Commit | Line | Data |
d73e5302 |
1 | #!/usr/bin/perl -w |
d73e5302 |
2 | use strict; |
cf25bb62 |
3 | use Carp; |
5beb625e |
4 | |
cf25bb62 |
5 | ## |
6 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) |
7 | ## from the Unicode database files (lib/unicore/*.txt). |
8 | ## |
d2d499f5 |
9 | |
5beb625e |
10 | mkdir("lib", 0755); |
11 | mkdir("To", 0755); |
d73e5302 |
12 | |
cf25bb62 |
13 | ## |
14 | ## Process any args. |
15 | ## |
5beb625e |
16 | my $Verbose = 0; |
17 | my $MakeTestScript = 0; |
cf25bb62 |
18 | |
19 | while (@ARGV) |
20 | { |
21 | my $arg = shift @ARGV; |
22 | if ($arg eq '-v') { |
23 | $Verbose = 1; |
24 | } elsif ($arg eq '-q') { |
25 | $Verbose = 0; |
5beb625e |
26 | } elsif ($arg eq '-maketest') { |
27 | $MakeTestScript = 1; |
d73e5302 |
28 | } else { |
5beb625e |
29 | die "usage: $0 [-v|-q] [-maketest]"; |
d73e5302 |
30 | } |
31 | } |
32 | |
cf25bb62 |
33 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 3.1.1. |
d73e5302 |
34 | |
cf25bb62 |
35 | my $HEADER=<<"EOF"; |
d73e5302 |
36 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! |
551b6b6f |
37 | # This file is built by $0 from e.g. UnicodeData.txt. |
d73e5302 |
38 | # Any changes made here will be lost! |
cf25bb62 |
39 | |
40 | EOF |
41 | |
5beb625e |
42 | |
43 | ## |
44 | ## Given a filename and a reference to an array of lines, |
45 | ## write the lines to the file only if the contents have not changed. |
46 | ## |
47 | sub WriteIfChanged($\@) |
48 | { |
49 | my $file = shift; |
50 | my $lines = shift; |
51 | |
52 | my $TextToWrite = join '', @$lines; |
53 | if (open IN, $file) { |
54 | local($/) = undef; |
55 | my $PreviousText = <IN>; |
56 | close IN; |
57 | if ($PreviousText eq $TextToWrite) { |
58 | print "$file unchanged.\n" if $Verbose; |
59 | return; |
60 | } |
61 | } |
62 | if (not open OUT, ">$file") { |
63 | die "$0: can't open $file for output: $!\n"; |
64 | } |
65 | print "$file written.\n" if $Verbose; |
66 | |
67 | print OUT $TextToWrite; |
68 | close OUT; |
69 | } |
70 | |
cf25bb62 |
71 | ## |
72 | ## The main datastructure (a "Table") represents a set of code points that |
73 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, |
74 | ## etc.). They are kept as ranges of code points (starting and ending of |
75 | ## each range). |
76 | ## |
77 | ## For example, a range ASCII LETTERS would be represented as: |
78 | ## [ [ 0x41 => 0x5A, 'UPPER' ], |
79 | ## [ 0x61 => 0x7A, 'LOWER, ] ] |
80 | ## |
81 | sub RANGE_START() { 0 } ## index into range element |
82 | sub RANGE_END() { 1 } ## index into range element |
83 | sub RANGE_NAME() { 2 } ## index into range element |
84 | |
99598c8c |
85 | ## Conceptually, these should really be folded into the 'Table' objects |
cf25bb62 |
86 | my %TableInfo; |
99598c8c |
87 | my %TableDesc; |
cf25bb62 |
88 | my %FuzzyNames; |
89 | my %AliasInfo; |
5beb625e |
90 | my %CanonicalToOrig; |
cf25bb62 |
91 | |
92 | ## |
99598c8c |
93 | ## Turn something like |
94 | ## OLD-ITALIC |
5beb625e |
95 | ## into |
99598c8c |
96 | ## OldItalic |
97 | ## |
98 | sub CanonicalName($) |
99 | { |
5beb625e |
100 | my $orig = shift; |
101 | my $name = lc $orig; |
99598c8c |
102 | $name =~ s/(?<![a-z])(\w)/\u$1/g; |
5beb625e |
103 | $name =~ s/[-_\s]+//g; |
99598c8c |
104 | |
5beb625e |
105 | $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name}; |
99598c8c |
106 | return $name; |
107 | } |
108 | |
99598c8c |
109 | ## |
cf25bb62 |
110 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. |
111 | ## |
112 | ## Called like: |
99598c8c |
113 | ## New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1); |
cf25bb62 |
114 | ## |
115 | ## Normally, these parameters are set when the Table is created (when the |
116 | ## Table->New constructor is called), but there are times when it needs to |
117 | ## be done after-the-fact...) |
118 | ## |
119 | sub New_Prop($$$@) |
120 | { |
121 | my $Type = shift; ## "Is" or "In"; |
122 | my $Name = shift; |
123 | my $Table = shift; |
124 | |
125 | ## remaining args are optional key/val |
126 | my %Args = @_; |
127 | |
99598c8c |
128 | my $Fuzzy = delete $Args{Fuzzy}; |
129 | my $Desc = delete $Args{Desc}; # description |
130 | |
5beb625e |
131 | $Name = CanonicalName($Name) if $Fuzzy; |
cf25bb62 |
132 | |
133 | ## sanity check a few args |
134 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { |
135 | confess "$0: bad args to New_Prop" |
136 | } |
137 | |
138 | if (not $TableInfo{$Type}->{$Name}) |
139 | { |
140 | $TableInfo{$Type}->{$Name} = $Table; |
99598c8c |
141 | $TableDesc{$Type}->{$Name} = $Desc; |
142 | if ($Fuzzy) { |
cf25bb62 |
143 | $FuzzyNames{$Type}->{$Name} = $Name; |
144 | } |
145 | } |
d73e5302 |
146 | } |
147 | |
d73e5302 |
148 | |
cf25bb62 |
149 | ## |
150 | ## Creates a new Table object. |
151 | ## |
152 | ## Args are key/value pairs: |
99598c8c |
153 | ## In => Name -- Name of "In" property to be associated with |
154 | ## Is => Name -- Name of "Is" property to be associated with |
155 | ## Fuzzy => Boolean -- True if name can be accessed "fuzzily" |
156 | ## Desc => String -- Description of the property |
cf25bb62 |
157 | ## |
158 | ## No args are required. |
159 | ## |
160 | sub Table::New |
161 | { |
162 | my $class = shift; |
163 | my %Args = @_; |
164 | |
165 | my $Table = bless [], $class; |
166 | |
99598c8c |
167 | my $Fuzzy = delete $Args{Fuzzy}; |
168 | my $Desc = delete $Args{Desc}; |
cf25bb62 |
169 | |
170 | for my $Type ('Is', 'In') |
171 | { |
172 | if (my $Name = delete $Args{$Type}) { |
99598c8c |
173 | New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy); |
cf25bb62 |
174 | } |
175 | } |
176 | |
177 | ## shouldn't have any left over |
178 | if (%Args) { |
179 | confess "$0: bad args to Table->New" |
180 | } |
181 | |
182 | return $Table; |
d73e5302 |
183 | } |
184 | |
cf25bb62 |
185 | ## |
186 | ## Returns true if the Table has no code points |
187 | ## |
188 | sub Table::IsEmpty |
189 | { |
190 | my $Table = shift; #self |
191 | return not @$Table; |
d73e5302 |
192 | } |
193 | |
cf25bb62 |
194 | ## |
195 | ## Returns true if the Table has code points |
196 | ## |
197 | sub Table::NotEmpty |
198 | { |
199 | my $Table = shift; #self |
200 | return @$Table; |
d73e5302 |
201 | } |
202 | |
cf25bb62 |
203 | ## |
204 | ## Returns the maximum code point currently in the table. |
205 | ## |
206 | sub Table::Max |
207 | { |
208 | my $Table = shift; #self |
209 | confess "oops" if $Table->IsEmpty; ## must have code points to have a max |
210 | return $Table->[-1]->[RANGE_END]; |
211 | } |
d73e5302 |
212 | |
cf25bb62 |
213 | ## |
214 | ## Replaces the codepoints in the Table with those in the Table given |
215 | ## as an arg. (NOTE: this is not a "deep copy"). |
216 | ## |
217 | sub Table::Replace($$) |
218 | { |
219 | my $Table = shift; #self |
220 | my $New = shift; |
d73e5302 |
221 | |
cf25bb62 |
222 | @$Table = @$New; |
223 | } |
71d929cb |
224 | |
cf25bb62 |
225 | ## |
226 | ## Given a new code point, make the last range of the Table extend to |
227 | ## include the new (and all intervening) code points. |
228 | ## |
229 | sub Table::Extend |
230 | { |
231 | my $Table = shift; #self |
232 | my $codepoint = shift; |
d73e5302 |
233 | |
cf25bb62 |
234 | my $PrevMax = $Table->Max; |
e904f995 |
235 | |
cf25bb62 |
236 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; |
e904f995 |
237 | |
cf25bb62 |
238 | $Table->[-1]->[RANGE_END] = $codepoint; |
239 | } |
c3a8a2b8 |
240 | |
cf25bb62 |
241 | ## |
242 | ## Given a code point range start and end (and optional name), blindly |
243 | ## append them to the list of ranges for the Table. |
244 | ## |
245 | ## NOTE: Code points must be added in strictly ascending numeric order. |
246 | ## |
247 | sub Table::RawAppendRange |
248 | { |
249 | my $Table = shift; #self |
250 | my $start = shift; |
251 | my $end = shift; |
252 | my $name = shift; |
253 | $name = "" if not defined $name; ## warning: $name can be "0" |
254 | |
255 | push @$Table, [ $start, # RANGE_START |
256 | $end, # RANGE_END |
257 | $name ]; # RANGE_NAME |
258 | } |
c3a8a2b8 |
259 | |
cf25bb62 |
260 | ## |
261 | ## Given a code point (and optional name), add it to the Table. |
262 | ## |
263 | ## NOTE: Code points must be added in strictly ascending numeric order. |
264 | ## |
265 | sub Table::Append |
266 | { |
267 | my $Table = shift; #self |
268 | my $codepoint = shift; |
269 | my $name = shift; |
270 | $name = "" if not defined $name; ## warning: $name can be "0" |
271 | |
272 | ## |
273 | ## If we've already got a range working, and this code point is the next |
274 | ## one in line, and if the name is the same, just extend the current range. |
275 | ## |
276 | if ($Table->NotEmpty |
277 | and |
278 | $Table->Max == $codepoint - 1 |
279 | and |
280 | $Table->[-1]->[RANGE_NAME] eq $name) |
281 | { |
282 | $Table->Extend($codepoint); |
c3a8a2b8 |
283 | } |
cf25bb62 |
284 | else |
285 | { |
286 | $Table->RawAppendRange($codepoint, $codepoint, $name); |
d73e5302 |
287 | } |
cf25bb62 |
288 | } |
d73e5302 |
289 | |
cf25bb62 |
290 | ## |
291 | ## Given a code point range starting value and ending value (and name), |
292 | ## Add the range to teh Table. |
293 | ## |
294 | ## NOTE: Code points must be added in strictly ascending numeric order. |
295 | ## |
296 | sub Table::AppendRange |
297 | { |
298 | my $Table = shift; #self |
299 | my $start = shift; |
300 | my $end = shift; |
301 | my $name = shift; |
302 | $name = "" if not defined $name; ## warning: $name can be "0" |
303 | |
304 | $Table->Append($start, $name); |
305 | $Table->Extend($end) if $end > $start; |
306 | } |
d73e5302 |
307 | |
cf25bb62 |
308 | ## |
309 | ## Return a new Table that represents all code points not in the Table. |
310 | ## |
311 | sub Table::Invert |
312 | { |
313 | my $Table = shift; #self |
314 | |
315 | my $New = Table->New(); |
316 | my $max = -1; |
317 | for my $range (@$Table) |
318 | { |
319 | my $start = $range->[RANGE_START]; |
320 | my $end = $range->[RANGE_END]; |
321 | if ($start-1 >= $max+1) { |
322 | $New->AppendRange($max+1, $start-1, ""); |
323 | } |
324 | $max = $end; |
d73e5302 |
325 | } |
cf25bb62 |
326 | if ($max+1 < $LastUnicodeCodepoint) { |
327 | $New->AppendRange($max+1, $LastUnicodeCodepoint); |
d73e5302 |
328 | } |
cf25bb62 |
329 | return $New; |
330 | } |
d73e5302 |
331 | |
cf25bb62 |
332 | ## |
333 | ## Merges any number of other tables with $self, returning the new table. |
334 | ## (existing tables are not modified) |
335 | ## |
a3a8c5f0 |
336 | ## |
337 | ## Args may be Tables, or individual code points (as integers). |
338 | ## |
cf25bb62 |
339 | ## Can be called as either a constructor or a method. |
340 | ## |
341 | sub Table::Merge |
342 | { |
343 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class |
344 | my @Tables = @_; |
345 | |
346 | ## Accumulate all records from all tables |
347 | my @Records; |
a3a8c5f0 |
348 | for my $Arg (@Tables) |
349 | { |
350 | if (ref $Arg) { |
351 | ## arg is a table -- get its ranges |
352 | push @Records, @$Arg; |
353 | } else { |
354 | ## arg is a codepoint, make a range |
355 | push @Records, [ $Arg, $Arg ] |
356 | } |
d73e5302 |
357 | } |
358 | |
cf25bb62 |
359 | ## sort by range start, with longer ranges coming first. |
360 | my ($first, @Rest) = sort { |
361 | ($a->[RANGE_START] <=> $b->[RANGE_START]) |
362 | or |
363 | ($b->[RANGE_END] <=> $b->[RANGE_END]) |
364 | } @Records; |
365 | |
366 | my $New = Table->New(); |
367 | |
368 | ## Ensuring the first range is there makes the subsequent loop easier |
369 | $New->AppendRange($first->[RANGE_START], |
370 | $first->[RANGE_END]); |
371 | |
372 | ## Fold in records so long as they add new information. |
373 | for my $set (@Rest) |
374 | { |
375 | my $start = $set->[RANGE_START]; |
376 | my $end = $set->[RANGE_END]; |
377 | if ($start > $New->Max) { |
378 | $New->AppendRange($start, $end); |
379 | } elsif ($end > $New->Max) { |
380 | $New->Extend($end); |
381 | } |
d73e5302 |
382 | } |
d73e5302 |
383 | |
cf25bb62 |
384 | return $New; |
d73e5302 |
385 | } |
386 | |
cf25bb62 |
387 | ## |
388 | ## Given a filename, write a representation of the Table to a file. |
99598c8c |
389 | ## May have an optional comment as a 2nd arg. |
cf25bb62 |
390 | ## |
391 | sub Table::Write |
392 | { |
99598c8c |
393 | my $Table = shift; #self |
cf25bb62 |
394 | my $filename = shift; |
99598c8c |
395 | my $comment = shift; |
d73e5302 |
396 | |
5beb625e |
397 | my @OUT = $HEADER; |
99598c8c |
398 | if (defined $comment) { |
399 | $comment =~ s/\s+\Z//; |
400 | $comment =~ s/^/# /gm; |
5beb625e |
401 | push @OUT, "#\n$comment\n#\n"; |
99598c8c |
402 | } |
5beb625e |
403 | push @OUT, "return <<'END';\n"; |
d73e5302 |
404 | |
cf25bb62 |
405 | for my $set (@$Table) |
406 | { |
407 | my $start = $set->[RANGE_START]; |
408 | my $end = $set->[RANGE_END]; |
409 | my $name = $set->[RANGE_NAME]; |
d73e5302 |
410 | |
cf25bb62 |
411 | if ($start == $end) { |
5beb625e |
412 | push @OUT, sprintf "%04X\t\t%s\n", $start, $name; |
cf25bb62 |
413 | } else { |
5beb625e |
414 | push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name; |
cf25bb62 |
415 | } |
416 | } |
d73e5302 |
417 | |
5beb625e |
418 | push @OUT, "END\n"; |
419 | |
420 | WriteIfChanged($filename, @OUT); |
421 | } |
422 | |
423 | ## This used only for making the test script. |
424 | ## helper function |
425 | sub IsUsable($) |
426 | { |
427 | my $code = shift; |
428 | return 0 if $code <= 0x0000; ## don't use null |
429 | return 0 if $code >= $LastUnicodeCodepoint; ## keep in range |
430 | return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates |
431 | return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good |
432 | return 0 if (($code & 0xFFFF) == 0xFFFE); ## utf8.c says no good |
433 | return 0 if (($code & 0xFFFF) == 0xFFFF); ## utf8.c says no good |
434 | return 1; |
435 | } |
436 | |
437 | ## Return a code point that's part of the table. |
438 | ## Returns nothing if the table is empty (or covers only surrogates). |
439 | ## This used only for making the test script. |
440 | sub Table::ValidCode |
441 | { |
442 | my $Table = shift; #self |
443 | for my $set (@$Table) { |
444 | return $set->[RANGE_END] if IsUsable($set->[RANGE_END]); |
445 | } |
446 | return (); |
447 | } |
448 | |
449 | ## Return a code point that's not part of the table |
450 | ## Returns nothing if the table covers all code points. |
451 | ## This used only for making the test script. |
452 | sub Table::InvalidCode |
453 | { |
454 | my $Table = shift; #self |
455 | |
456 | return 0x1234 if $Table->IsEmpty(); |
457 | |
458 | for my $set (@$Table) |
459 | { |
460 | if (IsUsable($set->[RANGE_END] + 1)) |
461 | { |
462 | return $set->[RANGE_END] + 1; |
463 | } |
464 | |
465 | if (IsUsable($set->[RANGE_START] - 1)) |
466 | { |
467 | return $set->[RANGE_START] - 1; |
468 | } |
469 | } |
470 | return (); |
cf25bb62 |
471 | } |
472 | |
473 | ########################################################################### |
474 | ########################################################################### |
475 | ########################################################################### |
476 | |
477 | |
478 | ## |
479 | ## Called like: |
99598c8c |
480 | ## New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1); |
cf25bb62 |
481 | ## |
99598c8c |
482 | ## The args must be in that order, although the Fuzzy pair may be omitted. |
cf25bb62 |
483 | ## |
484 | ## This creates 'IsAll' as an alias for 'IsAny' |
485 | ## |
486 | sub New_Alias($$$@) |
487 | { |
488 | my $Type = shift; ## "Is" or "In" |
489 | my $Alias = shift; |
99598c8c |
490 | my $SameAs = shift; # expecting "SameAs" -- just ignored |
cf25bb62 |
491 | my $Name = shift; |
492 | |
493 | ## remaining args are optional key/val |
494 | my %Args = @_; |
495 | |
99598c8c |
496 | my $Fuzzy = delete $Args{Fuzzy}; |
cf25bb62 |
497 | |
498 | ## sanity check a few args |
499 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { |
500 | confess "$0: bad args to New_Alias" |
d73e5302 |
501 | } |
502 | |
5beb625e |
503 | $Alias = CanonicalName($Alias) if $Fuzzy; |
504 | |
505 | if (not $TableInfo{$Type}->{$Name}) |
506 | { |
507 | my $CName = CanonicalName($Name); |
508 | if ($TableInfo{$Type}->{$CName}) { |
509 | confess "$0: Use canonical form '$CName' instead of '$Name' for alias."; |
510 | } else { |
511 | confess "$0: don't have orignial $Type => $Name to make alias"; |
512 | } |
cf25bb62 |
513 | } |
514 | if ($TableInfo{$Alias}) { |
515 | confess "$0: already have original $Type => $Alias; can't make alias"; |
d73e5302 |
516 | } |
cf25bb62 |
517 | $AliasInfo{$Type}->{$Name} = $Alias; |
99598c8c |
518 | if ($Fuzzy) { |
cf25bb62 |
519 | $FuzzyNames{$Type}->{$Alias} = $Name; |
520 | } |
521 | |
d73e5302 |
522 | } |
523 | |
d73e5302 |
524 | |
cf25bb62 |
525 | ## All assigned code points |
99598c8c |
526 | my $Assigned = Table->New(Is => 'Assigned', |
527 | Desc => "All assigned code points", |
5beb625e |
528 | Fuzzy => 0); |
d2d499f5 |
529 | |
cf25bb62 |
530 | my $Name = Table->New(); ## all characters, individually by name |
531 | my $General = Table->New(); ## all characters, grouped by category |
532 | my %General; |
533 | my %Cat; |
d73e5302 |
534 | |
cf25bb62 |
535 | ## |
551b6b6f |
536 | ## Process UnicodeData.txt (Categories, etc.) |
cf25bb62 |
537 | ## |
538 | sub Unicode_Txt() |
539 | { |
540 | my $Bidi = Table->New(); |
541 | my $Deco = Table->New(); |
542 | my $Comb = Table->New(); |
543 | my $Number = Table->New(); |
99598c8c |
544 | my $Mirrored = Table->New(Is => 'Mirrored', |
545 | Desc => "Mirrored in bidirectional text", |
546 | Fuzzy => 0); |
d73e5302 |
547 | |
cf25bb62 |
548 | my %DC; |
549 | my %Bidi; |
550 | my %Deco; |
99598c8c |
551 | $Deco{Canon} = Table->New(Is => 'Canon', |
552 | Desc => 'Decomposes to multiple characters', |
553 | Fuzzy => 0); |
554 | $Deco{Compat} = Table->New(Is => 'Compat', |
555 | Desc => 'Compatible with a more-basic character', |
556 | Fuzzy => 0); |
cf25bb62 |
557 | |
558 | ## Initialize Perl-generated categories |
551b6b6f |
559 | ## (Categories from UnicodeData.txt are auto-initialized in gencat) |
99598c8c |
560 | $Cat{Alnum} = Table->New(Is => 'Alnum', Desc => "[[:Alnum:]]", Fuzzy => 0); |
561 | $Cat{Alpha} = Table->New(Is => 'Alpha', Desc => "[[:Alpha:]]", Fuzzy => 0); |
562 | $Cat{ASCII} = Table->New(Is => 'ASCII', Desc => "[[:ASCII:]]", Fuzzy => 0); |
563 | $Cat{Blank} = Table->New(Is => 'Blank', Desc => "[[:Blank:]]", Fuzzy => 0); |
564 | $Cat{Cntrl} = Table->New(Is => 'Cntrl', Desc => "[[:Cntrl:]]", Fuzzy => 0); |
565 | $Cat{Digit} = Table->New(Is => 'Digit', Desc => "[[:Digit:]]", Fuzzy => 0); |
566 | $Cat{Graph} = Table->New(Is => 'Graph', Desc => "[[:Graph:]]", Fuzzy => 0); |
567 | $Cat{Lower} = Table->New(Is => 'Lower', Desc => "[[:Lower:]]", Fuzzy => 0); |
568 | $Cat{Print} = Table->New(Is => 'Print', Desc => "[[:Print:]]", Fuzzy => 0); |
569 | $Cat{Punct} = Table->New(Is => 'Punct', Desc => "[[:Punct:]]", Fuzzy => 0); |
570 | $Cat{Space} = Table->New(Is => 'Space', Desc => "[[:Space:]]", Fuzzy => 0); |
571 | $Cat{Title} = Table->New(Is => 'Title', Desc => "[[:Title:]]", Fuzzy => 0); |
572 | $Cat{Upper} = Table->New(Is => 'Upper', Desc => "[[:Upper:]]", Fuzzy => 0); |
573 | $Cat{XDigit} = Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0); |
574 | $Cat{Word} = Table->New(Is => 'Word', Desc => "[[:Word:]]", Fuzzy => 0); |
575 | $Cat{SpacePerl} = Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0); |
d73e5302 |
576 | |
cf25bb62 |
577 | my %To; |
578 | $To{Upper} = Table->New(); |
579 | $To{Lower} = Table->New(); |
580 | $To{Title} = Table->New(); |
581 | $To{Digit} = Table->New(); |
582 | |
583 | sub gencat($$$$) |
584 | { |
585 | my ($name, ## Name ("LATIN CAPITAL LETTER A") |
586 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) |
587 | $code, ## Code point (as an integer) |
588 | $op) = @_; |
589 | |
590 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc |
591 | |
592 | $Assigned->$op($code); |
593 | $Name->$op($code, $name); |
594 | $General->$op($code, $cat); |
595 | |
596 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) |
99598c8c |
597 | $Cat{$cat} ||= Table->New(Is => $cat, |
598 | Desc => "General Category '$cat'", |
599 | Fuzzy => 0); |
cf25bb62 |
600 | $Cat{$cat}->$op($code); |
601 | |
602 | ## add to the major category (e.g. "L", "N", "C", ...) |
99598c8c |
603 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, |
604 | Desc => "Major Category '$MajorCat'", |
605 | Fuzzy => 0); |
cf25bb62 |
606 | $Cat{$MajorCat}->$op($code); |
607 | |
608 | ($General{$name} ||= Table->New)->$op($code, $name); |
609 | |
610 | # 005F: SPACING UNDERSCORE |
611 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]/ || $code == 0x005F; |
612 | $Cat{Alnum}->$op($code) if $cat =~ /^[LMN]/; |
613 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; |
614 | |
615 | |
616 | |
617 | $Cat{Space}->$op($code) if $cat =~ /^Z/ |
618 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
619 | || $code == 0x000A # 000A: LINE FEED |
620 | || $code == 0x000B # 000B: VERTICAL TAB |
621 | || $code == 0x000C # 000C: FORM FEED |
622 | || $code == 0x000D; # 000D: CARRIAGE RETURN |
623 | |
624 | |
625 | $Cat{SpacePerl}->$op($code) if $cat =~ /^Z/ |
626 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
627 | || $code == 0x000A # 000A: LINE FEED |
628 | || $code == 0x000C # 000C: FORM FEED |
629 | || $code == 0x000D # 000D: CARRIAGE RETURN |
630 | || $code == 0x0085 # 0085: <NEXT LINE> |
631 | || $code == 0x2028 # 2028: LINE SEPARATOR |
632 | || $code == 0x2029;# 2029: PARAGRAPH SEP. |
633 | |
634 | $Cat{Blank}->$op($code) if $cat =~ /^Z[^lp]$/ |
635 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
636 | || $code == 0x0020; # 0020: SPACE |
637 | |
638 | $Cat{Digit}->$op($code) if $cat eq "Nd"; |
639 | $Cat{Upper}->$op($code) if $cat eq "Lu"; |
640 | $Cat{Lower}->$op($code) if $cat eq "Ll"; |
641 | $Cat{Title}->$op($code) if $cat eq "Lt"; |
642 | $Cat{ASCII}->$op($code) if $code <= 0x007F; |
643 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; |
644 | $Cat{Graph}->$op($code) if $cat =~ /^([LMNPS]|Co)/; |
645 | $Cat{Print}->$op($code) if $cat =~ /^([LMNPS]|Co|Zs)/; |
646 | $Cat{Punct}->$op($code) if $cat =~ /^P/; |
647 | |
648 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 |
649 | || ($code >= 0x41 && $code <= 0x46) ## A..F |
650 | || ($code >= 0x61 && $code <= 0x66); ## a..f |
651 | } |
d73e5302 |
652 | |
cf25bb62 |
653 | ## open ane read file..... |
551b6b6f |
654 | if (not open IN, "UnicodeData.txt") { |
655 | die "$0: UnicodeData.txt: $!\n"; |
cf25bb62 |
656 | } |
d73e5302 |
657 | |
a3a8c5f0 |
658 | ## |
659 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} |
660 | ## |
661 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") |
662 | |
663 | my %CodeToDeco; ## Maps code to decomp. list for chars with first |
664 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) |
665 | |
666 | ## This is filled in as we go.... |
99598c8c |
667 | my $CombAbove = Table->New(Is => '_CombAbove', |
668 | Desc => '(for internal casefolding use)', |
669 | Fuzzy => 0); |
a3a8c5f0 |
670 | |
cf25bb62 |
671 | while (<IN>) |
672 | { |
673 | next unless /^[0-9A-Fa-f]+;/; |
674 | s/\s+$//; |
675 | |
676 | my ($hexcode, ## code point in hex (e.g. "0041") |
677 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") |
678 | $cat, ## category (e.g. "Lu") |
679 | $comb, ## Canonical combining class (e.t. "230") |
680 | $bidi, ## directional category (e.g. "L") |
681 | $deco, ## decomposition mapping |
682 | $decimal, ## decimal digit value |
683 | $digit, ## digit value |
684 | $number, ## numeric value |
685 | $mirrored, ## mirrored |
686 | $unicode10, ## name in Unicode 1.0 |
687 | $comment, ## comment field |
688 | $upper, ## uppercase mapping |
689 | $lower, ## lowercase mapping |
690 | $title, ## titlecase mapping |
691 | ) = split(/\s*;\s*/); |
692 | |
693 | my $code = hex($hexcode); |
694 | |
a3a8c5f0 |
695 | if ($comb and $comb == 230) { |
696 | $CombAbove->Append($code); |
697 | $_Above_HexCodes{$hexcode} = 1; |
698 | } |
699 | |
700 | ## Used in building \p{_CanonDCIJ} |
701 | if ($deco and $deco =~ m/^006[9A]\b/) { |
702 | $CodeToDeco{$code} = $deco; |
703 | } |
704 | |
cf25bb62 |
705 | ## |
706 | ## There are a few pairs of lines like: |
707 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; |
708 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; |
709 | ## that define ranges. |
710 | ## |
711 | if ($name =~ /^<(.+), (First|Last)>$/) |
712 | { |
713 | $name = $1; |
714 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); |
99598c8c |
715 | #New_Prop(In => $name, $General{$name}, Fuzzy => 1); |
cf25bb62 |
716 | } |
717 | else |
718 | { |
719 | ## normal (single-character) lines |
720 | gencat($name, $cat, $code, 'Append'); |
721 | |
722 | # No Append() here since since several codes may map into one. |
723 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; |
724 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; |
725 | $To{Title}->RawAppendRange($code, $code, $title) if $title; |
726 | $To{Digit}->Append($code, $decimal) if length $decimal; |
727 | |
728 | $Bidi->Append($code, $bidi); |
729 | $Comb->Append($code, $comb) if $comb; |
730 | $Number->Append($code, $number) if length $number; |
731 | |
732 | $Mirrored->Append($code) if $mirrored eq "Y"; |
733 | |
99598c8c |
734 | $Bidi{$bidi} ||= Table->New(Is => "Bidi$bidi", |
735 | Desc => "Bi-directional category '$bidi'", |
736 | Fuzzy => 0); |
cf25bb62 |
737 | $Bidi{$bidi}->Append($code); |
738 | |
739 | if ($deco) |
740 | { |
741 | $Deco->Append($code, $deco); |
742 | if ($deco =~/^<(\w+)>/) |
743 | { |
744 | $Deco{Compat}->Append($code); |
745 | |
99598c8c |
746 | $DC{$1} ||= Table->New(Is => "DC$1", |
747 | Desc => "Compatible with '$1'", |
748 | Fuzzy => 0); |
cf25bb62 |
749 | $DC{$1}->Append($code); |
750 | } |
751 | else |
752 | { |
753 | $Deco{Canon}->Append($code); |
754 | } |
755 | } |
756 | } |
757 | } |
758 | close IN; |
d2d499f5 |
759 | |
cf25bb62 |
760 | ## |
761 | ## Tidy up a few special cases.... |
762 | ## |
d73e5302 |
763 | |
cf25bb62 |
764 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist |
99598c8c |
765 | New_Prop(Is => 'Cn', |
766 | $Cat{Cn}, |
767 | Desc => "General Category 'Cn' [not functional in Perl]", |
768 | Fuzzy => 0); |
d73e5302 |
769 | |
cf25bb62 |
770 | ## Unassigned is the same as 'Cn' |
5beb625e |
771 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0); |
d73e5302 |
772 | |
cf25bb62 |
773 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C |
d73e5302 |
774 | |
d73e5302 |
775 | |
cf25bb62 |
776 | # L& is Ll, Lu, and Lt. |
777 | New_Prop(Is => 'L&', |
778 | Table->Merge(@Cat{qw[Ll Lu Lt]}), |
99598c8c |
779 | Desc => '[\p{Ll}\p{Lu}\p{Lt}]', |
780 | Fuzzy => 0); |
d73e5302 |
781 | |
cf25bb62 |
782 | ## Any and All are all code points. |
99598c8c |
783 | my $Any = Table->New(Is => 'Any', |
784 | Desc => sprintf("[\\x{0000}-\\x{%X}]", |
785 | $LastUnicodeCodepoint), |
5beb625e |
786 | Fuzzy => 0); |
cf25bb62 |
787 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); |
d73e5302 |
788 | |
5beb625e |
789 | New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0); |
d73e5302 |
790 | |
a3a8c5f0 |
791 | ## |
792 | ## Build special properties for Perl's internal case-folding needs: |
793 | ## \p{_CaseIgnorable} |
794 | ## \p{_CanonDCIJ} |
795 | ## \p{_CombAbove} |
796 | ## _CombAbove was built above. Others are built here.... |
797 | ## |
798 | |
799 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] |
800 | New_Prop(Is => '_CaseIgnorable', |
801 | Table->Merge($Cat{Mn}, |
802 | 0x00AD, #SOFT HYPHEN |
803 | 0x2010), #HYPHEN |
99598c8c |
804 | Desc => '(for internal casefolding use)', |
805 | Fuzzy => 0); |
a3a8c5f0 |
806 | |
807 | |
808 | ## \p{_CanonDCIJ} is fairly complex... |
99598c8c |
809 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', |
810 | Desc => '(for internal casefolding use)', |
811 | Fuzzy => 0); |
a3a8c5f0 |
812 | ## It contains the ASCII 'i' and 'j'.... |
813 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") |
814 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") |
815 | ## ...and any character with a decomposition that starts with either of |
816 | ## those code points, but only if the decomposition does not have any |
817 | ## combining character with the "ABOVE" canonical combining class. |
818 | for my $code (sort { $a <=> $b} keys %CodeToDeco) |
819 | { |
820 | ## Need to ensure that all decomposition characters do not have |
821 | ## a %HexCodeToComb in %AboveCombClasses. |
822 | my $want = 1; |
823 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) |
824 | { |
825 | if (exists $_Above_HexCodes{$deco_hexcode}) { |
826 | ## one of the decmposition chars has an ABOVE combination |
827 | ## class, so we're not interested in this one |
828 | $want = 0; |
829 | last; |
830 | } |
831 | } |
832 | if ($want) { |
833 | $CanonCDIJ->Append($code); |
834 | } |
835 | } |
836 | |
837 | |
d73e5302 |
838 | |
cf25bb62 |
839 | ## |
840 | ## Now dump the files. |
841 | ## |
842 | $Name->Write("Name.pl"); |
843 | $Bidi->Write("Bidirectional.pl"); |
844 | $Comb->Write("CombiningClass.pl"); |
845 | $Deco->Write("Decomposition.pl"); |
846 | $Number->Write("Number.pl"); |
847 | $General->Write("Category.pl"); |
848 | |
849 | for my $to (sort keys %To) { |
850 | $To{$to}->Write("To/$to.pl"); |
d73e5302 |
851 | } |
852 | } |
853 | |
cf25bb62 |
854 | ## |
551b6b6f |
855 | ## Process LineBreak.txt |
cf25bb62 |
856 | ## |
551b6b6f |
857 | sub LineBreak_Txt() |
cf25bb62 |
858 | { |
551b6b6f |
859 | if (not open IN, "LineBreak.txt") { |
860 | die "$0: LineBreak.txt: $!\n"; |
cf25bb62 |
861 | } |
d73e5302 |
862 | |
cf25bb62 |
863 | my $Lbrk = Table->New(); |
864 | my %Lbrk; |
d73e5302 |
865 | |
cf25bb62 |
866 | while (<IN>) |
867 | { |
868 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; |
d73e5302 |
869 | |
cf25bb62 |
870 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); |
d73e5302 |
871 | |
cf25bb62 |
872 | $Lbrk->Append($first, $lbrk); |
d73e5302 |
873 | |
99598c8c |
874 | $Lbrk{$lbrk} ||= Table->New(Is => "Lbrk$lbrk", |
875 | Desc => "Linebreak category '$lbrk'", |
876 | Fuzzy => 0); |
cf25bb62 |
877 | $Lbrk{$lbrk}->Append($first); |
d73e5302 |
878 | |
cf25bb62 |
879 | if ($last) { |
880 | $Lbrk->Extend($last); |
881 | $Lbrk{$lbrk}->Extend($last); |
d73e5302 |
882 | } |
883 | } |
cf25bb62 |
884 | close IN; |
d73e5302 |
885 | |
cf25bb62 |
886 | $Lbrk->Write("Lbrk.pl"); |
887 | } |
d73e5302 |
888 | |
cf25bb62 |
889 | ## |
551b6b6f |
890 | ## Process ArabicShaping.txt. |
cf25bb62 |
891 | ## |
551b6b6f |
892 | sub ArabicShaping_txt() |
cf25bb62 |
893 | { |
551b6b6f |
894 | if (not open IN, "ArabicShaping.txt") { |
895 | die "$0: ArabicShaping.txt: $!\n"; |
cf25bb62 |
896 | } |
d73e5302 |
897 | |
cf25bb62 |
898 | my $ArabLink = Table->New(); |
899 | my $ArabLinkGroup = Table->New(); |
d73e5302 |
900 | |
cf25bb62 |
901 | while (<IN>) |
902 | { |
903 | next unless /^[0-9A-Fa-f]+;/; |
904 | s/\s+$//; |
d73e5302 |
905 | |
cf25bb62 |
906 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); |
907 | my $code = hex($hexcode); |
908 | $ArabLink->Append($code, $link); |
909 | $ArabLinkGroup->Append($code, $linkgroup); |
d73e5302 |
910 | } |
cf25bb62 |
911 | close IN; |
912 | |
913 | $ArabLink->Write("ArabLink.pl"); |
914 | $ArabLinkGroup->Write("ArabLnkGrp.pl"); |
d73e5302 |
915 | } |
916 | |
cf25bb62 |
917 | ## |
918 | ## Process Jamo.txt. |
919 | ## |
920 | sub Jamo_txt() |
921 | { |
922 | if (not open IN, "Jamo.txt") { |
923 | die "$0: Jamo.txt: $!\n"; |
924 | } |
925 | my $Short = Table->New(); |
d73e5302 |
926 | |
cf25bb62 |
927 | while (<IN>) |
928 | { |
929 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; |
930 | my ($code, $short) = (hex($1), $2); |
d73e5302 |
931 | |
cf25bb62 |
932 | $Short->Append($code, $short); |
d73e5302 |
933 | } |
cf25bb62 |
934 | close IN; |
935 | $Short->Write("JamoShort.pl"); |
d73e5302 |
936 | } |
937 | |
cf25bb62 |
938 | ## |
939 | ## Process Scripts.txt. |
940 | ## |
941 | sub Scripts_txt() |
942 | { |
943 | my @ScriptInfo; |
d73e5302 |
944 | |
cf25bb62 |
945 | if (not open(IN, "Scripts.txt")) { |
946 | die "$0: Scripts.txt: $!\n"; |
947 | } |
948 | while (<IN>) { |
949 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
950 | |
cf25bb62 |
951 | # Wait until all the scripts have been read since |
952 | # they are not listed in numeric order. |
953 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; |
954 | } |
955 | close IN; |
d73e5302 |
956 | |
cf25bb62 |
957 | # Now append the scripts properties in their code point order. |
d73e5302 |
958 | |
cf25bb62 |
959 | my %Script; |
960 | my $Scripts = Table->New(); |
d73e5302 |
961 | |
cf25bb62 |
962 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) |
963 | { |
964 | my ($first, $last, $name) = @$script; |
965 | $Scripts->Append($first, $name); |
d73e5302 |
966 | |
99598c8c |
967 | $Script{$name} ||= Table->New(Is => $name, |
968 | Desc => "Script '$name'", |
969 | Fuzzy => 1); |
cf25bb62 |
970 | $Script{$name}->Append($first, $name); |
d73e5302 |
971 | |
cf25bb62 |
972 | if ($last) { |
973 | $Scripts->Extend($last); |
974 | $Script{$name}->Extend($last); |
975 | } |
976 | } |
d73e5302 |
977 | |
cf25bb62 |
978 | $Scripts->Write("Scripts.pl"); |
d73e5302 |
979 | |
cf25bb62 |
980 | ## Common is everything not explicitly assigned to a Script |
981 | ## |
982 | ## ***shouldn't this be intersected with \p{Assigned}? ****** |
983 | ## |
99598c8c |
984 | New_Prop(Is => 'Common', |
985 | $Scripts->Invert, |
986 | Desc => 'Pseudo-Script of codepoints not in other Unicode scripts', |
987 | Fuzzy => 1); |
cf25bb62 |
988 | } |
d73e5302 |
989 | |
cf25bb62 |
990 | ## |
991 | ## Given a name like "Close Punctuation", return a regex (that when applied |
992 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", |
993 | ## "Close-Punctuation", etc.) |
994 | ## |
995 | ## Accept any space, dash, or underbar where in the official name there is |
996 | ## space or a dash (or underbar, but there never is). |
997 | ## |
998 | ## |
999 | sub NameToRegex($) |
1000 | { |
1001 | my $Name = shift; |
1002 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; |
1003 | return $Name; |
1004 | } |
d73e5302 |
1005 | |
cf25bb62 |
1006 | ## |
1007 | ## Process Blocks.txt. |
1008 | ## |
1009 | sub Blocks_txt() |
1010 | { |
1011 | my $Blocks = Table->New(); |
1012 | my %Blocks; |
d73e5302 |
1013 | |
cf25bb62 |
1014 | if (not open IN, "Blocks.txt") { |
1015 | die "$0: Blocks.txt: $!\n"; |
1016 | } |
d73e5302 |
1017 | |
cf25bb62 |
1018 | while (<IN>) |
1019 | { |
1020 | #next if not /Private Use$/; |
1021 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; |
d73e5302 |
1022 | |
cf25bb62 |
1023 | my ($first, $last, $name) = (hex($1), hex($2), $3); |
d73e5302 |
1024 | |
cf25bb62 |
1025 | $Blocks->Append($first, $name); |
76ccdbe2 |
1026 | |
99598c8c |
1027 | $Blocks{$name} ||= Table->New(In => $name, |
1028 | Desc => "Block '$name'", |
1029 | Fuzzy => 1); |
cf25bb62 |
1030 | $Blocks{$name}->Append($first, $name); |
76ccdbe2 |
1031 | |
cf25bb62 |
1032 | if ($last and $last != $first) { |
1033 | $Blocks->Extend($last); |
1034 | $Blocks{$name}->Extend($last); |
d73e5302 |
1035 | } |
d73e5302 |
1036 | } |
cf25bb62 |
1037 | close IN; |
1038 | |
1039 | $Blocks->Write("Blocks.pl"); |
d73e5302 |
1040 | } |
1041 | |
cf25bb62 |
1042 | ## |
1043 | ## Read in the PropList.txt. It contains extended properties not |
551b6b6f |
1044 | ## listed in the UnicodeData.txt, such as 'Other_Alphabetic': |
cf25bb62 |
1045 | ## alphabetic but not of the general category L; many modifiers |
1046 | ## belong to this extended property category: while they are not |
1047 | ## alphabets, they are alphabetic in nature. |
1048 | ## |
1049 | sub PropList_txt() |
1050 | { |
1051 | my @PropInfo; |
1052 | |
1053 | if (not open IN, "PropList.txt") { |
1054 | die "$0: PropList.txt: $!\n"; |
1055 | } |
d73e5302 |
1056 | |
cf25bb62 |
1057 | while (<IN>) |
1058 | { |
1059 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
1060 | |
cf25bb62 |
1061 | # Wait until all the extended properties have been read since |
1062 | # they are not listed in numeric order. |
1063 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; |
1064 | } |
1065 | close IN; |
71d929cb |
1066 | |
cf25bb62 |
1067 | # Now append the extended properties in their code point order. |
1068 | my $Props = Table->New(); |
1069 | my %Prop; |
71d929cb |
1070 | |
cf25bb62 |
1071 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) |
1072 | { |
1073 | my ($first, $last, $name) = @$prop; |
1074 | $Props->Append($first, $name); |
71d929cb |
1075 | |
99598c8c |
1076 | $Prop{$name} ||= Table->New(Is => $name, |
1077 | Desc => "Extended property '$name'", |
1078 | Fuzzy => 1); |
cf25bb62 |
1079 | $Prop{$name}->Append($first, $name); |
71d929cb |
1080 | |
cf25bb62 |
1081 | if ($last) { |
1082 | $Props->Extend($last); |
1083 | $Prop{$name}->Extend($last); |
1084 | } |
71d929cb |
1085 | } |
d73e5302 |
1086 | |
cf25bb62 |
1087 | # Alphabetic is L and Other_Alphabetic. |
99598c8c |
1088 | New_Prop(Is => 'Alphabetic', |
cf25bb62 |
1089 | Table->Merge($Cat{L}, $Prop{Other_Alphabetic}), |
99598c8c |
1090 | Desc => '[\p{L}\p{OtherAlphabetic}]', # use canonical names here |
1091 | Fuzzy => 1); |
cf25bb62 |
1092 | |
1093 | # Lowercase is Ll and Other_Lowercase. |
99598c8c |
1094 | New_Prop(Is => 'Lowercase', |
cf25bb62 |
1095 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), |
99598c8c |
1096 | Desc => '[\p{Ll}\p{OtherLowercase}]', # use canonical names here |
1097 | Fuzzy => 1); |
cf25bb62 |
1098 | |
1099 | # Uppercase is Lu and Other_Uppercase. |
1100 | New_Prop(Is => 'Uppercase', |
1101 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), |
99598c8c |
1102 | Desc => '[\p{Lu}\p{Other_Uppercase}]', # use canonical names here |
1103 | Fuzzy => 1); |
cf25bb62 |
1104 | |
1105 | # Math is Sm and Other_Math. |
1106 | New_Prop(Is => 'Math', |
1107 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), |
99598c8c |
1108 | Desc => '[\p{Sm}\p{OtherMath}]', # use canonical names here |
1109 | Fuzzy => 1); |
cf25bb62 |
1110 | |
1111 | # ID_Start is Ll, Lu, Lt, Lm, Lo, and Nl. |
1112 | New_Prop(Is => 'ID_Start', |
1113 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}), |
99598c8c |
1114 | Desc => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}]', |
1115 | Fuzzy => 1); |
cf25bb62 |
1116 | |
1117 | # ID_Continue is ID_Start, Mn, Mc, Nd, and Pc. |
1118 | New_Prop(Is => 'ID_Continue', |
1119 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}), |
99598c8c |
1120 | Desc => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}]', |
1121 | Fuzzy => 1); |
d73e5302 |
1122 | } |
1123 | |
cf25bb62 |
1124 | sub Make_GC_Aliases() |
1125 | { |
1126 | ## |
1127 | ## The mapping from General Category long forms to short forms is |
1128 | ## currently hardwired here since no simple data file in the UCD |
1129 | ## seems to do that. Unicode 3.2 will assumedly correct this. |
1130 | ## |
1131 | my %Is = ( |
d73e5302 |
1132 | 'Letter' => 'L', |
e150c829 |
1133 | 'Uppercase_Letter' => 'Lu', |
1134 | 'Lowercase_Letter' => 'Ll', |
1135 | 'Titlecase_Letter' => 'Lt', |
1136 | 'Modifier_Letter' => 'Lm', |
1137 | 'Other_Letter' => 'Lo', |
d73e5302 |
1138 | |
1139 | 'Mark' => 'M', |
e150c829 |
1140 | 'Non_Spacing_Mark' => 'Mn', |
1141 | 'Spacing_Mark' => 'Mc', |
1142 | 'Enclosing_Mark' => 'Me', |
d73e5302 |
1143 | |
1144 | 'Separator' => 'Z', |
e150c829 |
1145 | 'Space_Separator' => 'Zs', |
1146 | 'Line_Separator' => 'Zl', |
1147 | 'Paragraph_Separator' => 'Zp', |
d73e5302 |
1148 | |
1149 | 'Number' => 'N', |
e150c829 |
1150 | 'Decimal_Number' => 'Nd', |
1151 | 'Letter_Number' => 'Nl', |
1152 | 'Other_Number' => 'No', |
d73e5302 |
1153 | |
1154 | 'Punctuation' => 'P', |
e150c829 |
1155 | 'Connector_Punctuation' => 'Pc', |
1156 | 'Dash_Punctuation' => 'Pd', |
1157 | 'Open_Punctuation' => 'Ps', |
1158 | 'Close_Punctuation' => 'Pe', |
1159 | 'Initial_Punctuation' => 'Pi', |
1160 | 'Final_Punctuation' => 'Pf', |
1161 | 'Other_Punctuation' => 'Po', |
d73e5302 |
1162 | |
1163 | 'Symbol' => 'S', |
e150c829 |
1164 | 'Math_Symbol' => 'Sm', |
1165 | 'Currency_Symbol' => 'Sc', |
1166 | 'Modifier_Symbol' => 'Sk', |
1167 | 'Other_Symbol' => 'So', |
d73e5302 |
1168 | |
1169 | 'Other' => 'C', |
1170 | 'Control' => 'Cc', |
1171 | 'Format' => 'Cf', |
1172 | 'Surrogate' => 'Cs', |
1173 | 'Private Use' => 'Co', |
e150c829 |
1174 | 'Unassigned' => 'Cn', |
cf25bb62 |
1175 | ); |
d2d499f5 |
1176 | |
cf25bb62 |
1177 | ## make the aliases.... |
1178 | while (my ($Alias, $Name) = each %Is) { |
99598c8c |
1179 | New_Alias(Is => $Alias, SameAs => $Name, Fuzzy => 1); |
cf25bb62 |
1180 | } |
1181 | } |
d2d499f5 |
1182 | |
5beb625e |
1183 | |
1184 | ## |
1185 | ## These are used in: |
1186 | ## MakePropTestScript() |
1187 | ## WriteAllMappings() |
1188 | ## for making the test script. |
1189 | ## |
1190 | my %FuzzyNameToTest; |
1191 | my %ExactNameToTest; |
1192 | |
1193 | |
1194 | ## This used only for making the test script |
1195 | sub GenTests($$$$) |
1196 | { |
1197 | my $FH = shift; |
1198 | my $Prop = shift; |
1199 | my $MatchCode = shift; |
1200 | my $FailCode = shift; |
1201 | |
1202 | if (defined $MatchCode) { |
1203 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode; |
1204 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode; |
1205 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode; |
1206 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode; |
1207 | } |
1208 | if (defined $FailCode) { |
1209 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode; |
1210 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode; |
1211 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode; |
1212 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode; |
1213 | } |
1214 | } |
1215 | |
1216 | ## This used only for making the test script |
1217 | sub ExpectError($$) |
1218 | { |
1219 | my $FH = shift; |
1220 | my $prop = shift; |
1221 | |
1222 | print $FH qq/Error('\\p{$prop}');\n/; |
1223 | print $FH qq/Error('\\P{$prop}');\n/; |
1224 | } |
1225 | |
1226 | ## This used only for making the test script |
1227 | my @GoodSeps = ( |
1228 | " ", |
1229 | "-", |
1230 | " \t ", |
1231 | "", |
1232 | "", |
1233 | "_", |
1234 | ); |
1235 | my @BadSeps = ( |
1236 | "--", |
1237 | "__", |
1238 | " _", |
1239 | "/" |
1240 | ); |
1241 | |
1242 | ## This used only for making the test script |
1243 | sub RandomlyFuzzifyName($;$) |
1244 | { |
1245 | my $Name = shift; |
1246 | my $WantError = shift; ## if true, make an error |
1247 | |
1248 | my @parts; |
1249 | for my $part (split /[-\s_]+/, $Name) |
1250 | { |
1251 | if (@parts) { |
1252 | if ($WantError and rand() < 0.3) { |
1253 | push @parts, $BadSeps[rand(@BadSeps)]; |
1254 | $WantError = 0; |
1255 | } else { |
1256 | push @parts, $GoodSeps[rand(@GoodSeps)]; |
1257 | } |
1258 | } |
1259 | my $switch = int rand(4); |
1260 | if ($switch == 0) { |
1261 | push @parts, uc $part; |
1262 | } elsif ($switch == 1) { |
1263 | push @parts, lc $part; |
1264 | } elsif ($switch == 2) { |
1265 | push @parts, ucfirst $part; |
1266 | } else { |
1267 | push @parts, $part; |
1268 | } |
1269 | } |
1270 | my $new = join('', @parts); |
1271 | |
1272 | if ($WantError) { |
1273 | if (rand() >= 0.5) { |
1274 | $new .= $BadSeps[rand(@BadSeps)]; |
1275 | } else { |
1276 | $new = $BadSeps[rand(@BadSeps)] . $new; |
1277 | } |
1278 | } |
1279 | return $new; |
1280 | } |
1281 | |
1282 | ## This used only for making the test script |
1283 | sub MakePropTestScript() |
1284 | { |
1285 | ## this written directly -- it's huge. |
1286 | if (not open OUT, ">TestProp.pl") { |
1287 | die "$0: TestProp.pl: $!\n"; |
1288 | } |
1289 | print OUT <DATA>; |
1290 | |
1291 | while (my ($Name, $Table) = each %ExactNameToTest) |
1292 | { |
1293 | GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode); |
1294 | ExpectError(*OUT, uc $Name) if uc $Name ne $Name; |
1295 | ExpectError(*OUT, lc $Name) if lc $Name ne $Name; |
1296 | } |
1297 | |
1298 | |
1299 | while (my ($Name, $Table) = each %FuzzyNameToTest) |
1300 | { |
1301 | my $Orig = $CanonicalToOrig{$Name}; |
1302 | my %Names = ( |
1303 | $Name => 1, |
1304 | $Orig => 1, |
1305 | RandomlyFuzzifyName($Orig) => 1 |
1306 | ); |
1307 | |
1308 | for my $N (keys %Names) { |
1309 | GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode); |
1310 | } |
1311 | |
1312 | ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR')); |
1313 | } |
1314 | |
1315 | print OUT "Finished();\n"; |
1316 | close OUT; |
1317 | } |
1318 | |
1319 | |
1320 | ## |
1321 | ## These are used only in: |
1322 | ## RegisterFileForName() |
1323 | ## WriteAllMappings() |
1324 | ## |
1325 | my %Exact; ## will become %utf8::Exact; |
1326 | my %Canonical; ## will become %utf8::Canonical; |
1327 | my %CaComment; ## Comment for %Canonical entry of same key |
1328 | |
1329 | ## |
1330 | ## Given info about a name and a datafile that it should be associated with, |
1331 | ## register that assocation in %Exact and %Canonical. |
1332 | sub RegisterFileForName($$$$) |
1333 | { |
1334 | my $Type = shift; |
1335 | my $Name = shift; |
1336 | my $IsFuzzy = shift; |
1337 | my $filename = shift; |
1338 | |
1339 | ## |
1340 | ## Now in details for the mapping. $Type eq 'Is' has the |
1341 | ## Is removed, as it will be removed in utf8_heavy when this |
1342 | ## data is being checked. In keeps its "In", but a second |
1343 | ## sans-In record is written if it doesn't conflict with |
1344 | ## anything already there. |
1345 | ## |
1346 | if (not $IsFuzzy) |
1347 | { |
1348 | if ($Type eq 'Is') { |
1349 | die "oops[$Name]" if $Exact{$Name}; |
1350 | $Exact{$Name} = $filename; |
1351 | } else { |
1352 | die "oops[$Type$Name]" if $Exact{"$Type$Name"}; |
1353 | $Exact{"$Type$Name"} = $filename; |
1354 | $Exact{$Name} = $filename if not $Exact{$Name}; |
1355 | } |
1356 | } |
1357 | else |
1358 | { |
1359 | my $CName = lc $Name; |
1360 | if ($Type eq 'Is') { |
1361 | die "oops[$CName]" if $Canonical{$CName}; |
1362 | $Canonical{$CName} = $filename; |
1363 | $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2; |
1364 | } else { |
1365 | die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"}; |
1366 | $Canonical{lc "$Type$CName"} = $filename; |
1367 | $CaComment{lc "$Type$CName"} = "$Type$Name"; |
1368 | if (not $Canonical{$CName}) { |
1369 | $Canonical{$CName} = $filename; |
1370 | $CaComment{$CName} = "$Type$Name"; |
1371 | } |
1372 | } |
1373 | } |
1374 | } |
1375 | |
cf25bb62 |
1376 | ## |
1377 | ## Writes the info accumulated in |
1378 | ## |
1379 | ## %TableInfo; |
1380 | ## %FuzzyNames; |
1381 | ## %AliasInfo; |
1382 | ## |
1383 | ## |
1384 | sub WriteAllMappings() |
1385 | { |
99598c8c |
1386 | my @MAP; |
1387 | |
5beb625e |
1388 | my %BaseNames; ## Base names already used (for avoiding 8.3 conflicts) |
cf25bb62 |
1389 | |
5beb625e |
1390 | ## 'Is' *MUST* come first, so its names have precidence over 'In's |
1391 | for my $Type ('Is', 'In') |
1392 | { |
1393 | my %RawNameToFile; ## a per-$Type cache |
cf25bb62 |
1394 | |
5beb625e |
1395 | for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}}) |
cf25bb62 |
1396 | { |
5beb625e |
1397 | ## Note: $Name is already canonical |
99598c8c |
1398 | my $Table = $TableInfo{$Type}->{$Name}; |
5beb625e |
1399 | my $IsFuzzy = $FuzzyNames{$Type}->{$Name}; |
99598c8c |
1400 | |
1401 | ## Need an 8.3 safe filename (which means "an 8 safe" $filename) |
5beb625e |
1402 | my $filename; |
cf25bb62 |
1403 | { |
5beb625e |
1404 | ## 'Is' items lose 'Is' from the basename. |
1405 | $filename = $Type eq 'Is' ? $Name : "$Type$Name"; |
1406 | |
1407 | $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_" |
1408 | substr($filename, 8) = '' if length($filename) > 8; |
1409 | |
1410 | ## |
1411 | ## Make sure the basename doesn't conflict with something we |
1412 | ## might have already written. If we have, say, |
1413 | ## InGreekExtended1 |
1414 | ## InGreekExtended2 |
1415 | ## they become |
1416 | ## InGreekE |
1417 | ## InGreek2 |
1418 | ## |
1419 | while (my $num = $BaseNames{lc $filename}++) |
1420 | { |
1421 | $num++; ## so basenames with numbers start with '2', which |
1422 | ## just looks more natural. |
1423 | ## Want to append $num, but if it'll make the basename longer |
1424 | ## than 8 characters, pre-truncate $filename so that the result |
1425 | ## is acceptable. |
1426 | my $delta = length($filename) + length($num) - 8; |
1427 | if ($delta > 0) { |
1428 | substr($filename, -$delta) = $num; |
1429 | } else { |
1430 | $filename .= $num; |
1431 | } |
99598c8c |
1432 | } |
5beb625e |
1433 | }; |
99598c8c |
1434 | |
1435 | ## |
1436 | ## Construct a nice comment to add to the file, and build data |
1437 | ## for the "./Properties" file along the way. |
1438 | ## |
1439 | my $Comment; |
1440 | { |
1441 | my $Desc = $TableDesc{$Type}->{$Name} || ""; |
1442 | ## get list of names this table is reference by |
1443 | my @Supported = $Name; |
1444 | while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} }) |
1445 | { |
1446 | if ($Orig eq $Name) { |
1447 | push @Supported, $Alias; |
1448 | } |
1449 | } |
1450 | |
1451 | my $TypeToShow = $Type eq 'Is' ? "" : $Type; |
1452 | my $OrigProp; |
1453 | |
1454 | $Comment = "This file supports:\n"; |
1455 | for my $N (@Supported) |
1456 | { |
1457 | my $IsFuzzy = $FuzzyNames{$Type}->{$N}; |
5beb625e |
1458 | my $Prop = "\\p{$TypeToShow$Name}"; |
99598c8c |
1459 | $OrigProp = $Prop if not $OrigProp; #cache for aliases |
1460 | if ($IsFuzzy) { |
1461 | $Comment .= "\t$Prop (and fuzzy permutations)\n"; |
1462 | } else { |
1463 | $Comment .= "\t$Prop\n"; |
1464 | } |
1465 | my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)"; |
1466 | |
1467 | push @MAP, sprintf("%s %-42s %s\n", |
1468 | $IsFuzzy ? '*' : ' ', $Prop, $MyDesc); |
1469 | } |
1470 | if ($Desc) { |
1471 | $Comment .= "\nMeaning: $Desc\n"; |
1472 | } |
1473 | |
1474 | } |
cf25bb62 |
1475 | ## |
1476 | ## Okay, write the file... |
1477 | ## |
5beb625e |
1478 | $Table->Write("lib/$filename.pl", $Comment); |
99598c8c |
1479 | |
5beb625e |
1480 | ## and register it |
1481 | $RawNameToFile{$Name} = $filename; |
1482 | RegisterFileForName($Type => $Name, $IsFuzzy, $filename); |
cf25bb62 |
1483 | |
5beb625e |
1484 | if ($IsFuzzy) |
1485 | { |
1486 | my $CName = CanonicalName($Type . '_'. $Name); |
1487 | $FuzzyNameToTest{$Name} = $Table if !$FuzzyNameToTest{$Name}; |
1488 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1489 | } else { |
1490 | $ExactNameToTest{$Name} = $Table; |
cf25bb62 |
1491 | } |
1492 | |
cf25bb62 |
1493 | } |
1494 | |
5beb625e |
1495 | ## Register aliase info |
1496 | for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}}) |
cf25bb62 |
1497 | { |
5beb625e |
1498 | my $Alias = $AliasInfo{$Type}->{$Name}; |
1499 | my $IsFuzzy = $FuzzyNames{$Type}->{$Alias}; |
1500 | my $filename = $RawNameToFile{$Name}; |
1501 | die "oops [$Alias]->[$Name]" if not $filename; |
1502 | RegisterFileForName($Type => $Alias, $IsFuzzy, $filename); |
1503 | |
1504 | my $Table = $TableInfo{$Type}->{$Name}; |
1505 | die "oops" if not $Table; |
1506 | if ($IsFuzzy) |
1507 | { |
1508 | my $CName = CanonicalName($Type .'_'. $Alias); |
1509 | $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias}; |
1510 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1511 | } else { |
1512 | $ExactNameToTest{$Alias} = $Table; |
1513 | } |
cf25bb62 |
1514 | } |
5beb625e |
1515 | } |
cf25bb62 |
1516 | |
5beb625e |
1517 | ## |
1518 | ## Write out the property list |
1519 | ## |
1520 | { |
1521 | my @OUT = ( |
1522 | "##\n", |
1523 | "## This file created by $0\n", |
1524 | "## List of built-in \\p{...}/\\P{...} properties.\n", |
1525 | "##\n", |
1526 | "## '*' means name may be 'fuzzy'\n", |
1527 | "##\n\n", |
1528 | sort { substr($a,2) cmp substr($b, 2) } @MAP, |
1529 | ); |
1530 | WriteIfChanged('Properties', @OUT); |
1531 | } |
cf25bb62 |
1532 | |
5beb625e |
1533 | use Text::Tabs (); ## using this makes the files about half the size |
1534 | |
1535 | ## Write Exact.pl |
1536 | { |
1537 | my @OUT = ( |
1538 | $HEADER, |
1539 | "##\n", |
1540 | "## Data in this file used by ../utf8_heavy.pl\n", |
1541 | "##\n\n", |
1542 | "## Mapping from name to filename in ./lib\n", |
1543 | "%utf8::Exact = (\n", |
1544 | ); |
cf25bb62 |
1545 | |
cf25bb62 |
1546 | for my $Name (sort keys %Exact) |
1547 | { |
1548 | my $File = $Exact{$Name}; |
5beb625e |
1549 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1550 | my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/); |
1551 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
1552 | } |
5beb625e |
1553 | push @OUT, ");\n1;\n"; |
1554 | |
1555 | WriteIfChanged('Exact.pl', @OUT); |
1556 | } |
cf25bb62 |
1557 | |
5beb625e |
1558 | ## Write Canonical.pl |
1559 | { |
1560 | my @OUT = ( |
1561 | $HEADER, |
1562 | "##\n", |
1563 | "## Data in this file used by ../utf8_heavy.pl\n", |
1564 | "##\n\n", |
1565 | "## Mapping from lc(canonical name) to filename in ./lib\n", |
1566 | "%utf8::Canonical = (\n", |
1567 | ); |
1568 | my $Trail = ""; ## used just to keep the spacing pretty |
1569 | for my $Name (sort keys %Canonical) |
cf25bb62 |
1570 | { |
5beb625e |
1571 | my $File = $Canonical{$Name}; |
1572 | if ($CaComment{$Name}) { |
1573 | push @OUT, "\n" if not $Trail; |
1574 | push @OUT, " # $CaComment{$Name}\n"; |
1575 | $Trail = "\n"; |
1576 | } else { |
1577 | $Trail = ""; |
cf25bb62 |
1578 | } |
5beb625e |
1579 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1580 | my $Text = sprintf(" %-41s => %s,\n$Trail", $Name, qq/'$File'/); |
1581 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
1582 | } |
5beb625e |
1583 | push @OUT, ");\n1\n"; |
1584 | WriteIfChanged('Canonical.pl', @OUT); |
d2d499f5 |
1585 | } |
5beb625e |
1586 | |
1587 | MakePropTestScript() if $MakeTestScript; |
d2d499f5 |
1588 | } |
1589 | |
5beb625e |
1590 | |
551b6b6f |
1591 | sub SpecialCasing_txt() |
cf25bb62 |
1592 | { |
1593 | # |
1594 | # Read in the special cases. |
1595 | # |
983ffd37 |
1596 | |
cf25bb62 |
1597 | my %CaseInfo; |
1598 | |
551b6b6f |
1599 | if (not open IN, "SpecialCasing.txt") { |
1600 | die "$0: SpecialCasing.txt: $!\n"; |
cf25bb62 |
1601 | } |
1602 | while (<IN>) { |
1603 | next unless /^[0-9A-Fa-f]+;/; |
1604 | s/\#.*//; |
1605 | s/\s+$//; |
1606 | |
1607 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); |
1608 | |
1609 | if ($condition) { # not implemented yet |
1610 | print "# SKIPPING $_\n" if $Verbose; |
1611 | next; |
1612 | } |
1613 | |
1614 | # Wait until all the special cases have been read since |
1615 | # they are not listed in numeric order. |
1616 | my $ix = hex($code); |
1617 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ]; |
1618 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ]; |
1619 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ]; |
1620 | } |
1621 | close IN; |
1622 | |
1623 | # Now write out the special cases properties in their code point order. |
1624 | # Prepend them to the To/{Upper,Lower,Title}.pl. |
1625 | |
1626 | for my $case (qw(Lower Title Upper)) |
1627 | { |
1628 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; |
cf25bb62 |
1629 | |
5beb625e |
1630 | my @OUT = ( |
1631 | $HEADER, "\n", |
1632 | "%utf8::ToSpec$case =\n(\n", |
1633 | ); |
cf25bb62 |
1634 | |
1635 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { |
1636 | my ($ix, $code, $to) = @$prop; |
1637 | my $tostr = |
1638 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; |
5beb625e |
1639 | push @OUT, sprintf qq['%04X' => "$tostr",\n], $ix; |
cf25bb62 |
1640 | } |
5beb625e |
1641 | push @OUT, ( |
1642 | ");\n\n", |
1643 | "return <<'END';\n", |
1644 | $NormalCase, |
1645 | "END\n" |
1646 | ); |
1647 | WriteIfChanged("To/$case.pl", @OUT); |
d2d499f5 |
1648 | } |
d2d499f5 |
1649 | } |
1650 | |
c4051cc5 |
1651 | # |
1652 | # Read in the case foldings. |
1653 | # |
551b6b6f |
1654 | # We will do full case folding, C + F + I (see CaseFolding.txt). |
c4051cc5 |
1655 | # |
551b6b6f |
1656 | sub CaseFolding_txt() |
cf25bb62 |
1657 | { |
551b6b6f |
1658 | if (not open IN, "CaseFolding.txt") { |
1659 | die "$0: CaseFolding.txt: $!\n"; |
cf25bb62 |
1660 | } |
c4051cc5 |
1661 | |
cf25bb62 |
1662 | my $Fold = Table->New(); |
c4051cc5 |
1663 | my %Fold; |
1664 | |
cf25bb62 |
1665 | while (<IN>) { |
254ba52a |
1666 | # Skip status 'S', simple case folding |
c4051cc5 |
1667 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; |
1668 | |
cf25bb62 |
1669 | my ($code, $status, $fold) = (hex($1), $2, $3); |
c4051cc5 |
1670 | |
1671 | if ($status eq 'C') { # Common: one-to-one folding |
254ba52a |
1672 | # No append() since several codes may fold into one. |
cf25bb62 |
1673 | $Fold->RawAppendRange($code, $code, $fold); |
c4051cc5 |
1674 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I |
cf25bb62 |
1675 | $Fold{$code} = $fold; |
c4051cc5 |
1676 | } |
1677 | } |
cf25bb62 |
1678 | close IN; |
c4051cc5 |
1679 | |
cf25bb62 |
1680 | $Fold->Write("To/Fold.pl"); |
c4051cc5 |
1681 | |
1682 | # |
1683 | # Prepend the special foldings to the common foldings. |
1684 | # |
c4051cc5 |
1685 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; |
5beb625e |
1686 | |
1687 | my @OUT = ( |
1688 | $HEADER, "\n", |
1689 | "%utf8::ToSpecFold =\n(\n", |
1690 | ); |
cf25bb62 |
1691 | for my $code (sort { $a <=> $b } keys %Fold) { |
1692 | my $foldstr = |
1693 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; |
5beb625e |
1694 | push @OUT, sprintf qq['%04X' => "$foldstr",\n], $code; |
c4051cc5 |
1695 | } |
5beb625e |
1696 | push @OUT, ( |
1697 | ");\n\n", |
1698 | "return <<'END';\n", |
1699 | $CommonFold, |
1700 | "END\n", |
1701 | ); |
1702 | |
1703 | WriteIfChanged("To/Fold.pl", @OUT); |
c4051cc5 |
1704 | } |
1705 | |
cf25bb62 |
1706 | ## Do it.... |
1707 | |
1708 | Unicode_Txt(); |
1709 | Make_GC_Aliases(); |
1710 | PropList_txt(); |
1711 | |
1712 | Scripts_txt(); |
1713 | Blocks_txt(); |
1714 | |
5beb625e |
1715 | WriteAllMappings(); |
1716 | |
551b6b6f |
1717 | LineBreak_Txt(); |
1718 | ArabicShaping_txt(); |
cf25bb62 |
1719 | Jamo_txt(); |
551b6b6f |
1720 | SpecialCasing_txt(); |
1721 | CaseFolding_txt(); |
cf25bb62 |
1722 | |
5beb625e |
1723 | exit(0); |
cf25bb62 |
1724 | |
5beb625e |
1725 | ## TRAILING CODE IS USED BY MakePropTestScript() |
1726 | __DATA__ |
1727 | use strict; |
1728 | use warnings; |
1729 | |
1730 | my $Tests = 0; |
1731 | my $Fails = 0; |
cf25bb62 |
1732 | |
5beb625e |
1733 | sub Expect($$$) |
1734 | { |
1735 | my $Expect = shift; |
1736 | my $String = shift; |
1737 | my $Regex = shift; |
1738 | my $Line = (caller)[2]; |
1739 | |
1740 | $Tests++; |
1741 | my $RegObj; |
1742 | my $result = eval { |
1743 | $RegObj = qr/$Regex/; |
1744 | $String =~ $RegObj ? 1 : 0 |
1745 | }; |
1746 | |
1747 | if (not defined $result) { |
1748 | print "couldn't compile /$Regex/ on $0 line $Line: $@\n"; |
1749 | $Fails++; |
1750 | } elsif ($result ^ $Expect) { |
1751 | print "bad result (expected $Expect) on $0 line $Line: $@\n"; |
1752 | $Fails++; |
1753 | } |
1754 | } |
d73e5302 |
1755 | |
5beb625e |
1756 | sub Error($) |
1757 | { |
1758 | my $Regex = shift; |
1759 | $Tests++; |
1760 | if (eval { 'x' =~ qr/$Regex/; 1 }) { |
1761 | $Fails++; |
1762 | my $Line = (caller)[2]; |
1763 | print "expected error for /$Regex/ on $0 line $Line: $@\n"; |
1764 | } |
1765 | } |
1766 | |
1767 | sub Finished() |
1768 | { |
1769 | if ($Fails == 0) { |
1770 | print "All $Tests tests passed.\n"; |
1771 | exit(0); |
1772 | } else { |
1773 | print "$Tests tests, $Fails failed!\n"; |
1774 | exit(-1); |
1775 | } |
1776 | } |