Commit | Line | Data |
d73e5302 |
1 | #!/usr/bin/perl -w |
4392c1cd |
2 | require 5.008; # Needs pack "U". Probably safest to run on 5.8.x |
d73e5302 |
3 | use strict; |
cf25bb62 |
4 | use Carp; |
5beb625e |
5 | |
cf25bb62 |
6 | ## |
7 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) |
8 | ## from the Unicode database files (lib/unicore/*.txt). |
9 | ## |
d2d499f5 |
10 | |
5beb625e |
11 | mkdir("lib", 0755); |
12ac2576 |
12 | mkdir("lib/gc_sc", 0755); |
5beb625e |
13 | mkdir("To", 0755); |
d73e5302 |
14 | |
cf25bb62 |
15 | ## |
16 | ## Process any args. |
17 | ## |
5beb625e |
18 | my $Verbose = 0; |
19 | my $MakeTestScript = 0; |
cf25bb62 |
20 | |
21 | while (@ARGV) |
22 | { |
23 | my $arg = shift @ARGV; |
24 | if ($arg eq '-v') { |
25 | $Verbose = 1; |
26 | } elsif ($arg eq '-q') { |
27 | $Verbose = 0; |
5beb625e |
28 | } elsif ($arg eq '-maketest') { |
29 | $MakeTestScript = 1; |
d73e5302 |
30 | } else { |
5beb625e |
31 | die "usage: $0 [-v|-q] [-maketest]"; |
d73e5302 |
32 | } |
33 | } |
34 | |
cf25bb62 |
35 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 3.1.1. |
d73e5302 |
36 | |
cf25bb62 |
37 | my $HEADER=<<"EOF"; |
d73e5302 |
38 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! |
551b6b6f |
39 | # This file is built by $0 from e.g. UnicodeData.txt. |
d73e5302 |
40 | # Any changes made here will be lost! |
cf25bb62 |
41 | |
42 | EOF |
43 | |
5beb625e |
44 | |
45 | ## |
46 | ## Given a filename and a reference to an array of lines, |
47 | ## write the lines to the file only if the contents have not changed. |
48 | ## |
49 | sub WriteIfChanged($\@) |
50 | { |
51 | my $file = shift; |
52 | my $lines = shift; |
53 | |
54 | my $TextToWrite = join '', @$lines; |
55 | if (open IN, $file) { |
56 | local($/) = undef; |
57 | my $PreviousText = <IN>; |
58 | close IN; |
59 | if ($PreviousText eq $TextToWrite) { |
60 | print "$file unchanged.\n" if $Verbose; |
61 | return; |
62 | } |
63 | } |
64 | if (not open OUT, ">$file") { |
65 | die "$0: can't open $file for output: $!\n"; |
66 | } |
67 | print "$file written.\n" if $Verbose; |
68 | |
69 | print OUT $TextToWrite; |
70 | close OUT; |
71 | } |
72 | |
cf25bb62 |
73 | ## |
74 | ## The main datastructure (a "Table") represents a set of code points that |
75 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, |
76 | ## etc.). They are kept as ranges of code points (starting and ending of |
77 | ## each range). |
78 | ## |
79 | ## For example, a range ASCII LETTERS would be represented as: |
80 | ## [ [ 0x41 => 0x5A, 'UPPER' ], |
81 | ## [ 0x61 => 0x7A, 'LOWER, ] ] |
82 | ## |
83 | sub RANGE_START() { 0 } ## index into range element |
84 | sub RANGE_END() { 1 } ## index into range element |
85 | sub RANGE_NAME() { 2 } ## index into range element |
86 | |
99598c8c |
87 | ## Conceptually, these should really be folded into the 'Table' objects |
cf25bb62 |
88 | my %TableInfo; |
99598c8c |
89 | my %TableDesc; |
cf25bb62 |
90 | my %FuzzyNames; |
91 | my %AliasInfo; |
5beb625e |
92 | my %CanonicalToOrig; |
cf25bb62 |
93 | |
94 | ## |
99598c8c |
95 | ## Turn something like |
96 | ## OLD-ITALIC |
5beb625e |
97 | ## into |
99598c8c |
98 | ## OldItalic |
99 | ## |
100 | sub CanonicalName($) |
101 | { |
5beb625e |
102 | my $orig = shift; |
103 | my $name = lc $orig; |
99598c8c |
104 | $name =~ s/(?<![a-z])(\w)/\u$1/g; |
5beb625e |
105 | $name =~ s/[-_\s]+//g; |
99598c8c |
106 | |
5beb625e |
107 | $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name}; |
99598c8c |
108 | return $name; |
109 | } |
110 | |
12ac2576 |
111 | |
112 | ## |
113 | ## Store the alias definitions for later use. |
114 | ## |
115 | my %PropertyAlias; |
116 | my %PropValueAlias; |
117 | |
118 | my %PA_reverse; |
119 | my %PVA_reverse; |
120 | |
121 | sub Build_Aliases() |
122 | { |
123 | ## |
124 | ## Most of the work with aliases doesn't occur here, |
125 | ## but rather in utf8_heavy.pl, which uses utf8_pva.pl, |
126 | ## which contains just this function. However, this one |
127 | ## |
128 | ## -- japhy (2004/04/13) |
129 | |
130 | open PA, "< PropertyAliases.txt" |
131 | or confess "Can't open PropertyAliases.txt: $!"; |
132 | while (<PA>) { |
133 | s/#.*//; |
134 | s/\s+$//; |
135 | next if /^$/; |
136 | |
137 | my ($abbrev, $name) = split /\s*;\s*/; |
138 | next if $abbrev eq "n/a"; |
139 | $PropertyAlias{$abbrev} = $name; |
140 | $PA_reverse{$name} = $abbrev; |
141 | } |
142 | close PA; |
143 | |
144 | open PVA, "< PropValueAliases.txt" |
145 | or confess "Can't open PropValueAliases.txt: $!"; |
146 | while (<PVA>) { |
147 | s/#.*//; |
148 | s/\s+$//; |
149 | next if /^$/; |
150 | |
151 | my ($prop, @data) = split /\s*;\s*/; |
152 | |
153 | if ($prop eq 'ccc') { |
154 | $PropValueAlias{$prop}{$data[1]} = [ @data[0,2] ]; |
155 | $PVA_reverse{$prop}{$data[2]} = [ @data[0,1] ]; |
156 | } |
157 | else { |
158 | next if $data[0] eq "n/a"; |
159 | $PropValueAlias{$prop}{$data[0]} = $data[1]; |
160 | $PVA_reverse{$prop}{$data[1]} = $data[0]; |
161 | } |
162 | } |
163 | close PVA; |
164 | } |
165 | |
166 | |
99598c8c |
167 | ## |
cf25bb62 |
168 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. |
169 | ## |
170 | ## Called like: |
99598c8c |
171 | ## New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1); |
cf25bb62 |
172 | ## |
173 | ## Normally, these parameters are set when the Table is created (when the |
174 | ## Table->New constructor is called), but there are times when it needs to |
175 | ## be done after-the-fact...) |
176 | ## |
177 | sub New_Prop($$$@) |
178 | { |
179 | my $Type = shift; ## "Is" or "In"; |
180 | my $Name = shift; |
181 | my $Table = shift; |
182 | |
183 | ## remaining args are optional key/val |
184 | my %Args = @_; |
185 | |
99598c8c |
186 | my $Fuzzy = delete $Args{Fuzzy}; |
187 | my $Desc = delete $Args{Desc}; # description |
188 | |
5beb625e |
189 | $Name = CanonicalName($Name) if $Fuzzy; |
cf25bb62 |
190 | |
191 | ## sanity check a few args |
192 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { |
193 | confess "$0: bad args to New_Prop" |
194 | } |
195 | |
196 | if (not $TableInfo{$Type}->{$Name}) |
197 | { |
198 | $TableInfo{$Type}->{$Name} = $Table; |
99598c8c |
199 | $TableDesc{$Type}->{$Name} = $Desc; |
200 | if ($Fuzzy) { |
cf25bb62 |
201 | $FuzzyNames{$Type}->{$Name} = $Name; |
202 | } |
203 | } |
d73e5302 |
204 | } |
205 | |
d73e5302 |
206 | |
cf25bb62 |
207 | ## |
208 | ## Creates a new Table object. |
209 | ## |
210 | ## Args are key/value pairs: |
99598c8c |
211 | ## In => Name -- Name of "In" property to be associated with |
212 | ## Is => Name -- Name of "Is" property to be associated with |
213 | ## Fuzzy => Boolean -- True if name can be accessed "fuzzily" |
214 | ## Desc => String -- Description of the property |
cf25bb62 |
215 | ## |
216 | ## No args are required. |
217 | ## |
218 | sub Table::New |
219 | { |
220 | my $class = shift; |
221 | my %Args = @_; |
222 | |
223 | my $Table = bless [], $class; |
224 | |
99598c8c |
225 | my $Fuzzy = delete $Args{Fuzzy}; |
226 | my $Desc = delete $Args{Desc}; |
cf25bb62 |
227 | |
228 | for my $Type ('Is', 'In') |
229 | { |
230 | if (my $Name = delete $Args{$Type}) { |
99598c8c |
231 | New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy); |
cf25bb62 |
232 | } |
233 | } |
234 | |
235 | ## shouldn't have any left over |
236 | if (%Args) { |
237 | confess "$0: bad args to Table->New" |
238 | } |
239 | |
240 | return $Table; |
d73e5302 |
241 | } |
242 | |
cf25bb62 |
243 | ## |
244 | ## Returns true if the Table has no code points |
245 | ## |
246 | sub Table::IsEmpty |
247 | { |
248 | my $Table = shift; #self |
249 | return not @$Table; |
d73e5302 |
250 | } |
251 | |
cf25bb62 |
252 | ## |
253 | ## Returns true if the Table has code points |
254 | ## |
255 | sub Table::NotEmpty |
256 | { |
257 | my $Table = shift; #self |
258 | return @$Table; |
d73e5302 |
259 | } |
260 | |
cf25bb62 |
261 | ## |
262 | ## Returns the maximum code point currently in the table. |
263 | ## |
264 | sub Table::Max |
265 | { |
266 | my $Table = shift; #self |
267 | confess "oops" if $Table->IsEmpty; ## must have code points to have a max |
268 | return $Table->[-1]->[RANGE_END]; |
269 | } |
d73e5302 |
270 | |
cf25bb62 |
271 | ## |
272 | ## Replaces the codepoints in the Table with those in the Table given |
273 | ## as an arg. (NOTE: this is not a "deep copy"). |
274 | ## |
275 | sub Table::Replace($$) |
276 | { |
277 | my $Table = shift; #self |
278 | my $New = shift; |
d73e5302 |
279 | |
cf25bb62 |
280 | @$Table = @$New; |
281 | } |
71d929cb |
282 | |
cf25bb62 |
283 | ## |
284 | ## Given a new code point, make the last range of the Table extend to |
285 | ## include the new (and all intervening) code points. |
286 | ## |
287 | sub Table::Extend |
288 | { |
289 | my $Table = shift; #self |
290 | my $codepoint = shift; |
d73e5302 |
291 | |
cf25bb62 |
292 | my $PrevMax = $Table->Max; |
e904f995 |
293 | |
cf25bb62 |
294 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; |
e904f995 |
295 | |
cf25bb62 |
296 | $Table->[-1]->[RANGE_END] = $codepoint; |
297 | } |
c3a8a2b8 |
298 | |
cf25bb62 |
299 | ## |
300 | ## Given a code point range start and end (and optional name), blindly |
301 | ## append them to the list of ranges for the Table. |
302 | ## |
303 | ## NOTE: Code points must be added in strictly ascending numeric order. |
304 | ## |
305 | sub Table::RawAppendRange |
306 | { |
307 | my $Table = shift; #self |
308 | my $start = shift; |
309 | my $end = shift; |
310 | my $name = shift; |
311 | $name = "" if not defined $name; ## warning: $name can be "0" |
312 | |
313 | push @$Table, [ $start, # RANGE_START |
314 | $end, # RANGE_END |
315 | $name ]; # RANGE_NAME |
316 | } |
c3a8a2b8 |
317 | |
cf25bb62 |
318 | ## |
319 | ## Given a code point (and optional name), add it to the Table. |
320 | ## |
321 | ## NOTE: Code points must be added in strictly ascending numeric order. |
322 | ## |
323 | sub Table::Append |
324 | { |
325 | my $Table = shift; #self |
326 | my $codepoint = shift; |
327 | my $name = shift; |
328 | $name = "" if not defined $name; ## warning: $name can be "0" |
329 | |
330 | ## |
331 | ## If we've already got a range working, and this code point is the next |
332 | ## one in line, and if the name is the same, just extend the current range. |
333 | ## |
334 | if ($Table->NotEmpty |
335 | and |
336 | $Table->Max == $codepoint - 1 |
337 | and |
338 | $Table->[-1]->[RANGE_NAME] eq $name) |
339 | { |
340 | $Table->Extend($codepoint); |
c3a8a2b8 |
341 | } |
cf25bb62 |
342 | else |
343 | { |
344 | $Table->RawAppendRange($codepoint, $codepoint, $name); |
d73e5302 |
345 | } |
cf25bb62 |
346 | } |
d73e5302 |
347 | |
cf25bb62 |
348 | ## |
349 | ## Given a code point range starting value and ending value (and name), |
350 | ## Add the range to teh Table. |
351 | ## |
352 | ## NOTE: Code points must be added in strictly ascending numeric order. |
353 | ## |
354 | sub Table::AppendRange |
355 | { |
356 | my $Table = shift; #self |
357 | my $start = shift; |
358 | my $end = shift; |
359 | my $name = shift; |
360 | $name = "" if not defined $name; ## warning: $name can be "0" |
361 | |
362 | $Table->Append($start, $name); |
363 | $Table->Extend($end) if $end > $start; |
364 | } |
d73e5302 |
365 | |
cf25bb62 |
366 | ## |
367 | ## Return a new Table that represents all code points not in the Table. |
368 | ## |
369 | sub Table::Invert |
370 | { |
371 | my $Table = shift; #self |
372 | |
373 | my $New = Table->New(); |
374 | my $max = -1; |
375 | for my $range (@$Table) |
376 | { |
377 | my $start = $range->[RANGE_START]; |
378 | my $end = $range->[RANGE_END]; |
379 | if ($start-1 >= $max+1) { |
380 | $New->AppendRange($max+1, $start-1, ""); |
381 | } |
382 | $max = $end; |
d73e5302 |
383 | } |
cf25bb62 |
384 | if ($max+1 < $LastUnicodeCodepoint) { |
385 | $New->AppendRange($max+1, $LastUnicodeCodepoint); |
d73e5302 |
386 | } |
cf25bb62 |
387 | return $New; |
388 | } |
d73e5302 |
389 | |
cf25bb62 |
390 | ## |
391 | ## Merges any number of other tables with $self, returning the new table. |
392 | ## (existing tables are not modified) |
393 | ## |
a3a8c5f0 |
394 | ## |
395 | ## Args may be Tables, or individual code points (as integers). |
396 | ## |
cf25bb62 |
397 | ## Can be called as either a constructor or a method. |
398 | ## |
399 | sub Table::Merge |
400 | { |
401 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class |
402 | my @Tables = @_; |
403 | |
404 | ## Accumulate all records from all tables |
405 | my @Records; |
a3a8c5f0 |
406 | for my $Arg (@Tables) |
407 | { |
408 | if (ref $Arg) { |
409 | ## arg is a table -- get its ranges |
410 | push @Records, @$Arg; |
411 | } else { |
412 | ## arg is a codepoint, make a range |
413 | push @Records, [ $Arg, $Arg ] |
414 | } |
d73e5302 |
415 | } |
416 | |
cf25bb62 |
417 | ## sort by range start, with longer ranges coming first. |
418 | my ($first, @Rest) = sort { |
419 | ($a->[RANGE_START] <=> $b->[RANGE_START]) |
420 | or |
421 | ($b->[RANGE_END] <=> $b->[RANGE_END]) |
422 | } @Records; |
423 | |
424 | my $New = Table->New(); |
425 | |
426 | ## Ensuring the first range is there makes the subsequent loop easier |
427 | $New->AppendRange($first->[RANGE_START], |
428 | $first->[RANGE_END]); |
429 | |
430 | ## Fold in records so long as they add new information. |
431 | for my $set (@Rest) |
432 | { |
433 | my $start = $set->[RANGE_START]; |
434 | my $end = $set->[RANGE_END]; |
435 | if ($start > $New->Max) { |
436 | $New->AppendRange($start, $end); |
437 | } elsif ($end > $New->Max) { |
438 | $New->Extend($end); |
439 | } |
d73e5302 |
440 | } |
d73e5302 |
441 | |
cf25bb62 |
442 | return $New; |
d73e5302 |
443 | } |
444 | |
cf25bb62 |
445 | ## |
446 | ## Given a filename, write a representation of the Table to a file. |
99598c8c |
447 | ## May have an optional comment as a 2nd arg. |
cf25bb62 |
448 | ## |
449 | sub Table::Write |
450 | { |
99598c8c |
451 | my $Table = shift; #self |
cf25bb62 |
452 | my $filename = shift; |
99598c8c |
453 | my $comment = shift; |
d73e5302 |
454 | |
5beb625e |
455 | my @OUT = $HEADER; |
99598c8c |
456 | if (defined $comment) { |
457 | $comment =~ s/\s+\Z//; |
458 | $comment =~ s/^/# /gm; |
5beb625e |
459 | push @OUT, "#\n$comment\n#\n"; |
99598c8c |
460 | } |
5beb625e |
461 | push @OUT, "return <<'END';\n"; |
d73e5302 |
462 | |
cf25bb62 |
463 | for my $set (@$Table) |
464 | { |
465 | my $start = $set->[RANGE_START]; |
466 | my $end = $set->[RANGE_END]; |
467 | my $name = $set->[RANGE_NAME]; |
d73e5302 |
468 | |
cf25bb62 |
469 | if ($start == $end) { |
5beb625e |
470 | push @OUT, sprintf "%04X\t\t%s\n", $start, $name; |
cf25bb62 |
471 | } else { |
5beb625e |
472 | push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name; |
cf25bb62 |
473 | } |
474 | } |
d73e5302 |
475 | |
5beb625e |
476 | push @OUT, "END\n"; |
477 | |
478 | WriteIfChanged($filename, @OUT); |
479 | } |
480 | |
481 | ## This used only for making the test script. |
482 | ## helper function |
483 | sub IsUsable($) |
484 | { |
485 | my $code = shift; |
486 | return 0 if $code <= 0x0000; ## don't use null |
487 | return 0 if $code >= $LastUnicodeCodepoint; ## keep in range |
488 | return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates |
489 | return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good |
490 | return 0 if (($code & 0xFFFF) == 0xFFFE); ## utf8.c says no good |
491 | return 0 if (($code & 0xFFFF) == 0xFFFF); ## utf8.c says no good |
492 | return 1; |
493 | } |
494 | |
495 | ## Return a code point that's part of the table. |
496 | ## Returns nothing if the table is empty (or covers only surrogates). |
497 | ## This used only for making the test script. |
498 | sub Table::ValidCode |
499 | { |
500 | my $Table = shift; #self |
501 | for my $set (@$Table) { |
502 | return $set->[RANGE_END] if IsUsable($set->[RANGE_END]); |
503 | } |
504 | return (); |
505 | } |
506 | |
507 | ## Return a code point that's not part of the table |
508 | ## Returns nothing if the table covers all code points. |
509 | ## This used only for making the test script. |
510 | sub Table::InvalidCode |
511 | { |
512 | my $Table = shift; #self |
513 | |
514 | return 0x1234 if $Table->IsEmpty(); |
515 | |
516 | for my $set (@$Table) |
517 | { |
518 | if (IsUsable($set->[RANGE_END] + 1)) |
519 | { |
520 | return $set->[RANGE_END] + 1; |
521 | } |
522 | |
523 | if (IsUsable($set->[RANGE_START] - 1)) |
524 | { |
525 | return $set->[RANGE_START] - 1; |
526 | } |
527 | } |
528 | return (); |
cf25bb62 |
529 | } |
530 | |
531 | ########################################################################### |
532 | ########################################################################### |
533 | ########################################################################### |
534 | |
535 | |
536 | ## |
537 | ## Called like: |
99598c8c |
538 | ## New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1); |
cf25bb62 |
539 | ## |
99598c8c |
540 | ## The args must be in that order, although the Fuzzy pair may be omitted. |
cf25bb62 |
541 | ## |
542 | ## This creates 'IsAll' as an alias for 'IsAny' |
543 | ## |
544 | sub New_Alias($$$@) |
545 | { |
546 | my $Type = shift; ## "Is" or "In" |
547 | my $Alias = shift; |
99598c8c |
548 | my $SameAs = shift; # expecting "SameAs" -- just ignored |
cf25bb62 |
549 | my $Name = shift; |
550 | |
551 | ## remaining args are optional key/val |
552 | my %Args = @_; |
553 | |
99598c8c |
554 | my $Fuzzy = delete $Args{Fuzzy}; |
cf25bb62 |
555 | |
556 | ## sanity check a few args |
557 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { |
558 | confess "$0: bad args to New_Alias" |
d73e5302 |
559 | } |
560 | |
5beb625e |
561 | $Alias = CanonicalName($Alias) if $Fuzzy; |
562 | |
563 | if (not $TableInfo{$Type}->{$Name}) |
564 | { |
565 | my $CName = CanonicalName($Name); |
566 | if ($TableInfo{$Type}->{$CName}) { |
567 | confess "$0: Use canonical form '$CName' instead of '$Name' for alias."; |
568 | } else { |
12ac2576 |
569 | confess "$0: don't have original $Type => $Name to make alias\n"; |
5beb625e |
570 | } |
cf25bb62 |
571 | } |
572 | if ($TableInfo{$Alias}) { |
573 | confess "$0: already have original $Type => $Alias; can't make alias"; |
d73e5302 |
574 | } |
cf25bb62 |
575 | $AliasInfo{$Type}->{$Name} = $Alias; |
99598c8c |
576 | if ($Fuzzy) { |
cf25bb62 |
577 | $FuzzyNames{$Type}->{$Alias} = $Name; |
578 | } |
579 | |
d73e5302 |
580 | } |
581 | |
d73e5302 |
582 | |
cf25bb62 |
583 | ## All assigned code points |
99598c8c |
584 | my $Assigned = Table->New(Is => 'Assigned', |
585 | Desc => "All assigned code points", |
5beb625e |
586 | Fuzzy => 0); |
d2d499f5 |
587 | |
cf25bb62 |
588 | my $Name = Table->New(); ## all characters, individually by name |
589 | my $General = Table->New(); ## all characters, grouped by category |
590 | my %General; |
591 | my %Cat; |
d73e5302 |
592 | |
cf25bb62 |
593 | ## |
551b6b6f |
594 | ## Process UnicodeData.txt (Categories, etc.) |
cf25bb62 |
595 | ## |
44da8cae |
596 | sub UnicodeData_Txt() |
cf25bb62 |
597 | { |
598 | my $Bidi = Table->New(); |
599 | my $Deco = Table->New(); |
600 | my $Comb = Table->New(); |
601 | my $Number = Table->New(); |
12ac2576 |
602 | my $Mirrored = Table->New();#Is => 'Mirrored', |
603 | #Desc => "Mirrored in bidirectional text", |
604 | #Fuzzy => 0); |
d73e5302 |
605 | |
cf25bb62 |
606 | my %DC; |
607 | my %Bidi; |
12ac2576 |
608 | my %Number; |
609 | $DC{can} = Table->New(); |
610 | $DC{com} = Table->New(); |
cf25bb62 |
611 | |
612 | ## Initialize Perl-generated categories |
551b6b6f |
613 | ## (Categories from UnicodeData.txt are auto-initialized in gencat) |
44da8cae |
614 | $Cat{Alnum} = |
615 | Table->New(Is => 'Alnum', Desc => "[[:Alnum:]]", Fuzzy => 0); |
616 | $Cat{Alpha} = |
617 | Table->New(Is => 'Alpha', Desc => "[[:Alpha:]]", Fuzzy => 0); |
618 | $Cat{ASCII} = |
619 | Table->New(Is => 'ASCII', Desc => "[[:ASCII:]]", Fuzzy => 0); |
620 | $Cat{Blank} = |
621 | Table->New(Is => 'Blank', Desc => "[[:Blank:]]", Fuzzy => 0); |
622 | $Cat{Cntrl} = |
623 | Table->New(Is => 'Cntrl', Desc => "[[:Cntrl:]]", Fuzzy => 0); |
624 | $Cat{Digit} = |
625 | Table->New(Is => 'Digit', Desc => "[[:Digit:]]", Fuzzy => 0); |
626 | $Cat{Graph} = |
627 | Table->New(Is => 'Graph', Desc => "[[:Graph:]]", Fuzzy => 0); |
628 | $Cat{Lower} = |
629 | Table->New(Is => 'Lower', Desc => "[[:Lower:]]", Fuzzy => 0); |
630 | $Cat{Print} = |
631 | Table->New(Is => 'Print', Desc => "[[:Print:]]", Fuzzy => 0); |
632 | $Cat{Punct} = |
633 | Table->New(Is => 'Punct', Desc => "[[:Punct:]]", Fuzzy => 0); |
634 | $Cat{Space} = |
635 | Table->New(Is => 'Space', Desc => "[[:Space:]]", Fuzzy => 0); |
636 | $Cat{Title} = |
637 | Table->New(Is => 'Title', Desc => "[[:Title:]]", Fuzzy => 0); |
638 | $Cat{Upper} = |
639 | Table->New(Is => 'Upper', Desc => "[[:Upper:]]", Fuzzy => 0); |
640 | $Cat{XDigit} = |
641 | Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0); |
642 | $Cat{Word} = |
643 | Table->New(Is => 'Word', Desc => "[[:Word:]]", Fuzzy => 0); |
644 | $Cat{SpacePerl} = |
645 | Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0); |
d73e5302 |
646 | |
cf25bb62 |
647 | my %To; |
648 | $To{Upper} = Table->New(); |
649 | $To{Lower} = Table->New(); |
650 | $To{Title} = Table->New(); |
651 | $To{Digit} = Table->New(); |
652 | |
653 | sub gencat($$$$) |
654 | { |
655 | my ($name, ## Name ("LATIN CAPITAL LETTER A") |
656 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) |
657 | $code, ## Code point (as an integer) |
658 | $op) = @_; |
659 | |
660 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc |
661 | |
662 | $Assigned->$op($code); |
663 | $Name->$op($code, $name); |
664 | $General->$op($code, $cat); |
665 | |
666 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) |
99598c8c |
667 | $Cat{$cat} ||= Table->New(Is => $cat, |
668 | Desc => "General Category '$cat'", |
669 | Fuzzy => 0); |
cf25bb62 |
670 | $Cat{$cat}->$op($code); |
671 | |
672 | ## add to the major category (e.g. "L", "N", "C", ...) |
99598c8c |
673 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, |
674 | Desc => "Major Category '$MajorCat'", |
675 | Fuzzy => 0); |
cf25bb62 |
676 | $Cat{$MajorCat}->$op($code); |
677 | |
678 | ($General{$name} ||= Table->New)->$op($code, $name); |
679 | |
680 | # 005F: SPACING UNDERSCORE |
d75d706f |
681 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]|Pc/; |
c65e4d19 |
682 | $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/; |
cf25bb62 |
683 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; |
684 | |
d75d706f |
685 | my $isspace = |
686 | ($cat =~ /Zs|Zl|Zp/ && |
687 | $code != 0x200B) # 200B is ZWSP which is for line break control |
688 | # and therefore it is not part of "space" even while it is "Zs". |
cf25bb62 |
689 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
690 | || $code == 0x000A # 000A: LINE FEED |
691 | || $code == 0x000B # 000B: VERTICAL TAB |
692 | || $code == 0x000C # 000C: FORM FEED |
e6f58734 |
693 | || $code == 0x000D # 000D: CARRIAGE RETURN |
d75d706f |
694 | || $code == 0x0085 # 0085: NEL |
695 | |
696 | ; |
cf25bb62 |
697 | |
d75d706f |
698 | $Cat{Space}->$op($code) if $isspace; |
cf25bb62 |
699 | |
d75d706f |
700 | $Cat{SpacePerl}->$op($code) if $isspace |
701 | && $code != 0x000B; # Backward compat. |
cf25bb62 |
702 | |
d75d706f |
703 | $Cat{Blank}->$op($code) if $isspace |
704 | && !($code == 0x000A || |
705 | $code == 0x000B || |
706 | $code == 0x000C || |
707 | $code == 0x000D || |
708 | $code == 0x0085 || |
709 | $cat =~ /^Z[lp]/); |
cf25bb62 |
710 | |
711 | $Cat{Digit}->$op($code) if $cat eq "Nd"; |
712 | $Cat{Upper}->$op($code) if $cat eq "Lu"; |
713 | $Cat{Lower}->$op($code) if $cat eq "Ll"; |
714 | $Cat{Title}->$op($code) if $cat eq "Lt"; |
715 | $Cat{ASCII}->$op($code) if $code <= 0x007F; |
716 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; |
d75d706f |
717 | my $isgraph = !$isspace && $cat !~ /Cc|Cs|Cn/; |
718 | $Cat{Graph}->$op($code) if $isgraph; |
719 | $Cat{Print}->$op($code) if $isgraph || $isspace; |
cf25bb62 |
720 | $Cat{Punct}->$op($code) if $cat =~ /^P/; |
721 | |
722 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 |
723 | || ($code >= 0x41 && $code <= 0x46) ## A..F |
724 | || ($code >= 0x61 && $code <= 0x66); ## a..f |
725 | } |
d73e5302 |
726 | |
cf25bb62 |
727 | ## open ane read file..... |
551b6b6f |
728 | if (not open IN, "UnicodeData.txt") { |
729 | die "$0: UnicodeData.txt: $!\n"; |
cf25bb62 |
730 | } |
d73e5302 |
731 | |
a3a8c5f0 |
732 | ## |
733 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} |
734 | ## |
735 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") |
736 | |
737 | my %CodeToDeco; ## Maps code to decomp. list for chars with first |
738 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) |
739 | |
740 | ## This is filled in as we go.... |
99598c8c |
741 | my $CombAbove = Table->New(Is => '_CombAbove', |
742 | Desc => '(for internal casefolding use)', |
743 | Fuzzy => 0); |
a3a8c5f0 |
744 | |
cf25bb62 |
745 | while (<IN>) |
746 | { |
747 | next unless /^[0-9A-Fa-f]+;/; |
748 | s/\s+$//; |
749 | |
750 | my ($hexcode, ## code point in hex (e.g. "0041") |
751 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") |
752 | $cat, ## category (e.g. "Lu") |
753 | $comb, ## Canonical combining class (e.t. "230") |
754 | $bidi, ## directional category (e.g. "L") |
755 | $deco, ## decomposition mapping |
756 | $decimal, ## decimal digit value |
757 | $digit, ## digit value |
758 | $number, ## numeric value |
759 | $mirrored, ## mirrored |
760 | $unicode10, ## name in Unicode 1.0 |
761 | $comment, ## comment field |
762 | $upper, ## uppercase mapping |
763 | $lower, ## lowercase mapping |
764 | $title, ## titlecase mapping |
765 | ) = split(/\s*;\s*/); |
766 | |
2eb5892f |
767 | # Note that in Unicode 3.2 there will be names like |
768 | # LINE FEED (LF), which probably means that \N{} needs |
769 | # to cope also with LINE FEED and LF. |
770 | $name = $unicode10 if $name eq '<control>' && $unicode10 ne ''; |
771 | |
cf25bb62 |
772 | my $code = hex($hexcode); |
773 | |
a3a8c5f0 |
774 | if ($comb and $comb == 230) { |
775 | $CombAbove->Append($code); |
776 | $_Above_HexCodes{$hexcode} = 1; |
777 | } |
778 | |
779 | ## Used in building \p{_CanonDCIJ} |
780 | if ($deco and $deco =~ m/^006[9A]\b/) { |
781 | $CodeToDeco{$code} = $deco; |
782 | } |
783 | |
cf25bb62 |
784 | ## |
785 | ## There are a few pairs of lines like: |
786 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; |
787 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; |
788 | ## that define ranges. |
789 | ## |
790 | if ($name =~ /^<(.+), (First|Last)>$/) |
791 | { |
792 | $name = $1; |
793 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); |
99598c8c |
794 | #New_Prop(In => $name, $General{$name}, Fuzzy => 1); |
cf25bb62 |
795 | } |
796 | else |
797 | { |
798 | ## normal (single-character) lines |
799 | gencat($name, $cat, $code, 'Append'); |
800 | |
801 | # No Append() here since since several codes may map into one. |
802 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; |
803 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; |
804 | $To{Title}->RawAppendRange($code, $code, $title) if $title; |
805 | $To{Digit}->Append($code, $decimal) if length $decimal; |
806 | |
807 | $Bidi->Append($code, $bidi); |
808 | $Comb->Append($code, $comb) if $comb; |
809 | $Number->Append($code, $number) if length $number; |
810 | |
12ac2576 |
811 | length($decimal) and ($Number{De} ||= Table->New())->Append($code) |
812 | or |
813 | length($digit) and ($Number{Di} ||= Table->New())->Append($code) |
814 | or |
815 | length($number) and ($Number{Nu} ||= Table->New())->Append($code); |
816 | |
cf25bb62 |
817 | $Mirrored->Append($code) if $mirrored eq "Y"; |
818 | |
12ac2576 |
819 | $Bidi{$bidi} ||= Table->New();#Is => "bt/$bidi", |
820 | #Desc => "Bi-directional category '$bidi'", |
821 | #Fuzzy => 0); |
cf25bb62 |
822 | $Bidi{$bidi}->Append($code); |
823 | |
824 | if ($deco) |
825 | { |
826 | $Deco->Append($code, $deco); |
827 | if ($deco =~/^<(\w+)>/) |
828 | { |
12ac2576 |
829 | my $dshort = $PVA_reverse{dt}{ucfirst lc $1}; |
830 | $DC{com}->Append($code); |
cf25bb62 |
831 | |
12ac2576 |
832 | $DC{$dshort} ||= Table->New(); |
833 | $DC{$dshort}->Append($code); |
cf25bb62 |
834 | } |
835 | else |
836 | { |
12ac2576 |
837 | $DC{can}->Append($code); |
cf25bb62 |
838 | } |
839 | } |
840 | } |
841 | } |
842 | close IN; |
d2d499f5 |
843 | |
cf25bb62 |
844 | ## |
845 | ## Tidy up a few special cases.... |
846 | ## |
d73e5302 |
847 | |
cf25bb62 |
848 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist |
99598c8c |
849 | New_Prop(Is => 'Cn', |
850 | $Cat{Cn}, |
851 | Desc => "General Category 'Cn' [not functional in Perl]", |
852 | Fuzzy => 0); |
d73e5302 |
853 | |
cf25bb62 |
854 | ## Unassigned is the same as 'Cn' |
5beb625e |
855 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0); |
d73e5302 |
856 | |
cf25bb62 |
857 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C |
d73e5302 |
858 | |
d73e5302 |
859 | |
12ac2576 |
860 | # LC is Ll, Lu, and Lt. |
861 | # (used to be L& or L_, but PropValueAliases.txt defines it as LC) |
862 | New_Prop(Is => 'LC', |
cf25bb62 |
863 | Table->Merge(@Cat{qw[Ll Lu Lt]}), |
99598c8c |
864 | Desc => '[\p{Ll}\p{Lu}\p{Lt}]', |
865 | Fuzzy => 0); |
d73e5302 |
866 | |
cf25bb62 |
867 | ## Any and All are all code points. |
99598c8c |
868 | my $Any = Table->New(Is => 'Any', |
869 | Desc => sprintf("[\\x{0000}-\\x{%X}]", |
870 | $LastUnicodeCodepoint), |
5beb625e |
871 | Fuzzy => 0); |
cf25bb62 |
872 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); |
d73e5302 |
873 | |
5beb625e |
874 | New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0); |
d73e5302 |
875 | |
a3a8c5f0 |
876 | ## |
877 | ## Build special properties for Perl's internal case-folding needs: |
878 | ## \p{_CaseIgnorable} |
879 | ## \p{_CanonDCIJ} |
880 | ## \p{_CombAbove} |
881 | ## _CombAbove was built above. Others are built here.... |
882 | ## |
883 | |
884 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] |
885 | New_Prop(Is => '_CaseIgnorable', |
886 | Table->Merge($Cat{Mn}, |
887 | 0x00AD, #SOFT HYPHEN |
888 | 0x2010), #HYPHEN |
99598c8c |
889 | Desc => '(for internal casefolding use)', |
890 | Fuzzy => 0); |
a3a8c5f0 |
891 | |
892 | |
893 | ## \p{_CanonDCIJ} is fairly complex... |
99598c8c |
894 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', |
895 | Desc => '(for internal casefolding use)', |
896 | Fuzzy => 0); |
a3a8c5f0 |
897 | ## It contains the ASCII 'i' and 'j'.... |
898 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") |
899 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") |
900 | ## ...and any character with a decomposition that starts with either of |
901 | ## those code points, but only if the decomposition does not have any |
902 | ## combining character with the "ABOVE" canonical combining class. |
903 | for my $code (sort { $a <=> $b} keys %CodeToDeco) |
904 | { |
905 | ## Need to ensure that all decomposition characters do not have |
906 | ## a %HexCodeToComb in %AboveCombClasses. |
907 | my $want = 1; |
908 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) |
909 | { |
910 | if (exists $_Above_HexCodes{$deco_hexcode}) { |
911 | ## one of the decmposition chars has an ABOVE combination |
912 | ## class, so we're not interested in this one |
913 | $want = 0; |
914 | last; |
915 | } |
916 | } |
917 | if ($want) { |
918 | $CanonCDIJ->Append($code); |
919 | } |
920 | } |
921 | |
922 | |
d73e5302 |
923 | |
cf25bb62 |
924 | ## |
925 | ## Now dump the files. |
926 | ## |
927 | $Name->Write("Name.pl"); |
12ac2576 |
928 | |
929 | # $Bidi->Write("Bidirectional.pl"); |
930 | mkdir("lib/bc", 0755); |
931 | for (keys %Bidi) { |
932 | $Bidi{$_}->Write( |
933 | "lib/bc/$_.pl", |
934 | "BidiClass category '$PropValueAlias{bc}{$_}'" |
935 | ); |
936 | } |
937 | |
cf25bb62 |
938 | $Comb->Write("CombiningClass.pl"); |
12ac2576 |
939 | mkdir("lib/ccc", 0755); |
940 | for (keys %{ $PropValueAlias{ccc} }) { |
941 | my ($code, $name) = @{ $PropValueAlias{ccc}{$_} }; |
942 | (my $c = Table->New())->Append($code); |
943 | $c->Write( |
944 | "lib/ccc/$_.pl", |
945 | "CombiningClass category '$name'" |
946 | ); |
947 | } |
948 | |
cf25bb62 |
949 | $Deco->Write("Decomposition.pl"); |
12ac2576 |
950 | mkdir("lib/dt", 0755); |
951 | for (keys %DC) { |
952 | $DC{$_}->Write( |
953 | "lib/dt/$_.pl", |
954 | "DecompositionType category '$PropValueAlias{dt}{$_}'" |
955 | ); |
956 | } |
957 | |
958 | # $Number->Write("Number.pl"); |
959 | mkdir("lib/nt", 0755); |
960 | for (keys %Number) { |
961 | $Number{$_}->Write( |
962 | "lib/nt/$_.pl", |
963 | "NumericType category '$PropValueAlias{nt}{$_}'" |
964 | ); |
965 | } |
966 | |
967 | # $General->Write("Category.pl"); |
cf25bb62 |
968 | |
969 | for my $to (sort keys %To) { |
970 | $To{$to}->Write("To/$to.pl"); |
d73e5302 |
971 | } |
12ac2576 |
972 | |
973 | for (keys %{ $PropValueAlias{gc} }) { |
974 | New_Alias(Is => $PropValueAlias{gc}{$_}, SameAs => $_, Fuzzy => 1); |
975 | } |
d73e5302 |
976 | } |
977 | |
cf25bb62 |
978 | ## |
551b6b6f |
979 | ## Process LineBreak.txt |
cf25bb62 |
980 | ## |
551b6b6f |
981 | sub LineBreak_Txt() |
cf25bb62 |
982 | { |
551b6b6f |
983 | if (not open IN, "LineBreak.txt") { |
984 | die "$0: LineBreak.txt: $!\n"; |
cf25bb62 |
985 | } |
d73e5302 |
986 | |
cf25bb62 |
987 | my $Lbrk = Table->New(); |
988 | my %Lbrk; |
d73e5302 |
989 | |
cf25bb62 |
990 | while (<IN>) |
991 | { |
992 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; |
d73e5302 |
993 | |
cf25bb62 |
994 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); |
d73e5302 |
995 | |
cf25bb62 |
996 | $Lbrk->Append($first, $lbrk); |
d73e5302 |
997 | |
12ac2576 |
998 | $Lbrk{$lbrk} ||= Table->New(); |
cf25bb62 |
999 | $Lbrk{$lbrk}->Append($first); |
d73e5302 |
1000 | |
cf25bb62 |
1001 | if ($last) { |
1002 | $Lbrk->Extend($last); |
1003 | $Lbrk{$lbrk}->Extend($last); |
d73e5302 |
1004 | } |
1005 | } |
cf25bb62 |
1006 | close IN; |
d73e5302 |
1007 | |
12ac2576 |
1008 | # $Lbrk->Write("Lbrk.pl"); |
1009 | |
1010 | mkdir("lib/lb", 0755); |
1011 | |
1012 | for (keys %Lbrk) { |
1013 | $Lbrk{$_}->Write( |
1014 | "lib/lb/$_.pl", |
1015 | "Linebreak category '$PropValueAlias{lb}{$_}'" |
1016 | ); |
1017 | } |
cf25bb62 |
1018 | } |
d73e5302 |
1019 | |
cf25bb62 |
1020 | ## |
551b6b6f |
1021 | ## Process ArabicShaping.txt. |
cf25bb62 |
1022 | ## |
551b6b6f |
1023 | sub ArabicShaping_txt() |
cf25bb62 |
1024 | { |
551b6b6f |
1025 | if (not open IN, "ArabicShaping.txt") { |
1026 | die "$0: ArabicShaping.txt: $!\n"; |
cf25bb62 |
1027 | } |
d73e5302 |
1028 | |
cf25bb62 |
1029 | my $ArabLink = Table->New(); |
1030 | my $ArabLinkGroup = Table->New(); |
d73e5302 |
1031 | |
12ac2576 |
1032 | my %JoinType; |
1033 | |
cf25bb62 |
1034 | while (<IN>) |
1035 | { |
1036 | next unless /^[0-9A-Fa-f]+;/; |
1037 | s/\s+$//; |
d73e5302 |
1038 | |
cf25bb62 |
1039 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); |
1040 | my $code = hex($hexcode); |
1041 | $ArabLink->Append($code, $link); |
1042 | $ArabLinkGroup->Append($code, $linkgroup); |
12ac2576 |
1043 | |
1044 | $JoinType{$link} ||= Table->New(Is => "JoinType$link"); |
1045 | $JoinType{$link}->Append($code); |
d73e5302 |
1046 | } |
cf25bb62 |
1047 | close IN; |
1048 | |
12ac2576 |
1049 | # $ArabLink->Write("ArabLink.pl"); |
1050 | # $ArabLinkGroup->Write("ArabLnkGrp.pl"); |
1051 | |
1052 | mkdir("lib/jt", 0755); |
1053 | |
1054 | for (keys %JoinType) { |
1055 | $JoinType{$_}->Write( |
1056 | "lib/jt/$_.pl", |
1057 | "JoiningType category '$PropValueAlias{jt}{$_}'" |
1058 | ); |
1059 | } |
1060 | } |
1061 | |
1062 | ## |
1063 | ## Process EastAsianWidth.txt. |
1064 | ## |
1065 | sub EastAsianWidth_txt() |
1066 | { |
1067 | if (not open IN, "EastAsianWidth.txt") { |
1068 | die "$0: EastAsianWidth.txt: $!\n"; |
1069 | } |
1070 | |
1071 | my %EAW; |
1072 | |
1073 | while (<IN>) |
1074 | { |
1075 | next unless /^[0-9A-Fa-f]+;/; |
1076 | s/#.*//; |
1077 | s/\s+$//; |
1078 | |
1079 | my ($hexcode, $pv) = split(/\s*;\s*/); |
1080 | my $code = hex($hexcode); |
1081 | $EAW{$pv} ||= Table->New(Is => "EastAsianWidth$pv"); |
1082 | $EAW{$pv}->Append($code); |
1083 | } |
1084 | close IN; |
1085 | |
1086 | mkdir("lib/ea", 0755); |
1087 | |
1088 | for (keys %EAW) { |
1089 | $EAW{$_}->Write( |
1090 | "lib/ea/$_.pl", |
1091 | "EastAsianWidth category '$PropValueAlias{ea}{$_}'" |
1092 | ); |
1093 | } |
1094 | } |
1095 | |
1096 | ## |
1097 | ## Process HangulSyllableType.txt. |
1098 | ## |
1099 | sub HangulSyllableType_txt() |
1100 | { |
1101 | if (not open IN, "HangulSyllableType.txt") { |
1102 | die "$0: HangulSyllableType.txt: $!\n"; |
1103 | } |
1104 | |
1105 | my %HST; |
1106 | |
1107 | while (<IN>) |
1108 | { |
1109 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; |
1110 | my ($first, $last, $pv) = (hex($1), hex($2||""), $3); |
1111 | |
1112 | $HST{$pv} ||= Table->New(Is => "HangulSyllableType$pv"); |
1113 | $HST{$pv}->Append($first); |
1114 | |
1115 | if ($last) { $HST{$pv}->Extend($last) } |
1116 | } |
1117 | close IN; |
1118 | |
1119 | mkdir("lib/hst", 0755); |
1120 | |
1121 | for (keys %HST) { |
1122 | $HST{$_}->Write( |
1123 | "lib/hst/$_.pl", |
1124 | "HangulSyllableType category '$PropValueAlias{hst}{$_}'" |
1125 | ); |
1126 | } |
d73e5302 |
1127 | } |
1128 | |
cf25bb62 |
1129 | ## |
1130 | ## Process Jamo.txt. |
1131 | ## |
1132 | sub Jamo_txt() |
1133 | { |
1134 | if (not open IN, "Jamo.txt") { |
1135 | die "$0: Jamo.txt: $!\n"; |
1136 | } |
1137 | my $Short = Table->New(); |
d73e5302 |
1138 | |
cf25bb62 |
1139 | while (<IN>) |
1140 | { |
1141 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; |
1142 | my ($code, $short) = (hex($1), $2); |
d73e5302 |
1143 | |
cf25bb62 |
1144 | $Short->Append($code, $short); |
d73e5302 |
1145 | } |
cf25bb62 |
1146 | close IN; |
12ac2576 |
1147 | # $Short->Write("JamoShort.pl"); |
d73e5302 |
1148 | } |
1149 | |
cf25bb62 |
1150 | ## |
1151 | ## Process Scripts.txt. |
1152 | ## |
1153 | sub Scripts_txt() |
1154 | { |
1155 | my @ScriptInfo; |
d73e5302 |
1156 | |
cf25bb62 |
1157 | if (not open(IN, "Scripts.txt")) { |
1158 | die "$0: Scripts.txt: $!\n"; |
1159 | } |
1160 | while (<IN>) { |
1161 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
1162 | |
cf25bb62 |
1163 | # Wait until all the scripts have been read since |
1164 | # they are not listed in numeric order. |
1165 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; |
1166 | } |
1167 | close IN; |
d73e5302 |
1168 | |
cf25bb62 |
1169 | # Now append the scripts properties in their code point order. |
d73e5302 |
1170 | |
cf25bb62 |
1171 | my %Script; |
1172 | my $Scripts = Table->New(); |
d73e5302 |
1173 | |
cf25bb62 |
1174 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) |
1175 | { |
1176 | my ($first, $last, $name) = @$script; |
1177 | $Scripts->Append($first, $name); |
d73e5302 |
1178 | |
99598c8c |
1179 | $Script{$name} ||= Table->New(Is => $name, |
1180 | Desc => "Script '$name'", |
1181 | Fuzzy => 1); |
cf25bb62 |
1182 | $Script{$name}->Append($first, $name); |
d73e5302 |
1183 | |
cf25bb62 |
1184 | if ($last) { |
1185 | $Scripts->Extend($last); |
1186 | $Script{$name}->Extend($last); |
1187 | } |
1188 | } |
d73e5302 |
1189 | |
12ac2576 |
1190 | # $Scripts->Write("Scripts.pl"); |
d73e5302 |
1191 | |
cf25bb62 |
1192 | ## Common is everything not explicitly assigned to a Script |
1193 | ## |
1194 | ## ***shouldn't this be intersected with \p{Assigned}? ****** |
1195 | ## |
99598c8c |
1196 | New_Prop(Is => 'Common', |
1197 | $Scripts->Invert, |
1198 | Desc => 'Pseudo-Script of codepoints not in other Unicode scripts', |
1199 | Fuzzy => 1); |
cf25bb62 |
1200 | } |
d73e5302 |
1201 | |
cf25bb62 |
1202 | ## |
1203 | ## Given a name like "Close Punctuation", return a regex (that when applied |
1204 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", |
1205 | ## "Close-Punctuation", etc.) |
1206 | ## |
1207 | ## Accept any space, dash, or underbar where in the official name there is |
1208 | ## space or a dash (or underbar, but there never is). |
1209 | ## |
1210 | ## |
1211 | sub NameToRegex($) |
1212 | { |
1213 | my $Name = shift; |
1214 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; |
1215 | return $Name; |
1216 | } |
d73e5302 |
1217 | |
cf25bb62 |
1218 | ## |
1219 | ## Process Blocks.txt. |
1220 | ## |
1221 | sub Blocks_txt() |
1222 | { |
1223 | my $Blocks = Table->New(); |
1224 | my %Blocks; |
d73e5302 |
1225 | |
cf25bb62 |
1226 | if (not open IN, "Blocks.txt") { |
1227 | die "$0: Blocks.txt: $!\n"; |
1228 | } |
d73e5302 |
1229 | |
cf25bb62 |
1230 | while (<IN>) |
1231 | { |
1232 | #next if not /Private Use$/; |
1233 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; |
d73e5302 |
1234 | |
cf25bb62 |
1235 | my ($first, $last, $name) = (hex($1), hex($2), $3); |
d73e5302 |
1236 | |
cf25bb62 |
1237 | $Blocks->Append($first, $name); |
76ccdbe2 |
1238 | |
99598c8c |
1239 | $Blocks{$name} ||= Table->New(In => $name, |
1240 | Desc => "Block '$name'", |
1241 | Fuzzy => 1); |
cf25bb62 |
1242 | $Blocks{$name}->Append($first, $name); |
76ccdbe2 |
1243 | |
cf25bb62 |
1244 | if ($last and $last != $first) { |
1245 | $Blocks->Extend($last); |
1246 | $Blocks{$name}->Extend($last); |
d73e5302 |
1247 | } |
d73e5302 |
1248 | } |
cf25bb62 |
1249 | close IN; |
1250 | |
12ac2576 |
1251 | # $Blocks->Write("Blocks.pl"); |
d73e5302 |
1252 | } |
1253 | |
cf25bb62 |
1254 | ## |
1255 | ## Read in the PropList.txt. It contains extended properties not |
551b6b6f |
1256 | ## listed in the UnicodeData.txt, such as 'Other_Alphabetic': |
cf25bb62 |
1257 | ## alphabetic but not of the general category L; many modifiers |
1258 | ## belong to this extended property category: while they are not |
1259 | ## alphabets, they are alphabetic in nature. |
1260 | ## |
1261 | sub PropList_txt() |
1262 | { |
1263 | my @PropInfo; |
1264 | |
1265 | if (not open IN, "PropList.txt") { |
1266 | die "$0: PropList.txt: $!\n"; |
1267 | } |
d73e5302 |
1268 | |
cf25bb62 |
1269 | while (<IN>) |
1270 | { |
1271 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
1272 | |
cf25bb62 |
1273 | # Wait until all the extended properties have been read since |
1274 | # they are not listed in numeric order. |
1275 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; |
1276 | } |
1277 | close IN; |
71d929cb |
1278 | |
cf25bb62 |
1279 | # Now append the extended properties in their code point order. |
1280 | my $Props = Table->New(); |
1281 | my %Prop; |
71d929cb |
1282 | |
cf25bb62 |
1283 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) |
1284 | { |
1285 | my ($first, $last, $name) = @$prop; |
1286 | $Props->Append($first, $name); |
71d929cb |
1287 | |
99598c8c |
1288 | $Prop{$name} ||= Table->New(Is => $name, |
1289 | Desc => "Extended property '$name'", |
1290 | Fuzzy => 1); |
cf25bb62 |
1291 | $Prop{$name}->Append($first, $name); |
71d929cb |
1292 | |
cf25bb62 |
1293 | if ($last) { |
1294 | $Props->Extend($last); |
1295 | $Prop{$name}->Extend($last); |
1296 | } |
71d929cb |
1297 | } |
d73e5302 |
1298 | |
12ac2576 |
1299 | for (keys %Prop) { |
1300 | (my $file = $PA_reverse{$_}) =~ tr/_//d; |
1301 | $Prop{$_}->Write( |
1302 | "lib/gc_sc/$file.pl", |
1303 | "Binary property '$_'" |
1304 | ); |
1305 | } |
1306 | |
cf25bb62 |
1307 | # Alphabetic is L and Other_Alphabetic. |
99598c8c |
1308 | New_Prop(Is => 'Alphabetic', |
cf25bb62 |
1309 | Table->Merge($Cat{L}, $Prop{Other_Alphabetic}), |
99598c8c |
1310 | Desc => '[\p{L}\p{OtherAlphabetic}]', # use canonical names here |
1311 | Fuzzy => 1); |
cf25bb62 |
1312 | |
1313 | # Lowercase is Ll and Other_Lowercase. |
99598c8c |
1314 | New_Prop(Is => 'Lowercase', |
cf25bb62 |
1315 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), |
99598c8c |
1316 | Desc => '[\p{Ll}\p{OtherLowercase}]', # use canonical names here |
1317 | Fuzzy => 1); |
cf25bb62 |
1318 | |
1319 | # Uppercase is Lu and Other_Uppercase. |
1320 | New_Prop(Is => 'Uppercase', |
1321 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), |
99598c8c |
1322 | Desc => '[\p{Lu}\p{Other_Uppercase}]', # use canonical names here |
1323 | Fuzzy => 1); |
cf25bb62 |
1324 | |
1325 | # Math is Sm and Other_Math. |
1326 | New_Prop(Is => 'Math', |
1327 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), |
99598c8c |
1328 | Desc => '[\p{Sm}\p{OtherMath}]', # use canonical names here |
1329 | Fuzzy => 1); |
cf25bb62 |
1330 | |
1331 | # ID_Start is Ll, Lu, Lt, Lm, Lo, and Nl. |
1332 | New_Prop(Is => 'ID_Start', |
1333 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}), |
99598c8c |
1334 | Desc => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}]', |
1335 | Fuzzy => 1); |
cf25bb62 |
1336 | |
1337 | # ID_Continue is ID_Start, Mn, Mc, Nd, and Pc. |
1338 | New_Prop(Is => 'ID_Continue', |
1339 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}), |
99598c8c |
1340 | Desc => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}]', |
1341 | Fuzzy => 1); |
d73e5302 |
1342 | } |
1343 | |
5beb625e |
1344 | |
1345 | ## |
1346 | ## These are used in: |
1347 | ## MakePropTestScript() |
1348 | ## WriteAllMappings() |
1349 | ## for making the test script. |
1350 | ## |
1351 | my %FuzzyNameToTest; |
1352 | my %ExactNameToTest; |
1353 | |
1354 | |
1355 | ## This used only for making the test script |
1356 | sub GenTests($$$$) |
1357 | { |
1358 | my $FH = shift; |
1359 | my $Prop = shift; |
1360 | my $MatchCode = shift; |
1361 | my $FailCode = shift; |
1362 | |
1363 | if (defined $MatchCode) { |
1364 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode; |
1365 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode; |
1366 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode; |
1367 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode; |
1368 | } |
1369 | if (defined $FailCode) { |
1370 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode; |
1371 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode; |
1372 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode; |
1373 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode; |
1374 | } |
1375 | } |
1376 | |
1377 | ## This used only for making the test script |
1378 | sub ExpectError($$) |
1379 | { |
1380 | my $FH = shift; |
1381 | my $prop = shift; |
1382 | |
1383 | print $FH qq/Error('\\p{$prop}');\n/; |
1384 | print $FH qq/Error('\\P{$prop}');\n/; |
1385 | } |
1386 | |
1387 | ## This used only for making the test script |
1388 | my @GoodSeps = ( |
1389 | " ", |
1390 | "-", |
1391 | " \t ", |
1392 | "", |
1393 | "", |
1394 | "_", |
1395 | ); |
1396 | my @BadSeps = ( |
1397 | "--", |
1398 | "__", |
1399 | " _", |
1400 | "/" |
1401 | ); |
1402 | |
1403 | ## This used only for making the test script |
1404 | sub RandomlyFuzzifyName($;$) |
1405 | { |
1406 | my $Name = shift; |
1407 | my $WantError = shift; ## if true, make an error |
1408 | |
1409 | my @parts; |
1410 | for my $part (split /[-\s_]+/, $Name) |
1411 | { |
1412 | if (@parts) { |
1413 | if ($WantError and rand() < 0.3) { |
1414 | push @parts, $BadSeps[rand(@BadSeps)]; |
1415 | $WantError = 0; |
1416 | } else { |
1417 | push @parts, $GoodSeps[rand(@GoodSeps)]; |
1418 | } |
1419 | } |
1420 | my $switch = int rand(4); |
1421 | if ($switch == 0) { |
1422 | push @parts, uc $part; |
1423 | } elsif ($switch == 1) { |
1424 | push @parts, lc $part; |
1425 | } elsif ($switch == 2) { |
1426 | push @parts, ucfirst $part; |
1427 | } else { |
1428 | push @parts, $part; |
1429 | } |
1430 | } |
1431 | my $new = join('', @parts); |
1432 | |
1433 | if ($WantError) { |
1434 | if (rand() >= 0.5) { |
1435 | $new .= $BadSeps[rand(@BadSeps)]; |
1436 | } else { |
1437 | $new = $BadSeps[rand(@BadSeps)] . $new; |
1438 | } |
1439 | } |
1440 | return $new; |
1441 | } |
1442 | |
1443 | ## This used only for making the test script |
1444 | sub MakePropTestScript() |
1445 | { |
1446 | ## this written directly -- it's huge. |
1447 | if (not open OUT, ">TestProp.pl") { |
1448 | die "$0: TestProp.pl: $!\n"; |
1449 | } |
1450 | print OUT <DATA>; |
1451 | |
1452 | while (my ($Name, $Table) = each %ExactNameToTest) |
1453 | { |
1454 | GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode); |
1455 | ExpectError(*OUT, uc $Name) if uc $Name ne $Name; |
1456 | ExpectError(*OUT, lc $Name) if lc $Name ne $Name; |
1457 | } |
1458 | |
1459 | |
1460 | while (my ($Name, $Table) = each %FuzzyNameToTest) |
1461 | { |
1462 | my $Orig = $CanonicalToOrig{$Name}; |
1463 | my %Names = ( |
1464 | $Name => 1, |
1465 | $Orig => 1, |
1466 | RandomlyFuzzifyName($Orig) => 1 |
1467 | ); |
1468 | |
1469 | for my $N (keys %Names) { |
1470 | GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode); |
1471 | } |
1472 | |
1473 | ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR')); |
1474 | } |
1475 | |
1476 | print OUT "Finished();\n"; |
1477 | close OUT; |
1478 | } |
1479 | |
1480 | |
1481 | ## |
1482 | ## These are used only in: |
1483 | ## RegisterFileForName() |
1484 | ## WriteAllMappings() |
1485 | ## |
1486 | my %Exact; ## will become %utf8::Exact; |
1487 | my %Canonical; ## will become %utf8::Canonical; |
1488 | my %CaComment; ## Comment for %Canonical entry of same key |
1489 | |
1490 | ## |
1491 | ## Given info about a name and a datafile that it should be associated with, |
1492 | ## register that assocation in %Exact and %Canonical. |
1493 | sub RegisterFileForName($$$$) |
1494 | { |
1495 | my $Type = shift; |
1496 | my $Name = shift; |
1497 | my $IsFuzzy = shift; |
1498 | my $filename = shift; |
1499 | |
1500 | ## |
1501 | ## Now in details for the mapping. $Type eq 'Is' has the |
1502 | ## Is removed, as it will be removed in utf8_heavy when this |
1503 | ## data is being checked. In keeps its "In", but a second |
1504 | ## sans-In record is written if it doesn't conflict with |
1505 | ## anything already there. |
1506 | ## |
1507 | if (not $IsFuzzy) |
1508 | { |
1509 | if ($Type eq 'Is') { |
1510 | die "oops[$Name]" if $Exact{$Name}; |
1511 | $Exact{$Name} = $filename; |
1512 | } else { |
1513 | die "oops[$Type$Name]" if $Exact{"$Type$Name"}; |
1514 | $Exact{"$Type$Name"} = $filename; |
1515 | $Exact{$Name} = $filename if not $Exact{$Name}; |
1516 | } |
1517 | } |
1518 | else |
1519 | { |
1520 | my $CName = lc $Name; |
1521 | if ($Type eq 'Is') { |
1522 | die "oops[$CName]" if $Canonical{$CName}; |
1523 | $Canonical{$CName} = $filename; |
1524 | $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2; |
1525 | } else { |
1526 | die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"}; |
1527 | $Canonical{lc "$Type$CName"} = $filename; |
1528 | $CaComment{lc "$Type$CName"} = "$Type$Name"; |
1529 | if (not $Canonical{$CName}) { |
1530 | $Canonical{$CName} = $filename; |
1531 | $CaComment{$CName} = "$Type$Name"; |
1532 | } |
1533 | } |
1534 | } |
1535 | } |
1536 | |
cf25bb62 |
1537 | ## |
1538 | ## Writes the info accumulated in |
1539 | ## |
1540 | ## %TableInfo; |
1541 | ## %FuzzyNames; |
1542 | ## %AliasInfo; |
1543 | ## |
1544 | ## |
1545 | sub WriteAllMappings() |
1546 | { |
99598c8c |
1547 | my @MAP; |
1548 | |
5beb625e |
1549 | my %BaseNames; ## Base names already used (for avoiding 8.3 conflicts) |
cf25bb62 |
1550 | |
5beb625e |
1551 | ## 'Is' *MUST* come first, so its names have precidence over 'In's |
1552 | for my $Type ('Is', 'In') |
1553 | { |
1554 | my %RawNameToFile; ## a per-$Type cache |
cf25bb62 |
1555 | |
5beb625e |
1556 | for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}}) |
cf25bb62 |
1557 | { |
5beb625e |
1558 | ## Note: $Name is already canonical |
99598c8c |
1559 | my $Table = $TableInfo{$Type}->{$Name}; |
5beb625e |
1560 | my $IsFuzzy = $FuzzyNames{$Type}->{$Name}; |
99598c8c |
1561 | |
1562 | ## Need an 8.3 safe filename (which means "an 8 safe" $filename) |
5beb625e |
1563 | my $filename; |
cf25bb62 |
1564 | { |
5beb625e |
1565 | ## 'Is' items lose 'Is' from the basename. |
12ac2576 |
1566 | $filename = $Type eq 'Is' ? |
1567 | ($PVA_reverse{sc}{$Name} || $Name) : |
1568 | "$Type$Name"; |
5beb625e |
1569 | |
1570 | $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_" |
1571 | substr($filename, 8) = '' if length($filename) > 8; |
1572 | |
1573 | ## |
1574 | ## Make sure the basename doesn't conflict with something we |
1575 | ## might have already written. If we have, say, |
1576 | ## InGreekExtended1 |
1577 | ## InGreekExtended2 |
1578 | ## they become |
1579 | ## InGreekE |
1580 | ## InGreek2 |
1581 | ## |
1582 | while (my $num = $BaseNames{lc $filename}++) |
1583 | { |
1584 | $num++; ## so basenames with numbers start with '2', which |
1585 | ## just looks more natural. |
1586 | ## Want to append $num, but if it'll make the basename longer |
1587 | ## than 8 characters, pre-truncate $filename so that the result |
1588 | ## is acceptable. |
1589 | my $delta = length($filename) + length($num) - 8; |
1590 | if ($delta > 0) { |
1591 | substr($filename, -$delta) = $num; |
1592 | } else { |
1593 | $filename .= $num; |
1594 | } |
99598c8c |
1595 | } |
5beb625e |
1596 | }; |
99598c8c |
1597 | |
1598 | ## |
1599 | ## Construct a nice comment to add to the file, and build data |
1600 | ## for the "./Properties" file along the way. |
1601 | ## |
1602 | my $Comment; |
1603 | { |
1604 | my $Desc = $TableDesc{$Type}->{$Name} || ""; |
1605 | ## get list of names this table is reference by |
1606 | my @Supported = $Name; |
1607 | while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} }) |
1608 | { |
1609 | if ($Orig eq $Name) { |
1610 | push @Supported, $Alias; |
1611 | } |
1612 | } |
1613 | |
1614 | my $TypeToShow = $Type eq 'Is' ? "" : $Type; |
1615 | my $OrigProp; |
1616 | |
1617 | $Comment = "This file supports:\n"; |
1618 | for my $N (@Supported) |
1619 | { |
1620 | my $IsFuzzy = $FuzzyNames{$Type}->{$N}; |
5beb625e |
1621 | my $Prop = "\\p{$TypeToShow$Name}"; |
99598c8c |
1622 | $OrigProp = $Prop if not $OrigProp; #cache for aliases |
1623 | if ($IsFuzzy) { |
1624 | $Comment .= "\t$Prop (and fuzzy permutations)\n"; |
1625 | } else { |
1626 | $Comment .= "\t$Prop\n"; |
1627 | } |
1628 | my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)"; |
1629 | |
1630 | push @MAP, sprintf("%s %-42s %s\n", |
1631 | $IsFuzzy ? '*' : ' ', $Prop, $MyDesc); |
1632 | } |
1633 | if ($Desc) { |
1634 | $Comment .= "\nMeaning: $Desc\n"; |
1635 | } |
1636 | |
1637 | } |
cf25bb62 |
1638 | ## |
1639 | ## Okay, write the file... |
1640 | ## |
12ac2576 |
1641 | $Table->Write("lib/gc_sc/$filename.pl", $Comment); |
99598c8c |
1642 | |
5beb625e |
1643 | ## and register it |
1644 | $RawNameToFile{$Name} = $filename; |
1645 | RegisterFileForName($Type => $Name, $IsFuzzy, $filename); |
cf25bb62 |
1646 | |
5beb625e |
1647 | if ($IsFuzzy) |
1648 | { |
1649 | my $CName = CanonicalName($Type . '_'. $Name); |
1650 | $FuzzyNameToTest{$Name} = $Table if !$FuzzyNameToTest{$Name}; |
1651 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1652 | } else { |
1653 | $ExactNameToTest{$Name} = $Table; |
cf25bb62 |
1654 | } |
1655 | |
cf25bb62 |
1656 | } |
1657 | |
5beb625e |
1658 | ## Register aliase info |
1659 | for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}}) |
cf25bb62 |
1660 | { |
5beb625e |
1661 | my $Alias = $AliasInfo{$Type}->{$Name}; |
1662 | my $IsFuzzy = $FuzzyNames{$Type}->{$Alias}; |
1663 | my $filename = $RawNameToFile{$Name}; |
1664 | die "oops [$Alias]->[$Name]" if not $filename; |
1665 | RegisterFileForName($Type => $Alias, $IsFuzzy, $filename); |
1666 | |
1667 | my $Table = $TableInfo{$Type}->{$Name}; |
1668 | die "oops" if not $Table; |
1669 | if ($IsFuzzy) |
1670 | { |
1671 | my $CName = CanonicalName($Type .'_'. $Alias); |
1672 | $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias}; |
1673 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1674 | } else { |
1675 | $ExactNameToTest{$Alias} = $Table; |
1676 | } |
cf25bb62 |
1677 | } |
5beb625e |
1678 | } |
cf25bb62 |
1679 | |
5beb625e |
1680 | ## |
1681 | ## Write out the property list |
1682 | ## |
1683 | { |
1684 | my @OUT = ( |
1685 | "##\n", |
1686 | "## This file created by $0\n", |
1687 | "## List of built-in \\p{...}/\\P{...} properties.\n", |
1688 | "##\n", |
1689 | "## '*' means name may be 'fuzzy'\n", |
1690 | "##\n\n", |
1691 | sort { substr($a,2) cmp substr($b, 2) } @MAP, |
1692 | ); |
1693 | WriteIfChanged('Properties', @OUT); |
1694 | } |
cf25bb62 |
1695 | |
5beb625e |
1696 | use Text::Tabs (); ## using this makes the files about half the size |
1697 | |
1698 | ## Write Exact.pl |
1699 | { |
1700 | my @OUT = ( |
1701 | $HEADER, |
1702 | "##\n", |
1703 | "## Data in this file used by ../utf8_heavy.pl\n", |
1704 | "##\n\n", |
12ac2576 |
1705 | "## Mapping from name to filename in ./lib/gc_sc\n", |
5beb625e |
1706 | "%utf8::Exact = (\n", |
1707 | ); |
cf25bb62 |
1708 | |
12ac2576 |
1709 | $Exact{InGreek} = 'InGreekA'; # this is evil kludge |
cf25bb62 |
1710 | for my $Name (sort keys %Exact) |
1711 | { |
1712 | my $File = $Exact{$Name}; |
5beb625e |
1713 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1714 | my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/); |
1715 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
1716 | } |
5beb625e |
1717 | push @OUT, ");\n1;\n"; |
1718 | |
1719 | WriteIfChanged('Exact.pl', @OUT); |
1720 | } |
cf25bb62 |
1721 | |
5beb625e |
1722 | ## Write Canonical.pl |
1723 | { |
1724 | my @OUT = ( |
1725 | $HEADER, |
1726 | "##\n", |
1727 | "## Data in this file used by ../utf8_heavy.pl\n", |
1728 | "##\n\n", |
1729 | "## Mapping from lc(canonical name) to filename in ./lib\n", |
1730 | "%utf8::Canonical = (\n", |
1731 | ); |
1732 | my $Trail = ""; ## used just to keep the spacing pretty |
1733 | for my $Name (sort keys %Canonical) |
cf25bb62 |
1734 | { |
5beb625e |
1735 | my $File = $Canonical{$Name}; |
1736 | if ($CaComment{$Name}) { |
1737 | push @OUT, "\n" if not $Trail; |
1738 | push @OUT, " # $CaComment{$Name}\n"; |
1739 | $Trail = "\n"; |
1740 | } else { |
1741 | $Trail = ""; |
cf25bb62 |
1742 | } |
5beb625e |
1743 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1744 | my $Text = sprintf(" %-41s => %s,\n$Trail", $Name, qq/'$File'/); |
1745 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
1746 | } |
5beb625e |
1747 | push @OUT, ");\n1\n"; |
1748 | WriteIfChanged('Canonical.pl', @OUT); |
d2d499f5 |
1749 | } |
5beb625e |
1750 | |
1751 | MakePropTestScript() if $MakeTestScript; |
d2d499f5 |
1752 | } |
1753 | |
5beb625e |
1754 | |
551b6b6f |
1755 | sub SpecialCasing_txt() |
cf25bb62 |
1756 | { |
1757 | # |
1758 | # Read in the special cases. |
1759 | # |
983ffd37 |
1760 | |
cf25bb62 |
1761 | my %CaseInfo; |
1762 | |
551b6b6f |
1763 | if (not open IN, "SpecialCasing.txt") { |
1764 | die "$0: SpecialCasing.txt: $!\n"; |
cf25bb62 |
1765 | } |
1766 | while (<IN>) { |
1767 | next unless /^[0-9A-Fa-f]+;/; |
1768 | s/\#.*//; |
1769 | s/\s+$//; |
1770 | |
1771 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); |
1772 | |
1773 | if ($condition) { # not implemented yet |
1774 | print "# SKIPPING $_\n" if $Verbose; |
1775 | next; |
1776 | } |
1777 | |
1778 | # Wait until all the special cases have been read since |
1779 | # they are not listed in numeric order. |
1780 | my $ix = hex($code); |
a6da3000 |
1781 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ] |
1782 | unless $code eq $lower; |
1783 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ] |
1784 | unless $code eq $title; |
1785 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ] |
1786 | unless $code eq $upper; |
cf25bb62 |
1787 | } |
1788 | close IN; |
1789 | |
1790 | # Now write out the special cases properties in their code point order. |
1791 | # Prepend them to the To/{Upper,Lower,Title}.pl. |
1792 | |
1793 | for my $case (qw(Lower Title Upper)) |
1794 | { |
1795 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; |
cf25bb62 |
1796 | |
b08cf34e |
1797 | my @OUT = |
1798 | ( |
1799 | $HEADER, "\n", |
1800 | "# The key UTF-8 _bytes_, the value UTF-8 (speed hack)\n", |
1801 | "%utf8::ToSpec$case =\n(\n", |
1802 | ); |
cf25bb62 |
1803 | |
1804 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { |
1805 | my ($ix, $code, $to) = @$prop; |
1806 | my $tostr = |
1807 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; |
b08cf34e |
1808 | push @OUT, sprintf qq["%s" => "$tostr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $ix))); |
5cb851a6 |
1809 | # Remove any single-character mappings for |
1810 | # the same character since we are going for |
1811 | # the special casing rules. |
1812 | $NormalCase =~ s/^$code\t\t\w+\n//m; |
cf25bb62 |
1813 | } |
5beb625e |
1814 | push @OUT, ( |
1815 | ");\n\n", |
1816 | "return <<'END';\n", |
1817 | $NormalCase, |
1818 | "END\n" |
1819 | ); |
1820 | WriteIfChanged("To/$case.pl", @OUT); |
d2d499f5 |
1821 | } |
d2d499f5 |
1822 | } |
1823 | |
c4051cc5 |
1824 | # |
1825 | # Read in the case foldings. |
1826 | # |
551b6b6f |
1827 | # We will do full case folding, C + F + I (see CaseFolding.txt). |
c4051cc5 |
1828 | # |
551b6b6f |
1829 | sub CaseFolding_txt() |
cf25bb62 |
1830 | { |
551b6b6f |
1831 | if (not open IN, "CaseFolding.txt") { |
1832 | die "$0: CaseFolding.txt: $!\n"; |
cf25bb62 |
1833 | } |
c4051cc5 |
1834 | |
cf25bb62 |
1835 | my $Fold = Table->New(); |
c4051cc5 |
1836 | my %Fold; |
1837 | |
cf25bb62 |
1838 | while (<IN>) { |
254ba52a |
1839 | # Skip status 'S', simple case folding |
c4051cc5 |
1840 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; |
1841 | |
cf25bb62 |
1842 | my ($code, $status, $fold) = (hex($1), $2, $3); |
c4051cc5 |
1843 | |
1844 | if ($status eq 'C') { # Common: one-to-one folding |
254ba52a |
1845 | # No append() since several codes may fold into one. |
cf25bb62 |
1846 | $Fold->RawAppendRange($code, $code, $fold); |
c4051cc5 |
1847 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I |
cf25bb62 |
1848 | $Fold{$code} = $fold; |
c4051cc5 |
1849 | } |
1850 | } |
cf25bb62 |
1851 | close IN; |
c4051cc5 |
1852 | |
cf25bb62 |
1853 | $Fold->Write("To/Fold.pl"); |
c4051cc5 |
1854 | |
1855 | # |
1856 | # Prepend the special foldings to the common foldings. |
1857 | # |
c4051cc5 |
1858 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; |
5beb625e |
1859 | |
b08cf34e |
1860 | my @OUT = |
1861 | ( |
1862 | $HEADER, "\n", |
1863 | "# The ke UTF-8 _bytes_, the value UTF-8 (speed hack)\n", |
1864 | "%utf8::ToSpecFold =\n(\n", |
1865 | ); |
cf25bb62 |
1866 | for my $code (sort { $a <=> $b } keys %Fold) { |
1867 | my $foldstr = |
1868 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; |
b08cf34e |
1869 | push @OUT, sprintf qq["%s" => "$foldstr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $code))); |
c4051cc5 |
1870 | } |
5beb625e |
1871 | push @OUT, ( |
1872 | ");\n\n", |
1873 | "return <<'END';\n", |
1874 | $CommonFold, |
1875 | "END\n", |
1876 | ); |
1877 | |
1878 | WriteIfChanged("To/Fold.pl", @OUT); |
c4051cc5 |
1879 | } |
1880 | |
cf25bb62 |
1881 | ## Do it.... |
1882 | |
12ac2576 |
1883 | Build_Aliases(); |
44da8cae |
1884 | UnicodeData_Txt(); |
cf25bb62 |
1885 | PropList_txt(); |
1886 | |
1887 | Scripts_txt(); |
1888 | Blocks_txt(); |
1889 | |
5beb625e |
1890 | WriteAllMappings(); |
1891 | |
551b6b6f |
1892 | LineBreak_Txt(); |
1893 | ArabicShaping_txt(); |
12ac2576 |
1894 | EastAsianWidth_txt(); |
1895 | HangulSyllableType_txt(); |
cf25bb62 |
1896 | Jamo_txt(); |
551b6b6f |
1897 | SpecialCasing_txt(); |
1898 | CaseFolding_txt(); |
cf25bb62 |
1899 | |
5beb625e |
1900 | exit(0); |
cf25bb62 |
1901 | |
5beb625e |
1902 | ## TRAILING CODE IS USED BY MakePropTestScript() |
1903 | __DATA__ |
1904 | use strict; |
1905 | use warnings; |
1906 | |
1907 | my $Tests = 0; |
1908 | my $Fails = 0; |
cf25bb62 |
1909 | |
5beb625e |
1910 | sub Expect($$$) |
1911 | { |
1912 | my $Expect = shift; |
1913 | my $String = shift; |
1914 | my $Regex = shift; |
1915 | my $Line = (caller)[2]; |
1916 | |
1917 | $Tests++; |
1918 | my $RegObj; |
1919 | my $result = eval { |
1920 | $RegObj = qr/$Regex/; |
1921 | $String =~ $RegObj ? 1 : 0 |
1922 | }; |
1923 | |
1924 | if (not defined $result) { |
1925 | print "couldn't compile /$Regex/ on $0 line $Line: $@\n"; |
1926 | $Fails++; |
1927 | } elsif ($result ^ $Expect) { |
1928 | print "bad result (expected $Expect) on $0 line $Line: $@\n"; |
1929 | $Fails++; |
1930 | } |
1931 | } |
d73e5302 |
1932 | |
5beb625e |
1933 | sub Error($) |
1934 | { |
1935 | my $Regex = shift; |
1936 | $Tests++; |
1937 | if (eval { 'x' =~ qr/$Regex/; 1 }) { |
1938 | $Fails++; |
1939 | my $Line = (caller)[2]; |
1940 | print "expected error for /$Regex/ on $0 line $Line: $@\n"; |
1941 | } |
1942 | } |
1943 | |
1944 | sub Finished() |
1945 | { |
1946 | if ($Fails == 0) { |
1947 | print "All $Tests tests passed.\n"; |
1948 | exit(0); |
1949 | } else { |
1950 | print "$Tests tests, $Fails failed!\n"; |
1951 | exit(-1); |
1952 | } |
1953 | } |