > $New->Max) { | |
321 | $New->AppendRange($start, $end); | |
322 | } elsif ($end > $New->Max) { | |
323 | $New->Extend($end); | |
324 | } | |
d73e5302 | 325 | } |
d73e5302 | 326 | |
cf25bb62 | 327 | return $New; |
d73e5302 JH |
328 | } |
329 | ||
cf25bb62 JH |
330 | ## |
331 | ## Given a filename, write a representation of the Table to a file. | |
332 | ## | |
333 | sub Table::Write | |
334 | { | |
335 | my $Table = shift; #self | |
336 | my $filename = shift; | |
d73e5302 | 337 | |
cf25bb62 | 338 | print "$filename\n" if $Verbose; |
d73e5302 | 339 | |
cf25bb62 JH |
340 | if (not open(OUT, ">$filename")) { |
341 | die "$0: can't write $filename: $!\n"; | |
342 | } | |
d73e5302 | 343 | |
cf25bb62 JH |
344 | print OUT $HEADER; |
345 | print OUT "return <<'END';\n"; | |
d73e5302 | 346 | |
cf25bb62 JH |
347 | for my $set (@$Table) |
348 | { | |
349 | my $start = $set->[RANGE_START]; | |
350 | my $end = $set->[RANGE_END]; | |
351 | my $name = $set->[RANGE_NAME]; | |
d73e5302 | 352 | |
cf25bb62 JH |
353 | if ($start == $end) { |
354 | printf OUT "%04X\t\t%s\n", $start, $name; | |
355 | } else { | |
356 | printf OUT "%04X\t%04X\t%s\n", $start, $end, $name; | |
357 | } | |
358 | } | |
d73e5302 | 359 | |
cf25bb62 JH |
360 | print OUT "END\n"; |
361 | close OUT; | |
362 | } | |
363 | ||
364 | ########################################################################### | |
365 | ########################################################################### | |
366 | ########################################################################### | |
367 | ||
368 | ||
369 | ## | |
370 | ## Called like: | |
371 | ## New_Alias(Is => 'All', SameAs => 'Any', AllowFuzzy => 1); | |
372 | ## | |
373 | ## The args must be in that order, although the AllowFuzzy pair may be omitted. | |
374 | ## | |
375 | ## This creates 'IsAll' as an alias for 'IsAny' | |
376 | ## | |
377 | sub New_Alias($$$@) | |
378 | { | |
379 | my $Type = shift; ## "Is" or "In" | |
380 | my $Alias = shift; | |
381 | my $SameAs = shift; | |
382 | my $Name = shift; | |
383 | ||
384 | ## remaining args are optional key/val | |
385 | my %Args = @_; | |
386 | ||
387 | my $AllowFuzzy = delete $Args{AllowFuzzy}; | |
388 | ||
389 | ## sanity check a few args | |
390 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { | |
391 | confess "$0: bad args to New_Alias" | |
d73e5302 JH |
392 | } |
393 | ||
cf25bb62 JH |
394 | if (not $TableInfo{$Type}->{$Name}) { |
395 | confess "$0: don't have orignial $Type => $Name to make alias" | |
396 | } | |
397 | if ($TableInfo{$Alias}) { | |
398 | confess "$0: already have original $Type => $Alias; can't make alias"; | |
d73e5302 | 399 | } |
cf25bb62 JH |
400 | $AliasInfo{$Type}->{$Name} = $Alias; |
401 | if ($AllowFuzzy) { | |
402 | $FuzzyNames{$Type}->{$Alias} = $Name; | |
403 | } | |
404 | ||
d73e5302 JH |
405 | } |
406 | ||
cf25bb62 JH |
407 | ## |
408 | ## Turn something like | |
409 | ## OLD-ITALIC | |
410 | ## to | |
411 | ## Old_Italic | |
412 | ## | |
413 | sub CanonicalName($) | |
414 | { | |
415 | my $name = lc shift; | |
416 | $name =~ s/\W+/_/; | |
417 | $name =~ s/(?<![a-z])(\w)/\u$1/g; | |
418 | return $name; | |
419 | } | |
d73e5302 | 420 | |
d73e5302 | 421 | |
cf25bb62 JH |
422 | ## All assigned code points |
423 | my $Assigned = Table->New(Is => 'Assigned', AllowFuzzy => 1); | |
d2d499f5 | 424 | |
cf25bb62 JH |
425 | my $Name = Table->New(); ## all characters, individually by name |
426 | my $General = Table->New(); ## all characters, grouped by category | |
427 | my %General; | |
428 | my %Cat; | |
d73e5302 | 429 | |
cf25bb62 JH |
430 | ## |
431 | ## Process Unicode.txt (Categories, etc.) | |
432 | ## | |
433 | sub Unicode_Txt() | |
434 | { | |
435 | my $Bidi = Table->New(); | |
436 | my $Deco = Table->New(); | |
437 | my $Comb = Table->New(); | |
438 | my $Number = Table->New(); | |
439 | my $Mirrored = Table->New(Is => 'Mirrored', AllowFuzzy => 0); | |
d73e5302 | 440 | |
cf25bb62 JH |
441 | my %DC; |
442 | my %Bidi; | |
443 | my %Deco; | |
444 | $Deco{Canon} = Table->New(Is => 'Canon', AllowFuzzy => 0); | |
445 | $Deco{Compat} = Table->New(Is => 'Compat', AllowFuzzy => 0); | |
446 | ||
447 | ## Initialize Perl-generated categories | |
448 | $Cat{Alnum} = Table->New(Is => 'Alnum', AllowFuzzy => 0); | |
449 | $Cat{Alpha} = Table->New(Is => 'Alpha', AllowFuzzy => 0); | |
450 | $Cat{ASCII} = Table->New(Is => 'ASCII', AllowFuzzy => 0); | |
451 | $Cat{Blank} = Table->New(Is => 'Blank', AllowFuzzy => 0); | |
452 | $Cat{Cntrl} = Table->New(Is => 'Cntrl', AllowFuzzy => 0); | |
453 | $Cat{Digit} = Table->New(Is => 'Digit', AllowFuzzy => 0); | |
454 | $Cat{Graph} = Table->New(Is => 'Graph', AllowFuzzy => 0); | |
455 | $Cat{Lower} = Table->New(Is => 'Lower', AllowFuzzy => 0); | |
456 | $Cat{Print} = Table->New(Is => 'Print', AllowFuzzy => 0); | |
457 | $Cat{Punct} = Table->New(Is => 'Punct', AllowFuzzy => 0); | |
458 | $Cat{SpacePerl} = Table->New(Is => 'SpacePerl', AllowFuzzy => 0); | |
459 | $Cat{Space} = Table->New(Is => 'Space', AllowFuzzy => 0); | |
460 | $Cat{Title} = Table->New(Is => 'Title', AllowFuzzy => 0); | |
461 | $Cat{Upper} = Table->New(Is => 'Upper', AllowFuzzy => 0); | |
462 | $Cat{Word} = Table->New(Is => 'Word' , AllowFuzzy => 0); | |
463 | $Cat{XDigit} = Table->New(Is => 'XDigit', AllowFuzzy => 0); | |
464 | ## Categories from Unicode.txt are auto-initialized in gencat() | |
d73e5302 | 465 | |
cf25bb62 JH |
466 | my %To; |
467 | $To{Upper} = Table->New(); | |
468 | $To{Lower} = Table->New(); | |
469 | $To{Title} = Table->New(); | |
470 | $To{Digit} = Table->New(); | |
471 | ||
472 | sub gencat($$$$) | |
473 | { | |
474 | my ($name, ## Name ("LATIN CAPITAL LETTER A") | |
475 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) | |
476 | $code, ## Code point (as an integer) | |
477 | $op) = @_; | |
478 | ||
479 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc | |
480 | ||
481 | $Assigned->$op($code); | |
482 | $Name->$op($code, $name); | |
483 | $General->$op($code, $cat); | |
484 | ||
485 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) | |
486 | $Cat{$cat} ||= Table->New(Is => $cat, AllowFuzzy => 0); | |
487 | $Cat{$cat}->$op($code); | |
488 | ||
489 | ## add to the major category (e.g. "L", "N", "C", ...) | |
490 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, AllowFuzzy => 0); | |
491 | $Cat{$MajorCat}->$op($code); | |
492 | ||
493 | ($General{$name} ||= Table->New)->$op($code, $name); | |
494 | ||
495 | # 005F: SPACING UNDERSCORE | |
496 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]/ || $code == 0x005F; | |
497 | $Cat{Alnum}->$op($code) if $cat =~ /^[LMN]/; | |
498 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; | |
499 | ||
500 | ||
501 | ||
502 | $Cat{Space}->$op($code) if $cat =~ /^Z/ | |
503 | || $code == 0x0009 # 0009: HORIZONTAL TAB | |
504 | || $code == 0x000A # 000A: LINE FEED | |
505 | || $code == 0x000B # 000B: VERTICAL TAB | |
506 | || $code == 0x000C # 000C: FORM FEED | |
507 | || $code == 0x000D; # 000D: CARRIAGE RETURN | |
508 | ||
509 | ||
510 | $Cat{SpacePerl}->$op($code) if $cat =~ /^Z/ | |
511 | || $code == 0x0009 # 0009: HORIZONTAL TAB | |
512 | || $code == 0x000A # 000A: LINE FEED | |
513 | || $code == 0x000C # 000C: FORM FEED | |
514 | || $code == 0x000D # 000D: CARRIAGE RETURN | |
515 | || $code == 0x0085 # 0085: <NEXT LINE> | |
516 | || $code == 0x2028 # 2028: LINE SEPARATOR | |
517 | || $code == 0x2029;# 2029: PARAGRAPH SEP. | |
518 | ||
519 | $Cat{Blank}->$op($code) if $cat =~ /^Z[^lp]$/ | |
520 | || $code == 0x0009 # 0009: HORIZONTAL TAB | |
521 | || $code == 0x0020; # 0020: SPACE | |
522 | ||
523 | $Cat{Digit}->$op($code) if $cat eq "Nd"; | |
524 | $Cat{Upper}->$op($code) if $cat eq "Lu"; | |
525 | $Cat{Lower}->$op($code) if $cat eq "Ll"; | |
526 | $Cat{Title}->$op($code) if $cat eq "Lt"; | |
527 | $Cat{ASCII}->$op($code) if $code <= 0x007F; | |
528 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; | |
529 | $Cat{Graph}->$op($code) if $cat =~ /^([LMNPS]|Co)/; | |
530 | $Cat{Print}->$op($code) if $cat =~ /^([LMNPS]|Co|Zs)/; | |
531 | $Cat{Punct}->$op($code) if $cat =~ /^P/; | |
532 | ||
533 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 | |
534 | || ($code >= 0x41 && $code <= 0x46) ## A..F | |
535 | || ($code >= 0x61 && $code <= 0x66); ## a..f | |
536 | } | |
d73e5302 | 537 | |
cf25bb62 JH |
538 | ## open ane read file..... |
539 | if (not open IN, "Unicode.txt") { | |
540 | die "$0: Unicode.txt: $!\n"; | |
541 | } | |
d73e5302 | 542 | |
a3a8c5f0 JH |
543 | ## |
544 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} | |
545 | ## | |
546 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") | |
547 | ||
548 | my %CodeToDeco; ## Maps code to decomp. list for chars with first | |
549 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) | |
550 | ||
551 | ## This is filled in as we go.... | |
552 | my $CombAbove = Table->New(Is => '_CombAbove', AllowFuzzy => 0); | |
553 | ||
cf25bb62 JH |
554 | while (<IN>) |
555 | { | |
556 | next unless /^[0-9A-Fa-f]+;/; | |
557 | s/\s+$//; | |
558 | ||
559 | my ($hexcode, ## code point in hex (e.g. "0041") | |
560 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") | |
561 | $cat, ## category (e.g. "Lu") | |
562 | $comb, ## Canonical combining class (e.t. "230") | |
563 | $bidi, ## directional category (e.g. "L") | |
564 | $deco, ## decomposition mapping | |
565 | $decimal, ## decimal digit value | |
566 | $digit, ## digit value | |
567 | $number, ## numeric value | |
568 | $mirrored, ## mirrored | |
569 | $unicode10, ## name in Unicode 1.0 | |
570 | $comment, ## comment field | |
571 | $upper, ## uppercase mapping | |
572 | $lower, ## lowercase mapping | |
573 | $title, ## titlecase mapping | |
574 | ) = split(/\s*;\s*/); | |
575 | ||
576 | my $code = hex($hexcode); | |
577 | ||
a3a8c5f0 JH |
578 | if ($comb and $comb == 230) { |
579 | $CombAbove->Append($code); | |
580 | $_Above_HexCodes{$hexcode} = 1; | |
581 | } | |
582 | ||
583 | ## Used in building \p{_CanonDCIJ} | |
584 | if ($deco and $deco =~ m/^006[9A]\b/) { | |
585 | $CodeToDeco{$code} = $deco; | |
586 | } | |
587 | ||
cf25bb62 JH |
588 | ## |
589 | ## There are a few pairs of lines like: | |
590 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; | |
591 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; | |
592 | ## that define ranges. | |
593 | ## | |
594 | if ($name =~ /^<(.+), (First|Last)>$/) | |
595 | { | |
596 | $name = $1; | |
597 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); | |
598 | #New_Prop(In => $name, $General{$name}, AllowFuzzy => 1); | |
599 | } | |
600 | else | |
601 | { | |
602 | ## normal (single-character) lines | |
603 | gencat($name, $cat, $code, 'Append'); | |
604 | ||
605 | # No Append() here since since several codes may map into one. | |
606 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; | |
607 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; | |
608 | $To{Title}->RawAppendRange($code, $code, $title) if $title; | |
609 | $To{Digit}->Append($code, $decimal) if length $decimal; | |
610 | ||
611 | $Bidi->Append($code, $bidi); | |
612 | $Comb->Append($code, $comb) if $comb; | |
613 | $Number->Append($code, $number) if length $number; | |
614 | ||
615 | $Mirrored->Append($code) if $mirrored eq "Y"; | |
616 | ||
617 | $Bidi{$bidi} ||= Table->New(Is => "Bidi$bidi", AllowFuzzy => 0); | |
618 | $Bidi{$bidi}->Append($code); | |
619 | ||
620 | if ($deco) | |
621 | { | |
622 | $Deco->Append($code, $deco); | |
623 | if ($deco =~/^<(\w+)>/) | |
624 | { | |
625 | $Deco{Compat}->Append($code); | |
626 | ||
627 | $DC{$1} ||= Table->New(Is => "DC$1", AllowFuzzy => 0); | |
628 | $DC{$1}->Append($code); | |
629 | } | |
630 | else | |
631 | { | |
632 | $Deco{Canon}->Append($code); | |
633 | } | |
634 | } | |
635 | } | |
636 | } | |
637 | close IN; | |
d2d499f5 | 638 | |
cf25bb62 JH |
639 | ## |
640 | ## Tidy up a few special cases.... | |
641 | ## | |
d73e5302 | 642 | |
cf25bb62 JH |
643 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist |
644 | New_Prop(Is => 'Cn', $Cat{Cn}, AllowFuzzy => 0); | |
d73e5302 | 645 | |
cf25bb62 JH |
646 | ## Unassigned is the same as 'Cn' |
647 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', AllowFuzzy => 1); | |
d73e5302 | 648 | |
cf25bb62 | 649 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C |
d73e5302 | 650 | |
d73e5302 | 651 | |
cf25bb62 JH |
652 | # L& is Ll, Lu, and Lt. |
653 | New_Prop(Is => 'L&', | |
654 | Table->Merge(@Cat{qw[Ll Lu Lt]}), | |
655 | AllowFuzzy => 0); | |
d73e5302 | 656 | |
cf25bb62 JH |
657 | ## Any and All are all code points. |
658 | my $Any = Table->New(Is => 'Any', AllowFuzzy => 1); | |
659 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); | |
d73e5302 | 660 | |
cf25bb62 | 661 | New_Alias(Is => 'All', SameAs => 'Any', AllowFuzzy => 1); |
d73e5302 | 662 | |
a3a8c5f0 JH |
663 | ## |
664 | ## Build special properties for Perl's internal case-folding needs: | |
665 | ## \p{_CaseIgnorable} | |
666 | ## \p{_CanonDCIJ} | |
667 | ## \p{_CombAbove} | |
668 | ## _CombAbove was built above. Others are built here.... | |
669 | ## | |
670 | ||
671 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] | |
672 | New_Prop(Is => '_CaseIgnorable', | |
673 | Table->Merge($Cat{Mn}, | |
674 | 0x00AD, #SOFT HYPHEN | |
675 | 0x2010), #HYPHEN | |
676 | AllowFuzzy => 0); | |
677 | ||
678 | ||
679 | ## \p{_CanonDCIJ} is fairly complex... | |
680 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', AllowFuzzy => 0); | |
681 | ## It contains the ASCII 'i' and 'j'.... | |
682 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") | |
683 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") | |
684 | ## ...and any character with a decomposition that starts with either of | |
685 | ## those code points, but only if the decomposition does not have any | |
686 | ## combining character with the "ABOVE" canonical combining class. | |
687 | for my $code (sort { $a <=> $b} keys %CodeToDeco) | |
688 | { | |
689 | ## Need to ensure that all decomposition characters do not have | |
690 | ## a %HexCodeToComb in %AboveCombClasses. | |
691 | my $want = 1; | |
692 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) | |
693 | { | |
694 | if (exists $_Above_HexCodes{$deco_hexcode}) { | |
695 | ## one of the decmposition chars has an ABOVE combination | |
696 | ## class, so we're not interested in this one | |
697 | $want = 0; | |
698 | last; | |
699 | } | |
700 | } | |
701 | if ($want) { | |
702 | $CanonCDIJ->Append($code); | |
703 | } | |
704 | } | |
705 | ||
706 | ||
d73e5302 | 707 | |
cf25bb62 JH |
708 | ## |
709 | ## Now dump the files. | |
710 | ## | |
711 | $Name->Write("Name.pl"); | |
712 | $Bidi->Write("Bidirectional.pl"); | |
713 | $Comb->Write("CombiningClass.pl"); | |
714 | $Deco->Write("Decomposition.pl"); | |
715 | $Number->Write("Number.pl"); | |
716 | $General->Write("Category.pl"); | |
717 | ||
718 | for my $to (sort keys %To) { | |
719 | $To{$to}->Write("To/$to.pl"); | |
d73e5302 JH |
720 | } |
721 | } | |
722 | ||
cf25bb62 JH |
723 | ## |
724 | ## Process LineBrk.txt | |
725 | ## | |
726 | sub LineBrk_Txt() | |
727 | { | |
728 | if (not open IN, "LineBrk.txt") { | |
729 | die "$0: LineBrk.txt: $!\n"; | |
730 | } | |
d73e5302 | 731 | |
cf25bb62 JH |
732 | my $Lbrk = Table->New(); |
733 | my %Lbrk; | |
d73e5302 | 734 | |
cf25bb62 JH |
735 | while (<IN>) |
736 | { | |
737 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; | |
d73e5302 | 738 | |
cf25bb62 | 739 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); |
d73e5302 | 740 | |
cf25bb62 | 741 | $Lbrk->Append($first, $lbrk); |
d73e5302 | 742 | |
cf25bb62 JH |
743 | $Lbrk{$lbrk} ||= Table->New(Is => "Lbrk$lbrk", AllowFuzzy => 0); |
744 | $Lbrk{$lbrk}->Append($first); | |
d73e5302 | 745 | |
cf25bb62 JH |
746 | if ($last) { |
747 | $Lbrk->Extend($last); | |
748 | $Lbrk{$lbrk}->Extend($last); | |
d73e5302 JH |
749 | } |
750 | } | |
cf25bb62 | 751 | close IN; |
d73e5302 | 752 | |
cf25bb62 JH |
753 | $Lbrk->Write("Lbrk.pl"); |
754 | } | |
d73e5302 | 755 | |
cf25bb62 JH |
756 | ## |
757 | ## Process ArabShap.txt. | |
758 | ## | |
759 | sub ArabShap_txt() | |
760 | { | |
761 | if (not open IN, "ArabShap.txt") { | |
762 | die "$0: ArabShap.txt: $!\n"; | |
763 | } | |
d73e5302 | 764 | |
cf25bb62 JH |
765 | my $ArabLink = Table->New(); |
766 | my $ArabLinkGroup = Table->New(); | |
d73e5302 | 767 | |
cf25bb62 JH |
768 | while (<IN>) |
769 | { | |
770 | next unless /^[0-9A-Fa-f]+;/; | |
771 | s/\s+$//; | |
d73e5302 | 772 | |
cf25bb62 JH |
773 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); |
774 | my $code = hex($hexcode); | |
775 | $ArabLink->Append($code, $link); | |
776 | $ArabLinkGroup->Append($code, $linkgroup); | |
d73e5302 | 777 | } |
cf25bb62 JH |
778 | close IN; |
779 | ||
780 | $ArabLink->Write("ArabLink.pl"); | |
781 | $ArabLinkGroup->Write("ArabLnkGrp.pl"); | |
d73e5302 JH |
782 | } |
783 | ||
cf25bb62 JH |
784 | ## |
785 | ## Process Jamo.txt. | |
786 | ## | |
787 | sub Jamo_txt() | |
788 | { | |
789 | if (not open IN, "Jamo.txt") { | |
790 | die "$0: Jamo.txt: $!\n"; | |
791 | } | |
792 | my $Short = Table->New(); | |
d73e5302 | 793 | |
cf25bb62 JH |
794 | while (<IN>) |
795 | { | |
796 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; | |
797 | my ($code, $short) = (hex($1), $2); | |
d73e5302 | 798 | |
cf25bb62 | 799 | $Short->Append($code, $short); |
d73e5302 | 800 | } |
cf25bb62 JH |
801 | close IN; |
802 | $Short->Write("JamoShort.pl"); | |
d73e5302 JH |
803 | } |
804 | ||
cf25bb62 JH |
805 | ## |
806 | ## Process Scripts.txt. | |
807 | ## | |
808 | sub Scripts_txt() | |
809 | { | |
810 | my @ScriptInfo; | |
d73e5302 | 811 | |
cf25bb62 JH |
812 | if (not open(IN, "Scripts.txt")) { |
813 | die "$0: Scripts.txt: $!\n"; | |
814 | } | |
815 | while (<IN>) { | |
816 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
d73e5302 | 817 | |
cf25bb62 JH |
818 | # Wait until all the scripts have been read since |
819 | # they are not listed in numeric order. | |
820 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; | |
821 | } | |
822 | close IN; | |
d73e5302 | 823 | |
cf25bb62 | 824 | # Now append the scripts properties in their code point order. |
d73e5302 | 825 | |
cf25bb62 JH |
826 | my %Script; |
827 | my $Scripts = Table->New(); | |
d73e5302 | 828 | |
cf25bb62 JH |
829 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) |
830 | { | |
831 | my ($first, $last, $name) = @$script; | |
832 | $Scripts->Append($first, $name); | |
d73e5302 | 833 | |
cf25bb62 JH |
834 | $Script{$name} ||= Table->New(Is => CanonicalName($name), |
835 | AllowFuzzy => 1); | |
836 | $Script{$name}->Append($first, $name); | |
d73e5302 | 837 | |
cf25bb62 JH |
838 | if ($last) { |
839 | $Scripts->Extend($last); | |
840 | $Script{$name}->Extend($last); | |
841 | } | |
842 | } | |
d73e5302 | 843 | |
cf25bb62 | 844 | $Scripts->Write("Scripts.pl"); |
d73e5302 | 845 | |
cf25bb62 JH |
846 | ## Common is everything not explicitly assigned to a Script |
847 | ## | |
848 | ## ***shouldn't this be intersected with \p{Assigned}? ****** | |
849 | ## | |
850 | New_Prop(Is => 'Common', $Scripts->Invert, AllowFuzzy => 1); | |
851 | } | |
d73e5302 | 852 | |
cf25bb62 JH |
853 | ## |
854 | ## Given a name like "Close Punctuation", return a regex (that when applied | |
855 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", | |
856 | ## "Close-Punctuation", etc.) | |
857 | ## | |
858 | ## Accept any space, dash, or underbar where in the official name there is | |
859 | ## space or a dash (or underbar, but there never is). | |
860 | ## | |
861 | ## | |
862 | sub NameToRegex($) | |
863 | { | |
864 | my $Name = shift; | |
865 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; | |
866 | return $Name; | |
867 | } | |
d73e5302 | 868 | |
cf25bb62 JH |
869 | ## |
870 | ## Process Blocks.txt. | |
871 | ## | |
872 | sub Blocks_txt() | |
873 | { | |
874 | my $Blocks = Table->New(); | |
875 | my %Blocks; | |
d73e5302 | 876 | |
cf25bb62 JH |
877 | if (not open IN, "Blocks.txt") { |
878 | die "$0: Blocks.txt: $!\n"; | |
879 | } | |
d73e5302 | 880 | |
cf25bb62 JH |
881 | while (<IN>) |
882 | { | |
883 | #next if not /Private Use$/; | |
884 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; | |
d73e5302 | 885 | |
cf25bb62 | 886 | my ($first, $last, $name) = (hex($1), hex($2), $3); |
d73e5302 | 887 | |
cf25bb62 | 888 | $Blocks->Append($first, $name); |
76ccdbe2 | 889 | |
cf25bb62 JH |
890 | $Blocks{$name} ||= Table->New(In=>CanonicalName($name), AllowFuzzy=>1); |
891 | $Blocks{$name}->Append($first, $name); | |
76ccdbe2 | 892 | |
cf25bb62 JH |
893 | if ($last and $last != $first) { |
894 | $Blocks->Extend($last); | |
895 | $Blocks{$name}->Extend($last); | |
d73e5302 | 896 | } |
d73e5302 | 897 | } |
cf25bb62 JH |
898 | close IN; |
899 | ||
900 | $Blocks->Write("Blocks.pl"); | |
d73e5302 JH |
901 | } |
902 | ||
cf25bb62 JH |
903 | ## |
904 | ## Read in the PropList.txt. It contains extended properties not | |
905 | ## listed in the Unicode.txt, such as 'Other_Alphabetic': | |
906 | ## alphabetic but not of the general category L; many modifiers | |
907 | ## belong to this extended property category: while they are not | |
908 | ## alphabets, they are alphabetic in nature. | |
909 | ## | |
910 | sub PropList_txt() | |
911 | { | |
912 | my @PropInfo; | |
913 | ||
914 | if (not open IN, "PropList.txt") { | |
915 | die "$0: PropList.txt: $!\n"; | |
916 | } | |
d73e5302 | 917 | |
cf25bb62 JH |
918 | while (<IN>) |
919 | { | |
920 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
d73e5302 | 921 | |
cf25bb62 JH |
922 | # Wait until all the extended properties have been read since |
923 | # they are not listed in numeric order. | |
924 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; | |
925 | } | |
926 | close IN; | |
71d929cb | 927 | |
cf25bb62 JH |
928 | # Now append the extended properties in their code point order. |
929 | my $Props = Table->New(); | |
930 | my %Prop; | |
71d929cb | 931 | |
cf25bb62 JH |
932 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) |
933 | { | |
934 | my ($first, $last, $name) = @$prop; | |
935 | $Props->Append($first, $name); | |
71d929cb | 936 | |
cf25bb62 JH |
937 | $Prop{$name} ||= Table->New(Is => $name, AllowFuzzy => 1); |
938 | $Prop{$name}->Append($first, $name); | |
71d929cb | 939 | |
cf25bb62 JH |
940 | if ($last) { |
941 | $Props->Extend($last); | |
942 | $Prop{$name}->Extend($last); | |
943 | } | |
71d929cb | 944 | } |
d73e5302 | 945 | |
cf25bb62 JH |
946 | # Alphabetic is L and Other_Alphabetic. |
947 | New_Prop(Is => 'Alphabetic', | |
948 | Table->Merge($Cat{L}, $Prop{Other_Alphabetic}), | |
949 | AllowFuzzy => 1); | |
950 | ||
951 | # Lowercase is Ll and Other_Lowercase. | |
952 | New_Prop(Is => 'Lowercase', | |
953 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), | |
954 | AllowFuzzy => 1); | |
955 | ||
956 | # Uppercase is Lu and Other_Uppercase. | |
957 | New_Prop(Is => 'Uppercase', | |
958 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), | |
959 | AllowFuzzy => 1); | |
960 | ||
961 | # Math is Sm and Other_Math. | |
962 | New_Prop(Is => 'Math', | |
963 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), | |
964 | AllowFuzzy => 1); | |
965 | ||
966 | # ID_Start is Ll, Lu, Lt, Lm, Lo, and Nl. | |
967 | New_Prop(Is => 'ID_Start', | |
968 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}), | |
969 | AllowFuzzy => 1); | |
970 | ||
971 | # ID_Continue is ID_Start, Mn, Mc, Nd, and Pc. | |
972 | New_Prop(Is => 'ID_Continue', | |
973 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}), | |
974 | AllowFuzzy => 1); | |
d73e5302 JH |
975 | } |
976 | ||
cf25bb62 JH |
977 | sub Make_GC_Aliases() |
978 | { | |
979 | ## | |
980 | ## The mapping from General Category long forms to short forms is | |
981 | ## currently hardwired here since no simple data file in the UCD | |
982 | ## seems to do that. Unicode 3.2 will assumedly correct this. | |
983 | ## | |
984 | my %Is = ( | |
d73e5302 | 985 | 'Letter' => 'L', |
e150c829 JH |
986 | 'Uppercase_Letter' => 'Lu', |
987 | 'Lowercase_Letter' => 'Ll', | |
988 | 'Titlecase_Letter' => 'Lt', | |
989 | 'Modifier_Letter' => 'Lm', | |
990 | 'Other_Letter' => 'Lo', | |
d73e5302 JH |
991 | |
992 | 'Mark' => 'M', | |
e150c829 JH |
993 | 'Non_Spacing_Mark' => 'Mn', |
994 | 'Spacing_Mark' => 'Mc', | |
995 | 'Enclosing_Mark' => 'Me', | |
d73e5302 JH |
996 | |
997 | 'Separator' => 'Z', | |
e150c829 JH |
998 | 'Space_Separator' => 'Zs', |
999 | 'Line_Separator' => 'Zl', | |
1000 | 'Paragraph_Separator' => 'Zp', | |
d73e5302 JH |
1001 | |
1002 | 'Number' => 'N', | |
e150c829 JH |
1003 | 'Decimal_Number' => 'Nd', |
1004 | 'Letter_Number' => 'Nl', | |
1005 | 'Other_Number' => 'No', | |
d73e5302 JH |
1006 | |
1007 | 'Punctuation' => 'P', | |
e150c829 JH |
1008 | 'Connector_Punctuation' => 'Pc', |
1009 | 'Dash_Punctuation' => 'Pd', | |
1010 | 'Open_Punctuation' => 'Ps', | |
1011 | 'Close_Punctuation' => 'Pe', | |
1012 | 'Initial_Punctuation' => 'Pi', | |
1013 | 'Final_Punctuation' => 'Pf', | |
1014 | 'Other_Punctuation' => 'Po', | |
d73e5302 JH |
1015 | |
1016 | 'Symbol' => 'S', | |
e150c829 JH |
1017 | 'Math_Symbol' => 'Sm', |
1018 | 'Currency_Symbol' => 'Sc', | |
1019 | 'Modifier_Symbol' => 'Sk', | |
1020 | 'Other_Symbol' => 'So', | |
d73e5302 JH |
1021 | |
1022 | 'Other' => 'C', | |
1023 | 'Control' => 'Cc', | |
1024 | 'Format' => 'Cf', | |
1025 | 'Surrogate' => 'Cs', | |
1026 | 'Private Use' => 'Co', | |
e150c829 | 1027 | 'Unassigned' => 'Cn', |
cf25bb62 | 1028 | ); |
d2d499f5 | 1029 | |
cf25bb62 JH |
1030 | ## make the aliases.... |
1031 | while (my ($Alias, $Name) = each %Is) { | |
1032 | New_Alias(Is => $Alias, SameAs => $Name, AllowFuzzy => 1); | |
1033 | } | |
1034 | } | |
d2d499f5 | 1035 | |
cf25bb62 JH |
1036 | ## |
1037 | ## Writes the info accumulated in | |
1038 | ## | |
1039 | ## %TableInfo; | |
1040 | ## %FuzzyNames; | |
1041 | ## %AliasInfo; | |
1042 | ## | |
1043 | ## | |
1044 | sub WriteAllMappings() | |
1045 | { | |
1046 | for my $Type ('In', 'Is') | |
1047 | { | |
1048 | my %Filenames; | |
1049 | my %NameToFile; | |
1050 | ||
1051 | my %Exact; ## will become %utf8::Is or %utf8::In | |
1052 | my %Pat; ## will become %utf8::IsPat or %utf8::InPat | |
1053 | ||
1054 | ## | |
1055 | ## First write all the files to the $Type/ directory | |
1056 | ## | |
1057 | while (my ($Name, $Table) = each %{$TableInfo{$Type}}) | |
1058 | { | |
1059 | ## Need an 8.3 safe filename. | |
1060 | my $filename = $Name; | |
1061 | $filename =~ s/[_\W]+(\w*)/\u$1/g; | |
1062 | substr($filename, 8) = '' if length($filename) > 8; | |
1063 | ||
1064 | ## | |
1065 | ## Make sure the filename doesn't conflict with something we | |
1066 | ## might have already written. If we have, say, | |
1067 | ## Greek_Extended1 | |
1068 | ## Greek_Extended2 | |
1069 | ## they become | |
1070 | ## Greek_Ex | |
1071 | ## Greek_E2 | |
1072 | ## | |
1073 | while (my $num = $Filenames{lc $filename}++) | |
1074 | { | |
1075 | $num++; ## so filenames with numbers start with '2', which | |
1076 | ## just looks more natural. | |
1077 | substr($filename, -length($num)) = $num; | |
1078 | } | |
1079 | ||
1080 | ## | |
1081 | ## Okay, write the file... | |
1082 | ## | |
1083 | $Exact{$Name} = $filename; | |
1084 | $Table->Write("$Type/$filename.pl"); | |
1085 | } | |
1086 | ||
1087 | ## | |
1088 | ## Build %Pat | |
1089 | ## | |
1090 | while (my ($Fuzzy, $Real) = each %{$FuzzyNames{$Type}}) | |
1091 | { | |
1092 | my $File = $Exact{$Real}; | |
1093 | ||
1094 | if (not $File) { | |
1095 | die "$0: oops [$Real]"; | |
1096 | } | |
1097 | ||
1098 | ## The prefix length of 2 is enough spread, | |
1099 | ## and besides, we have 'Yi' as an In category. | |
1100 | my $Prefix = lc(substr($Fuzzy, 0, 2)); | |
1101 | my $Regex = NameToRegex($Fuzzy); | |
1102 | ||
1103 | if ($Pat{$Prefix}->{$Regex}) { | |
1104 | warn "WHOA, conflict with /$Regex/: $Pat{$Prefix}->{$Regex} vs $File\n"; | |
1105 | } | |
1106 | ||
1107 | $Pat{$Prefix}->{$Regex} = $File; | |
1108 | } | |
1109 | ||
1110 | ## | |
1111 | ## Since the fuzzy method will provide for a way to match $Fuzzy, | |
1112 | ## there's no need for $Fuzzy to be in %Exact as well. | |
1113 | ## This can't be done in the loop above because there could be | |
1114 | ## multiple $Fuzzys pointing at the same $Real, and we don't want | |
1115 | ## the first to delete the exact mapping out from under the second. | |
1116 | ## | |
1117 | for my $Fuzzy (keys %{$FuzzyNames{$Type}}) | |
1118 | { | |
1119 | delete $Exact{$Fuzzy}; | |
1120 | } | |
1121 | ||
1122 | ||
1123 | ||
1124 | ## | |
1125 | ## Now write In.pl / Is.pl | |
1126 | ## | |
1127 | if (not open OUT, ">$Type.pl") { | |
1128 | die "$0: $Type.pl: $!\n"; | |
1129 | } | |
1130 | print OUT $HEADER; | |
1131 | print OUT "##\n"; | |
1132 | print OUT "## Data in this file used by ../utf8_heavy.pl\n"; | |
1133 | print OUT "##\n"; | |
1134 | print OUT "\n"; | |
1135 | print OUT "## Mapping from name to filename in ./$Type\n"; | |
1136 | print OUT "%utf8::$Type = (\n"; | |
1137 | for my $Name (sort keys %Exact) | |
1138 | { | |
1139 | my $File = $Exact{$Name}; | |
1140 | printf OUT " %-41s => %s,\n", "'$Name'", "'$File'"; | |
1141 | } | |
1142 | print OUT ");\n\n"; | |
1143 | ||
1144 | print OUT "## Mappings from regex to filename in ./$Type/\n"; | |
1145 | print OUT "%utf8::${Type}Pat = (\n"; | |
1146 | for my $Prefix (sort keys %Pat) | |
1147 | { | |
1148 | print OUT " '$Prefix' => {\n"; | |
1149 | while (my ($Regex, $File) = each %{ $Pat{$Prefix} }) { | |
1150 | print OUT "\t'$Regex' => '$File',\n"; | |
1151 | } | |
1152 | print OUT " },\n"; | |
1153 | } | |
1154 | print OUT ");\n"; | |
1155 | ||
1156 | close(OUT); | |
d2d499f5 | 1157 | } |
d2d499f5 JH |
1158 | } |
1159 | ||
cf25bb62 JH |
1160 | sub SpecCase_txt() |
1161 | { | |
1162 | # | |
1163 | # Read in the special cases. | |
1164 | # | |
983ffd37 | 1165 | |
cf25bb62 JH |
1166 | my %CaseInfo; |
1167 | ||
1168 | if (not open IN, "SpecCase.txt") { | |
1169 | die "$0: SpecCase.txt: $!\n"; | |
1170 | } | |
1171 | while (<IN>) { | |
1172 | next unless /^[0-9A-Fa-f]+;/; | |
1173 | s/\#.*//; | |
1174 | s/\s+$//; | |
1175 | ||
1176 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); | |
1177 | ||
1178 | if ($condition) { # not implemented yet | |
1179 | print "# SKIPPING $_\n" if $Verbose; | |
1180 | next; | |
1181 | } | |
1182 | ||
1183 | # Wait until all the special cases have been read since | |
1184 | # they are not listed in numeric order. | |
1185 | my $ix = hex($code); | |
1186 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ]; | |
1187 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ]; | |
1188 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ]; | |
1189 | } | |
1190 | close IN; | |
1191 | ||
1192 | # Now write out the special cases properties in their code point order. | |
1193 | # Prepend them to the To/{Upper,Lower,Title}.pl. | |
1194 | ||
1195 | for my $case (qw(Lower Title Upper)) | |
1196 | { | |
1197 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; | |
1198 | if (not open OUT, ">To/$case.pl") { | |
1199 | die "$0: To/$case.txt: $!"; | |
1200 | } | |
1201 | ||
1202 | print OUT $HEADER, "\n"; | |
1203 | print OUT "%utf8::ToSpec$case =\n(\n"; | |
1204 | ||
1205 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { | |
1206 | my ($ix, $code, $to) = @$prop; | |
1207 | my $tostr = | |
1208 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; | |
1209 | printf OUT qq['%04X' => "$tostr",\n], $ix; | |
1210 | } | |
1211 | print OUT ");\n\n"; | |
1212 | print OUT "return <<'END';\n"; | |
1213 | print OUT $NormalCase; | |
1214 | print OUT "END\n"; | |
1215 | close OUT; | |
d2d499f5 | 1216 | } |
d2d499f5 JH |
1217 | } |
1218 | ||
c4051cc5 JH |
1219 | # |
1220 | # Read in the case foldings. | |
1221 | # | |
1222 | # We will do full case folding, C + F + I (see CaseFold.txt). | |
1223 | # | |
cf25bb62 JH |
1224 | sub CaseFold_txt() |
1225 | { | |
1226 | if (not open IN, "CaseFold.txt") { | |
1227 | die "$0: To/Fold.pl: $!\n"; | |
1228 | } | |
c4051cc5 | 1229 | |
cf25bb62 | 1230 | my $Fold = Table->New(); |
c4051cc5 JH |
1231 | my %Fold; |
1232 | ||
cf25bb62 | 1233 | while (<IN>) { |
254ba52a | 1234 | # Skip status 'S', simple case folding |
c4051cc5 JH |
1235 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; |
1236 | ||
cf25bb62 | 1237 | my ($code, $status, $fold) = (hex($1), $2, $3); |
c4051cc5 JH |
1238 | |
1239 | if ($status eq 'C') { # Common: one-to-one folding | |
254ba52a | 1240 | # No append() since several codes may fold into one. |
cf25bb62 | 1241 | $Fold->RawAppendRange($code, $code, $fold); |
c4051cc5 | 1242 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I |
cf25bb62 | 1243 | $Fold{$code} = $fold; |
c4051cc5 JH |
1244 | } |
1245 | } | |
cf25bb62 | 1246 | close IN; |
c4051cc5 | 1247 | |
cf25bb62 | 1248 | $Fold->Write("To/Fold.pl"); |
c4051cc5 JH |
1249 | |
1250 | # | |
1251 | # Prepend the special foldings to the common foldings. | |
1252 | # | |
1253 | ||
1254 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; | |
cf25bb62 JH |
1255 | if (not open OUT, ">To/Fold.pl") { |
1256 | die "$0: To/Fold.pl: $!\n"; | |
1257 | } | |
1258 | print OUT $HEADER, "\n"; | |
1259 | print OUT "%utf8::ToSpecFold =\n(\n"; | |
1260 | for my $code (sort { $a <=> $b } keys %Fold) { | |
1261 | my $foldstr = | |
1262 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; | |
1263 | printf OUT qq['%04X' => "$foldstr",\n], $code; | |
c4051cc5 | 1264 | } |
cf25bb62 JH |
1265 | print OUT ");\n\n"; |
1266 | print OUT "return <<'END';\n"; | |
1267 | print OUT $CommonFold; | |
1268 | print OUT "END\n"; | |
1269 | close OUT; | |
c4051cc5 JH |
1270 | } |
1271 | ||
cf25bb62 JH |
1272 | ## Do it.... |
1273 | ||
1274 | Unicode_Txt(); | |
1275 | Make_GC_Aliases(); | |
1276 | PropList_txt(); | |
1277 | ||
1278 | Scripts_txt(); | |
1279 | Blocks_txt(); | |
1280 | ||
1281 | LineBrk_Txt(); | |
1282 | ArabShap_txt(); | |
1283 | Jamo_txt(); | |
1284 | SpecCase_txt(); | |
1285 | ||
1286 | WriteAllMappings(); | |
1287 | ||
1288 | CaseFold_txt(); | |
1289 | ||
d73e5302 JH |
1290 | # That's all, folks! |
1291 | ||
cf25bb62 | 1292 | __END__ |