Commit | Line | Data |
---|---|---|
920dae64 AT |
1 | ## !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!! |
2 | ## Any files created or read by this program should be listed in 'mktables.lst' | |
3 | ||
4 | #!/usr/bin/perl -w | |
5 | require 5.008; # Needs pack "U". Probably safest to run on 5.8.x | |
6 | use strict; | |
7 | use Carp; | |
8 | use File::Spec; | |
9 | ||
10 | ## | |
11 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) | |
12 | ## from the Unicode database files (lib/unicore/*.txt). | |
13 | ## | |
14 | ||
15 | ## "Fuzzy" means this section in Unicode TR18: | |
16 | ## | |
17 | ## The recommended names for UCD properties and property values are in | |
18 | ## PropertyAliases.txt [Prop] and PropertyValueAliases.txt | |
19 | ## [PropValue]. There are both abbreviated names and longer, more | |
20 | ## descriptive names. It is strongly recommended that both names be | |
21 | ## recognized, and that loose matching of property names be used, | |
22 | ## whereby the case distinctions, whitespace, hyphens, and underbar | |
23 | ## are ignored. | |
24 | ||
25 | ## Base names already used in lib/gc_sc (for avoiding 8.3 conflicts) | |
26 | my %BaseNames; | |
27 | ||
28 | ## | |
29 | ## Process any args. | |
30 | ## | |
31 | my $Verbose = 0; | |
32 | my $MakeTestScript = 0; | |
33 | my $AlwaysWrite = 0; | |
34 | my $UseDir = ""; | |
35 | my $FileList = "$0.lst"; | |
36 | my $MakeList = 0; | |
37 | ||
38 | while (@ARGV) | |
39 | { | |
40 | my $arg = shift @ARGV; | |
41 | if ($arg eq '-v') { | |
42 | $Verbose = 1; | |
43 | } elsif ($arg eq '-q') { | |
44 | $Verbose = 0; | |
45 | } elsif ($arg eq '-w') { | |
46 | $AlwaysWrite = 1; # update the files even if they havent changed | |
47 | $FileList = ""; | |
48 | } elsif ($arg eq '-maketest') { | |
49 | $MakeTestScript = 1; | |
50 | } elsif ($arg eq '-makelist') { | |
51 | $MakeList = 1; | |
52 | } elsif ($arg eq '-C' && defined ($UseDir = shift)) { | |
53 | -d $UseDir or die "Unknown directory '$UseDir'"; | |
54 | } elsif ($arg eq '-L' && defined ($FileList = shift)) { | |
55 | -e $FileList or die "Filelist '$FileList' doesn't appear to exist!"; | |
56 | } else { | |
57 | die "usage: $0 [-v|-q|-w|-C dir|-L filelist] [-maketest] [-makelist]\n", | |
58 | " -v : Verbose Mode\n", | |
59 | " -q : Quiet Mode\n", | |
60 | " -w : Write files regardless\n", | |
61 | " -maketest : Make test script\n", | |
62 | " -makelist : Rewrite the file list based on current setup\n", | |
63 | " -L filelist : Use this file list, (defaults to $0)\n", | |
64 | " -C dir : Change to this directory before proceding\n"; | |
65 | } | |
66 | } | |
67 | ||
68 | if ($FileList) { | |
69 | print "Reading file list '$FileList'\n" | |
70 | if $Verbose; | |
71 | open my $fh,"<",$FileList or die "Failed to read '$FileList':$!"; | |
72 | my @input; | |
73 | my @output; | |
74 | for my $list ( \@input, \@output ) { | |
75 | while (<$fh>) { | |
76 | s/^ \s+ | \s+ $//xg; | |
77 | next if /^ \s* (?: \# .* )? $/x; | |
78 | last if /^ =+ $/x; | |
79 | my ( $file ) = split /\t/, $_; | |
80 | push @$list, $file; | |
81 | } | |
82 | my %dupe; | |
83 | @$list = grep !$dupe{ $_ }++, @$list; | |
84 | } | |
85 | close $fh; | |
86 | die "No input or output files in '$FileList'!" | |
87 | if !@input or !@output; | |
88 | if ( $MakeList ) { | |
89 | foreach my $file (@output) { | |
90 | unlink $file; | |
91 | } | |
92 | } | |
93 | if ( $Verbose ) { | |
94 | print "Expecting ".scalar( @input )." input files. ", | |
95 | "Checking ".scalar( @output )." output files.\n"; | |
96 | } | |
97 | # we set maxtime to be the youngest input file, including $0 itself. | |
98 | my $maxtime = -M $0; # do this before the chdir! | |
99 | if ($UseDir) { | |
100 | chdir $UseDir or die "Failed to chdir to '$UseDir':$!"; | |
101 | } | |
102 | foreach my $in (@input) { | |
103 | my $time = -M $in; | |
104 | die "Missing input file '$in'" unless defined $time; | |
105 | $maxtime = $time if $maxtime < $time; | |
106 | } | |
107 | ||
108 | # now we check to see if any output files are older than maxtime, if | |
109 | # they are we need to continue on, otherwise we can presumably bail. | |
110 | my $ok = 1; | |
111 | foreach my $out (@output) { | |
112 | if ( ! -e $out ) { | |
113 | print "'$out' is missing.\n" | |
114 | if $Verbose; | |
115 | $ok = 0; | |
116 | last; | |
117 | } | |
118 | if ( -M $out > $maxtime ) { | |
119 | print "'$out' is too old.\n" | |
120 | if $Verbose; | |
121 | $ok = 0; | |
122 | last; | |
123 | } | |
124 | } | |
125 | if ($ok) { | |
126 | print "Files seem to be ok, not bothering to rebuild.\n"; | |
127 | exit(0); | |
128 | } | |
129 | print "Must rebuild tables.\n" | |
130 | if $Verbose; | |
131 | } else { | |
132 | if ($Verbose) { | |
133 | print "Not checking filelist.\n"; | |
134 | } | |
135 | if ($UseDir) { | |
136 | chdir $UseDir or die "Failed to chdir to '$UseDir':$!"; | |
137 | } | |
138 | } | |
139 | ||
140 | foreach my $lib ('To', 'lib', | |
141 | map {File::Spec->catdir("lib",$_)} | |
142 | qw(gc_sc dt bc hst ea jt lb nt ccc)) { | |
143 | next if -d $lib; | |
144 | mkdir $lib, 0755 or die "mkdir '$lib': $!"; | |
145 | } | |
146 | ||
147 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 3.1.1. | |
148 | ||
149 | my $HEADER=<<"EOF"; | |
150 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! | |
151 | # This file is built by $0 from e.g. UnicodeData.txt. | |
152 | # Any changes made here will be lost! | |
153 | ||
154 | EOF | |
155 | ||
156 | sub force_unlink { | |
157 | my $filename = shift; | |
158 | return unless -e $filename; | |
159 | return if CORE::unlink($filename); | |
160 | # We might need write permission | |
161 | chmod 0777, $filename; | |
162 | CORE::unlink($filename) or die "Couldn't unlink $filename: $!\n"; | |
163 | } | |
164 | ||
165 | ## | |
166 | ## Given a filename and a reference to an array of lines, | |
167 | ## write the lines to the file only if the contents have not changed. | |
168 | ## Filename can be given as an arrayref of directory names | |
169 | ## | |
170 | sub WriteIfChanged($\@) | |
171 | { | |
172 | my $file = shift; | |
173 | my $lines = shift; | |
174 | ||
175 | $file = File::Spec->catfile(@$file) if ref $file; | |
176 | ||
177 | my $TextToWrite = join '', @$lines; | |
178 | if (open IN, $file) { | |
179 | local($/) = undef; | |
180 | my $PreviousText = <IN>; | |
181 | close IN; | |
182 | if ($PreviousText eq $TextToWrite) { | |
183 | print "$file unchanged.\n" if $Verbose; | |
184 | return unless $AlwaysWrite; | |
185 | } | |
186 | } | |
187 | force_unlink ($file); | |
188 | if (not open OUT, ">$file") { | |
189 | die "$0: can't open $file for output: $!\n"; | |
190 | } | |
191 | print "$file written.\n" if $Verbose; | |
192 | ||
193 | print OUT $TextToWrite; | |
194 | close OUT; | |
195 | } | |
196 | ||
197 | ## | |
198 | ## The main datastructure (a "Table") represents a set of code points that | |
199 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, | |
200 | ## etc.). They are kept as ranges of code points (starting and ending of | |
201 | ## each range). | |
202 | ## | |
203 | ## For example, a range ASCII LETTERS would be represented as: | |
204 | ## [ [ 0x41 => 0x5A, 'UPPER' ], | |
205 | ## [ 0x61 => 0x7A, 'LOWER, ] ] | |
206 | ## | |
207 | sub RANGE_START() { 0 } ## index into range element | |
208 | sub RANGE_END() { 1 } ## index into range element | |
209 | sub RANGE_NAME() { 2 } ## index into range element | |
210 | ||
211 | ## Conceptually, these should really be folded into the 'Table' objects | |
212 | my %TableInfo; | |
213 | my %TableDesc; | |
214 | my %FuzzyNames; | |
215 | my %AliasInfo; | |
216 | my %CanonicalToOrig; | |
217 | ||
218 | ## | |
219 | ## Turn something like | |
220 | ## OLD-ITALIC | |
221 | ## into | |
222 | ## OldItalic | |
223 | ## | |
224 | sub CanonicalName($) | |
225 | { | |
226 | my $orig = shift; | |
227 | my $name = lc $orig; | |
228 | $name =~ s/(?<![a-z])(\w)/\u$1/g; | |
229 | $name =~ s/[-_\s]+//g; | |
230 | ||
231 | $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name}; | |
232 | return $name; | |
233 | } | |
234 | ||
235 | ||
236 | ## | |
237 | ## Store the alias definitions for later use. | |
238 | ## | |
239 | my %PropertyAlias; | |
240 | my %PropValueAlias; | |
241 | ||
242 | my %PA_reverse; | |
243 | my %PVA_reverse; | |
244 | ||
245 | sub Build_Aliases() | |
246 | { | |
247 | ## | |
248 | ## Most of the work with aliases doesn't occur here, | |
249 | ## but rather in utf8_heavy.pl, which uses PVA.pl, | |
250 | ||
251 | # Placate the warnings about used only once. (They are used again, but | |
252 | # via a typeglob lookup) | |
253 | %utf8::PropertyAlias = (); | |
254 | %utf8::PA_reverse = (); | |
255 | %utf8::PropValueAlias = (); | |
256 | %utf8::PVA_reverse = (); | |
257 | %utf8::PVA_abbr_map = (); | |
258 | ||
259 | open PA, "< PropertyAliases.txt" | |
260 | or confess "Can't open PropertyAliases.txt: $!"; | |
261 | while (<PA>) { | |
262 | s/#.*//; | |
263 | s/\s+$//; | |
264 | next if /^$/; | |
265 | ||
266 | my ($abbrev, $name) = split /\s*;\s*/; | |
267 | next if $abbrev eq "n/a"; | |
268 | $PropertyAlias{$abbrev} = $name; | |
269 | $PA_reverse{$name} = $abbrev; | |
270 | ||
271 | # The %utf8::... versions use japhy's code originally from utf8_pva.pl | |
272 | # However, it's moved here so that we build the tables at runtime. | |
273 | tr/ _-//d for $abbrev, $name; | |
274 | $utf8::PropertyAlias{lc $abbrev} = $name; | |
275 | $utf8::PA_reverse{lc $name} = $abbrev; | |
276 | } | |
277 | close PA; | |
278 | ||
279 | open PVA, "< PropValueAliases.txt" | |
280 | or confess "Can't open PropValueAliases.txt: $!"; | |
281 | while (<PVA>) { | |
282 | s/#.*//; | |
283 | s/\s+$//; | |
284 | next if /^$/; | |
285 | ||
286 | my ($prop, @data) = split /\s*;\s*/; | |
287 | ||
288 | if ($prop eq 'ccc') { | |
289 | $PropValueAlias{$prop}{$data[1]} = [ @data[0,2] ]; | |
290 | $PVA_reverse{$prop}{$data[2]} = [ @data[0,1] ]; | |
291 | } | |
292 | else { | |
293 | next if $data[0] eq "n/a"; | |
294 | $PropValueAlias{$prop}{$data[0]} = $data[1]; | |
295 | $PVA_reverse{$prop}{$data[1]} = $data[0]; | |
296 | } | |
297 | ||
298 | shift @data if $prop eq 'ccc'; | |
299 | next if $data[0] eq "n/a"; | |
300 | ||
301 | $data[1] =~ tr/ _-//d; | |
302 | $utf8::PropValueAlias{$prop}{lc $data[0]} = $data[1]; | |
303 | $utf8::PVA_reverse{$prop}{lc $data[1]} = $data[0]; | |
304 | ||
305 | my $abbr_class = ($prop eq 'gc' or $prop eq 'sc') ? 'gc_sc' : $prop; | |
306 | $utf8::PVA_abbr_map{$abbr_class}{lc $data[0]} = $data[0]; | |
307 | } | |
308 | close PVA; | |
309 | ||
310 | # backwards compatibility for L& -> LC | |
311 | $utf8::PropValueAlias{gc}{'l&'} = $utf8::PropValueAlias{gc}{lc}; | |
312 | $utf8::PVA_abbr_map{gc_sc}{'l&'} = $utf8::PVA_abbr_map{gc_sc}{lc}; | |
313 | ||
314 | } | |
315 | ||
316 | ||
317 | ## | |
318 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. | |
319 | ## | |
320 | ## Called like: | |
321 | ## New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1); | |
322 | ## | |
323 | ## Normally, these parameters are set when the Table is created (when the | |
324 | ## Table->New constructor is called), but there are times when it needs to | |
325 | ## be done after-the-fact...) | |
326 | ## | |
327 | sub New_Prop($$$@) | |
328 | { | |
329 | my $Type = shift; ## "Is" or "In"; | |
330 | my $Name = shift; | |
331 | my $Table = shift; | |
332 | ||
333 | ## remaining args are optional key/val | |
334 | my %Args = @_; | |
335 | ||
336 | my $Fuzzy = delete $Args{Fuzzy}; | |
337 | my $Desc = delete $Args{Desc}; # description | |
338 | ||
339 | $Name = CanonicalName($Name) if $Fuzzy; | |
340 | ||
341 | ## sanity check a few args | |
342 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { | |
343 | confess "$0: bad args to New_Prop" | |
344 | } | |
345 | ||
346 | if (not $TableInfo{$Type}->{$Name}) | |
347 | { | |
348 | $TableInfo{$Type}->{$Name} = $Table; | |
349 | $TableDesc{$Type}->{$Name} = $Desc; | |
350 | if ($Fuzzy) { | |
351 | $FuzzyNames{$Type}->{$Name} = $Name; | |
352 | } | |
353 | } | |
354 | } | |
355 | ||
356 | ||
357 | ## | |
358 | ## Creates a new Table object. | |
359 | ## | |
360 | ## Args are key/value pairs: | |
361 | ## In => Name -- Name of "In" property to be associated with | |
362 | ## Is => Name -- Name of "Is" property to be associated with | |
363 | ## Fuzzy => Boolean -- True if name can be accessed "fuzzily" | |
364 | ## Desc => String -- Description of the property | |
365 | ## | |
366 | ## No args are required. | |
367 | ## | |
368 | sub Table::New | |
369 | { | |
370 | my $class = shift; | |
371 | my %Args = @_; | |
372 | ||
373 | my $Table = bless [], $class; | |
374 | ||
375 | my $Fuzzy = delete $Args{Fuzzy}; | |
376 | my $Desc = delete $Args{Desc}; | |
377 | ||
378 | for my $Type ('Is', 'In') | |
379 | { | |
380 | if (my $Name = delete $Args{$Type}) { | |
381 | New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy); | |
382 | } | |
383 | } | |
384 | ||
385 | ## shouldn't have any left over | |
386 | if (%Args) { | |
387 | confess "$0: bad args to Table->New" | |
388 | } | |
389 | ||
390 | return $Table; | |
391 | } | |
392 | ||
393 | ||
394 | ## | |
395 | ## Returns the maximum code point currently in the table. | |
396 | ## | |
397 | sub Table::Max | |
398 | { | |
399 | my $last = $_[0]->[-1]; ## last code point | |
400 | confess "oops" unless $last; ## must have code points to have a max | |
401 | return $last->[RANGE_END]; | |
402 | } | |
403 | ||
404 | ## | |
405 | ## Replaces the codepoints in the Table with those in the Table given | |
406 | ## as an arg. (NOTE: this is not a "deep copy"). | |
407 | ## | |
408 | sub Table::Replace($$) | |
409 | { | |
410 | my $Table = shift; #self | |
411 | my $New = shift; | |
412 | ||
413 | @$Table = @$New; | |
414 | } | |
415 | ||
416 | ## | |
417 | ## Given a new code point, make the last range of the Table extend to | |
418 | ## include the new (and all intervening) code points. | |
419 | ## | |
420 | ## Takes the time to make sure that the extension is valid. | |
421 | ## | |
422 | sub Table::Extend | |
423 | { | |
424 | my $Table = shift; #self | |
425 | my $codepoint = shift; | |
426 | ||
427 | my $PrevMax = $Table->Max; | |
428 | ||
429 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; | |
430 | ||
431 | $Table->ExtendNoCheck($codepoint); | |
432 | } | |
433 | ||
434 | ||
435 | ## | |
436 | ## Given a new code point, make the last range of the Table extend to | |
437 | ## include the new (and all intervening) code points. | |
438 | ## | |
439 | ## Does NOT check that the extension is valid. Assumes that the caller | |
440 | ## has already made this check. | |
441 | ## | |
442 | sub Table::ExtendNoCheck | |
443 | { | |
444 | ## Optmized adding: Assumes $Table and $codepoint as parms | |
445 | $_[0]->[-1]->[RANGE_END] = $_[1]; | |
446 | } | |
447 | ||
448 | ## | |
449 | ## Given a code point range start and end (and optional name), blindly | |
450 | ## append them to the list of ranges for the Table. | |
451 | ## | |
452 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
453 | ## | |
454 | sub Table::RawAppendRange | |
455 | { | |
456 | my $Table = shift; #self | |
457 | my $start = shift; | |
458 | my $end = shift; | |
459 | my $name = shift; | |
460 | $name = "" if not defined $name; ## warning: $name can be "0" | |
461 | ||
462 | push @$Table, [ $start, # RANGE_START | |
463 | $end, # RANGE_END | |
464 | $name ]; # RANGE_NAME | |
465 | } | |
466 | ||
467 | ## | |
468 | ## Given a code point (and optional name), add it to the Table. | |
469 | ## | |
470 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
471 | ## | |
472 | sub Table::Append | |
473 | { | |
474 | my $Table = shift; #self | |
475 | my $codepoint = shift; | |
476 | my $name = shift; | |
477 | $name = "" if not defined $name; ## warning: $name can be "0" | |
478 | ||
479 | ## | |
480 | ## If we've already got a range working, and this code point is the next | |
481 | ## one in line, and if the name is the same, just extend the current range. | |
482 | ## | |
483 | my $last = $Table->[-1]; | |
484 | if ($last | |
485 | and | |
486 | $last->[RANGE_END] == $codepoint - 1 | |
487 | and | |
488 | $last->[RANGE_NAME] eq $name) | |
489 | { | |
490 | $Table->ExtendNoCheck($codepoint); | |
491 | } | |
492 | else | |
493 | { | |
494 | $Table->RawAppendRange($codepoint, $codepoint, $name); | |
495 | } | |
496 | } | |
497 | ||
498 | ## | |
499 | ## Given a code point range starting value and ending value (and name), | |
500 | ## Add the range to teh Table. | |
501 | ## | |
502 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
503 | ## | |
504 | sub Table::AppendRange | |
505 | { | |
506 | my $Table = shift; #self | |
507 | my $start = shift; | |
508 | my $end = shift; | |
509 | my $name = shift; | |
510 | $name = "" if not defined $name; ## warning: $name can be "0" | |
511 | ||
512 | $Table->Append($start, $name); | |
513 | $Table->Extend($end) if $end > $start; | |
514 | } | |
515 | ||
516 | ## | |
517 | ## Return a new Table that represents all code points not in the Table. | |
518 | ## | |
519 | sub Table::Invert | |
520 | { | |
521 | my $Table = shift; #self | |
522 | ||
523 | my $New = Table->New(); | |
524 | my $max = -1; | |
525 | for my $range (@$Table) | |
526 | { | |
527 | my $start = $range->[RANGE_START]; | |
528 | my $end = $range->[RANGE_END]; | |
529 | if ($start-1 >= $max+1) { | |
530 | $New->AppendRange($max+1, $start-1, ""); | |
531 | } | |
532 | $max = $end; | |
533 | } | |
534 | if ($max+1 < $LastUnicodeCodepoint) { | |
535 | $New->AppendRange($max+1, $LastUnicodeCodepoint); | |
536 | } | |
537 | return $New; | |
538 | } | |
539 | ||
540 | ## | |
541 | ## Merges any number of other tables with $self, returning the new table. | |
542 | ## (existing tables are not modified) | |
543 | ## | |
544 | ## | |
545 | ## Args may be Tables, or individual code points (as integers). | |
546 | ## | |
547 | ## Can be called as either a constructor or a method. | |
548 | ## | |
549 | sub Table::Merge | |
550 | { | |
551 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class | |
552 | my @Tables = @_; | |
553 | ||
554 | ## Accumulate all records from all tables | |
555 | my @Records; | |
556 | for my $Arg (@Tables) | |
557 | { | |
558 | if (ref $Arg) { | |
559 | ## arg is a table -- get its ranges | |
560 | push @Records, @$Arg; | |
561 | } else { | |
562 | ## arg is a codepoint, make a range | |
563 | push @Records, [ $Arg, $Arg ] | |
564 | } | |
565 | } | |
566 | ||
567 | ## sort by range start, with longer ranges coming first. | |
568 | my ($first, @Rest) = sort { | |
569 | ($a->[RANGE_START] <=> $b->[RANGE_START]) | |
570 | or | |
571 | ($b->[RANGE_END] <=> $b->[RANGE_END]) | |
572 | } @Records; | |
573 | ||
574 | my $New = Table->New(); | |
575 | ||
576 | ## Ensuring the first range is there makes the subsequent loop easier | |
577 | $New->AppendRange($first->[RANGE_START], | |
578 | $first->[RANGE_END]); | |
579 | ||
580 | ## Fold in records so long as they add new information. | |
581 | for my $set (@Rest) | |
582 | { | |
583 | my $start = $set->[RANGE_START]; | |
584 | my $end = $set->[RANGE_END]; | |
585 | if ($start > $New->Max) { | |
586 | $New->AppendRange($start, $end); | |
587 | } elsif ($end > $New->Max) { | |
588 | $New->ExtendNoCheck($end); | |
589 | } | |
590 | } | |
591 | ||
592 | return $New; | |
593 | } | |
594 | ||
595 | ## | |
596 | ## Given a filename, write a representation of the Table to a file. | |
597 | ## May have an optional comment as a 2nd arg. | |
598 | ## Filename may actually be an arrayref of directories | |
599 | ## | |
600 | sub Table::Write | |
601 | { | |
602 | my $Table = shift; #self | |
603 | my $filename = shift; | |
604 | my $comment = shift; | |
605 | ||
606 | my @OUT = $HEADER; | |
607 | if (defined $comment) { | |
608 | $comment =~ s/\s+\Z//; | |
609 | $comment =~ s/^/# /gm; | |
610 | push @OUT, "#\n$comment\n#\n"; | |
611 | } | |
612 | push @OUT, "return <<'END';\n"; | |
613 | ||
614 | for my $set (@$Table) | |
615 | { | |
616 | my $start = $set->[RANGE_START]; | |
617 | my $end = $set->[RANGE_END]; | |
618 | my $name = $set->[RANGE_NAME]; | |
619 | ||
620 | if ($start == $end) { | |
621 | push @OUT, sprintf "%04X\t\t%s\n", $start, $name; | |
622 | } else { | |
623 | push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name; | |
624 | } | |
625 | } | |
626 | ||
627 | push @OUT, "END\n"; | |
628 | ||
629 | WriteIfChanged($filename, @OUT); | |
630 | } | |
631 | ||
632 | ## This used only for making the test script. | |
633 | ## helper function | |
634 | sub IsUsable($) | |
635 | { | |
636 | my $code = shift; | |
637 | return 0 if $code <= 0x0000; ## don't use null | |
638 | return 0 if $code >= $LastUnicodeCodepoint; ## keep in range | |
639 | return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates | |
640 | return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good | |
641 | return 0 if (($code & 0xFFFF) == 0xFFFE); ## utf8.c says no good | |
642 | return 0 if (($code & 0xFFFF) == 0xFFFF); ## utf8.c says no good | |
643 | return 1; | |
644 | } | |
645 | ||
646 | ## Return a code point that's part of the table. | |
647 | ## Returns nothing if the table is empty (or covers only surrogates). | |
648 | ## This used only for making the test script. | |
649 | sub Table::ValidCode | |
650 | { | |
651 | my $Table = shift; #self | |
652 | for my $set (@$Table) { | |
653 | return $set->[RANGE_END] if IsUsable($set->[RANGE_END]); | |
654 | } | |
655 | return (); | |
656 | } | |
657 | ||
658 | ## Return a code point that's not part of the table | |
659 | ## Returns nothing if the table covers all code points. | |
660 | ## This used only for making the test script. | |
661 | sub Table::InvalidCode | |
662 | { | |
663 | my $Table = shift; #self | |
664 | ||
665 | return 0x1234 if not @$Table; | |
666 | ||
667 | for my $set (@$Table) | |
668 | { | |
669 | if (IsUsable($set->[RANGE_END] + 1)) | |
670 | { | |
671 | return $set->[RANGE_END] + 1; | |
672 | } | |
673 | ||
674 | if (IsUsable($set->[RANGE_START] - 1)) | |
675 | { | |
676 | return $set->[RANGE_START] - 1; | |
677 | } | |
678 | } | |
679 | return (); | |
680 | } | |
681 | ||
682 | ########################################################################### | |
683 | ########################################################################### | |
684 | ########################################################################### | |
685 | ||
686 | ||
687 | ## | |
688 | ## Called like: | |
689 | ## New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1); | |
690 | ## | |
691 | ## The args must be in that order, although the Fuzzy pair may be omitted. | |
692 | ## | |
693 | ## This creates 'IsAll' as an alias for 'IsAny' | |
694 | ## | |
695 | sub New_Alias($$$@) | |
696 | { | |
697 | my $Type = shift; ## "Is" or "In" | |
698 | my $Alias = shift; | |
699 | my $SameAs = shift; # expecting "SameAs" -- just ignored | |
700 | my $Name = shift; | |
701 | ||
702 | ## remaining args are optional key/val | |
703 | my %Args = @_; | |
704 | ||
705 | my $Fuzzy = delete $Args{Fuzzy}; | |
706 | ||
707 | ## sanity check a few args | |
708 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { | |
709 | confess "$0: bad args to New_Alias" | |
710 | } | |
711 | ||
712 | $Alias = CanonicalName($Alias) if $Fuzzy; | |
713 | ||
714 | if (not $TableInfo{$Type}->{$Name}) | |
715 | { | |
716 | my $CName = CanonicalName($Name); | |
717 | if ($TableInfo{$Type}->{$CName}) { | |
718 | confess "$0: Use canonical form '$CName' instead of '$Name' for alias."; | |
719 | } else { | |
720 | confess "$0: don't have original $Type => $Name to make alias\n"; | |
721 | } | |
722 | } | |
723 | if ($TableInfo{$Alias}) { | |
724 | confess "$0: already have original $Type => $Alias; can't make alias"; | |
725 | } | |
726 | $AliasInfo{$Type}->{$Name} = $Alias; | |
727 | if ($Fuzzy) { | |
728 | $FuzzyNames{$Type}->{$Alias} = $Name; | |
729 | } | |
730 | ||
731 | } | |
732 | ||
733 | ||
734 | ## All assigned code points | |
735 | my $Assigned = Table->New(Is => 'Assigned', | |
736 | Desc => "All assigned code points", | |
737 | Fuzzy => 0); | |
738 | ||
739 | my $Name = Table->New(); ## all characters, individually by name | |
740 | my $General = Table->New(); ## all characters, grouped by category | |
741 | my %General; | |
742 | my %Cat; | |
743 | ||
744 | ## Simple Data::Dumper alike. Good enough for our needs. We can't use the real | |
745 | ## thing as we have to run under miniperl | |
746 | sub simple_dumper { | |
747 | my @lines; | |
748 | my $item; | |
749 | foreach $item (@_) { | |
750 | if (ref $item) { | |
751 | if (ref $item eq 'ARRAY') { | |
752 | push @lines, "[\n", simple_dumper (@$item), "],\n"; | |
753 | } elsif (ref $item eq 'HASH') { | |
754 | push @lines, "{\n", simple_dumper (%$item), "},\n"; | |
755 | } else { | |
756 | die "Can't cope with $item"; | |
757 | } | |
758 | } else { | |
759 | if (defined $item) { | |
760 | my $copy = $item; | |
761 | $copy =~ s/([\'\\])/\\$1/gs; | |
762 | push @lines, "'$copy',\n"; | |
763 | } else { | |
764 | push @lines, "undef,\n"; | |
765 | } | |
766 | } | |
767 | } | |
768 | @lines; | |
769 | } | |
770 | ||
771 | ## | |
772 | ## Process UnicodeData.txt (Categories, etc.) | |
773 | ## | |
774 | sub UnicodeData_Txt() | |
775 | { | |
776 | my $Bidi = Table->New(); | |
777 | my $Deco = Table->New(); | |
778 | my $Comb = Table->New(); | |
779 | my $Number = Table->New(); | |
780 | my $Mirrored = Table->New();#Is => 'Mirrored', | |
781 | #Desc => "Mirrored in bidirectional text", | |
782 | #Fuzzy => 0); | |
783 | ||
784 | my %DC; | |
785 | my %Bidi; | |
786 | my %Number; | |
787 | $DC{can} = Table->New(); | |
788 | $DC{com} = Table->New(); | |
789 | ||
790 | ## Initialize Perl-generated categories | |
791 | ## (Categories from UnicodeData.txt are auto-initialized in gencat) | |
792 | $Cat{Alnum} = | |
793 | Table->New(Is => 'Alnum', Desc => "[[:Alnum:]]", Fuzzy => 0); | |
794 | $Cat{Alpha} = | |
795 | Table->New(Is => 'Alpha', Desc => "[[:Alpha:]]", Fuzzy => 0); | |
796 | $Cat{ASCII} = | |
797 | Table->New(Is => 'ASCII', Desc => "[[:ASCII:]]", Fuzzy => 0); | |
798 | $Cat{Blank} = | |
799 | Table->New(Is => 'Blank', Desc => "[[:Blank:]]", Fuzzy => 0); | |
800 | $Cat{Cntrl} = | |
801 | Table->New(Is => 'Cntrl', Desc => "[[:Cntrl:]]", Fuzzy => 0); | |
802 | $Cat{Digit} = | |
803 | Table->New(Is => 'Digit', Desc => "[[:Digit:]]", Fuzzy => 0); | |
804 | $Cat{Graph} = | |
805 | Table->New(Is => 'Graph', Desc => "[[:Graph:]]", Fuzzy => 0); | |
806 | $Cat{Lower} = | |
807 | Table->New(Is => 'Lower', Desc => "[[:Lower:]]", Fuzzy => 0); | |
808 | $Cat{Print} = | |
809 | Table->New(Is => 'Print', Desc => "[[:Print:]]", Fuzzy => 0); | |
810 | $Cat{Punct} = | |
811 | Table->New(Is => 'Punct', Desc => "[[:Punct:]]", Fuzzy => 0); | |
812 | $Cat{Space} = | |
813 | Table->New(Is => 'Space', Desc => "[[:Space:]]", Fuzzy => 0); | |
814 | $Cat{Title} = | |
815 | Table->New(Is => 'Title', Desc => "[[:Title:]]", Fuzzy => 0); | |
816 | $Cat{Upper} = | |
817 | Table->New(Is => 'Upper', Desc => "[[:Upper:]]", Fuzzy => 0); | |
818 | $Cat{XDigit} = | |
819 | Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0); | |
820 | $Cat{Word} = | |
821 | Table->New(Is => 'Word', Desc => "[[:Word:]]", Fuzzy => 0); | |
822 | $Cat{SpacePerl} = | |
823 | Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0); | |
824 | ||
825 | my %To; | |
826 | $To{Upper} = Table->New(); | |
827 | $To{Lower} = Table->New(); | |
828 | $To{Title} = Table->New(); | |
829 | $To{Digit} = Table->New(); | |
830 | ||
831 | sub gencat($$$$) | |
832 | { | |
833 | my ($name, ## Name ("LATIN CAPITAL LETTER A") | |
834 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) | |
835 | $code, ## Code point (as an integer) | |
836 | $op) = @_; | |
837 | ||
838 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc | |
839 | ||
840 | $Assigned->$op($code); | |
841 | $Name->$op($code, $name); | |
842 | $General->$op($code, $cat); | |
843 | ||
844 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) | |
845 | $Cat{$cat} ||= Table->New(Is => $cat, | |
846 | Desc => "General Category '$cat'", | |
847 | Fuzzy => 0); | |
848 | $Cat{$cat}->$op($code); | |
849 | ||
850 | ## add to the major category (e.g. "L", "N", "C", ...) | |
851 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, | |
852 | Desc => "Major Category '$MajorCat'", | |
853 | Fuzzy => 0); | |
854 | $Cat{$MajorCat}->$op($code); | |
855 | ||
856 | ($General{$name} ||= Table->New)->$op($code, $name); | |
857 | ||
858 | # 005F: SPACING UNDERSCORE | |
859 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]|Pc/; | |
860 | $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/; | |
861 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; | |
862 | ||
863 | my $isspace = | |
864 | ($cat =~ /Zs|Zl|Zp/ && | |
865 | $code != 0x200B) # 200B is ZWSP which is for line break control | |
866 | # and therefore it is not part of "space" even while it is "Zs". | |
867 | || $code == 0x0009 # 0009: HORIZONTAL TAB | |
868 | || $code == 0x000A # 000A: LINE FEED | |
869 | || $code == 0x000B # 000B: VERTICAL TAB | |
870 | || $code == 0x000C # 000C: FORM FEED | |
871 | || $code == 0x000D # 000D: CARRIAGE RETURN | |
872 | || $code == 0x0085 # 0085: NEL | |
873 | ||
874 | ; | |
875 | ||
876 | $Cat{Space}->$op($code) if $isspace; | |
877 | ||
878 | $Cat{SpacePerl}->$op($code) if $isspace | |
879 | && $code != 0x000B; # Backward compat. | |
880 | ||
881 | $Cat{Blank}->$op($code) if $isspace | |
882 | && !($code == 0x000A || | |
883 | $code == 0x000B || | |
884 | $code == 0x000C || | |
885 | $code == 0x000D || | |
886 | $code == 0x0085 || | |
887 | $cat =~ /^Z[lp]/); | |
888 | ||
889 | $Cat{Digit}->$op($code) if $cat eq "Nd"; | |
890 | $Cat{Upper}->$op($code) if $cat eq "Lu"; | |
891 | $Cat{Lower}->$op($code) if $cat eq "Ll"; | |
892 | $Cat{Title}->$op($code) if $cat eq "Lt"; | |
893 | $Cat{ASCII}->$op($code) if $code <= 0x007F; | |
894 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; | |
895 | my $isgraph = !$isspace && $cat !~ /Cc|Cs|Cn/; | |
896 | $Cat{Graph}->$op($code) if $isgraph; | |
897 | $Cat{Print}->$op($code) if $isgraph || $isspace; | |
898 | $Cat{Punct}->$op($code) if $cat =~ /^P/; | |
899 | ||
900 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 | |
901 | || ($code >= 0x41 && $code <= 0x46) ## A..F | |
902 | || ($code >= 0x61 && $code <= 0x66); ## a..f | |
903 | } | |
904 | ||
905 | ## open ane read file..... | |
906 | if (not open IN, "UnicodeData.txt") { | |
907 | die "$0: UnicodeData.txt: $!\n"; | |
908 | } | |
909 | ||
910 | ## | |
911 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} | |
912 | ## | |
913 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") | |
914 | ||
915 | my %CodeToDeco; ## Maps code to decomp. list for chars with first | |
916 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) | |
917 | ||
918 | ## This is filled in as we go.... | |
919 | my $CombAbove = Table->New(Is => '_CombAbove', | |
920 | Desc => '(for internal casefolding use)', | |
921 | Fuzzy => 0); | |
922 | ||
923 | while (<IN>) | |
924 | { | |
925 | next unless /^[0-9A-Fa-f]+;/; | |
926 | s/\s+$//; | |
927 | ||
928 | my ($hexcode, ## code point in hex (e.g. "0041") | |
929 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") | |
930 | $cat, ## category (e.g. "Lu") | |
931 | $comb, ## Canonical combining class (e.t. "230") | |
932 | $bidi, ## directional category (e.g. "L") | |
933 | $deco, ## decomposition mapping | |
934 | $decimal, ## decimal digit value | |
935 | $digit, ## digit value | |
936 | $number, ## numeric value | |
937 | $mirrored, ## mirrored | |
938 | $unicode10, ## name in Unicode 1.0 | |
939 | $comment, ## comment field | |
940 | $upper, ## uppercase mapping | |
941 | $lower, ## lowercase mapping | |
942 | $title, ## titlecase mapping | |
943 | ) = split(/\s*;\s*/); | |
944 | ||
945 | # Note that in Unicode 3.2 there will be names like | |
946 | # LINE FEED (LF), which probably means that \N{} needs | |
947 | # to cope also with LINE FEED and LF. | |
948 | $name = $unicode10 if $name eq '<control>' && $unicode10 ne ''; | |
949 | ||
950 | my $code = hex($hexcode); | |
951 | ||
952 | if ($comb and $comb == 230) { | |
953 | $CombAbove->Append($code); | |
954 | $_Above_HexCodes{$hexcode} = 1; | |
955 | } | |
956 | ||
957 | ## Used in building \p{_CanonDCIJ} | |
958 | if ($deco and $deco =~ m/^006[9A]\b/) { | |
959 | $CodeToDeco{$code} = $deco; | |
960 | } | |
961 | ||
962 | ## | |
963 | ## There are a few pairs of lines like: | |
964 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; | |
965 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; | |
966 | ## that define ranges. | |
967 | ## | |
968 | if ($name =~ /^<(.+), (First|Last)>$/) | |
969 | { | |
970 | $name = $1; | |
971 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); | |
972 | #New_Prop(In => $name, $General{$name}, Fuzzy => 1); | |
973 | } | |
974 | else | |
975 | { | |
976 | ## normal (single-character) lines | |
977 | gencat($name, $cat, $code, 'Append'); | |
978 | ||
979 | # No Append() here since since several codes may map into one. | |
980 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; | |
981 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; | |
982 | $To{Title}->RawAppendRange($code, $code, $title) if $title; | |
983 | $To{Digit}->Append($code, $decimal) if length $decimal; | |
984 | ||
985 | $Bidi->Append($code, $bidi); | |
986 | $Comb->Append($code, $comb) if $comb; | |
987 | $Number->Append($code, $number) if length $number; | |
988 | ||
989 | length($decimal) and ($Number{De} ||= Table->New())->Append($code) | |
990 | or | |
991 | length($digit) and ($Number{Di} ||= Table->New())->Append($code) | |
992 | or | |
993 | length($number) and ($Number{Nu} ||= Table->New())->Append($code); | |
994 | ||
995 | $Mirrored->Append($code) if $mirrored eq "Y"; | |
996 | ||
997 | $Bidi{$bidi} ||= Table->New();#Is => "bt/$bidi", | |
998 | #Desc => "Bi-directional category '$bidi'", | |
999 | #Fuzzy => 0); | |
1000 | $Bidi{$bidi}->Append($code); | |
1001 | ||
1002 | if ($deco) | |
1003 | { | |
1004 | $Deco->Append($code, $deco); | |
1005 | if ($deco =~/^<(\w+)>/) | |
1006 | { | |
1007 | my $dshort = $PVA_reverse{dt}{ucfirst lc $1}; | |
1008 | $DC{com}->Append($code); | |
1009 | ||
1010 | $DC{$dshort} ||= Table->New(); | |
1011 | $DC{$dshort}->Append($code); | |
1012 | } | |
1013 | else | |
1014 | { | |
1015 | $DC{can}->Append($code); | |
1016 | } | |
1017 | } | |
1018 | } | |
1019 | } | |
1020 | close IN; | |
1021 | ||
1022 | ## | |
1023 | ## Tidy up a few special cases.... | |
1024 | ## | |
1025 | ||
1026 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist | |
1027 | New_Prop(Is => 'Cn', | |
1028 | $Cat{Cn}, | |
1029 | Desc => "General Category 'Cn' [not functional in Perl]", | |
1030 | Fuzzy => 0); | |
1031 | ||
1032 | ## Unassigned is the same as 'Cn' | |
1033 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0); | |
1034 | ||
1035 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C | |
1036 | ||
1037 | ||
1038 | # LC is Ll, Lu, and Lt. | |
1039 | # (used to be L& or L_, but PropValueAliases.txt defines it as LC) | |
1040 | New_Prop(Is => 'LC', | |
1041 | Table->Merge(@Cat{qw[Ll Lu Lt]}), | |
1042 | Desc => '[\p{Ll}\p{Lu}\p{Lt}]', | |
1043 | Fuzzy => 0); | |
1044 | ||
1045 | ## Any and All are all code points. | |
1046 | my $Any = Table->New(Is => 'Any', | |
1047 | Desc => sprintf("[\\x{0000}-\\x{%X}]", | |
1048 | $LastUnicodeCodepoint), | |
1049 | Fuzzy => 0); | |
1050 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); | |
1051 | ||
1052 | New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0); | |
1053 | ||
1054 | ## | |
1055 | ## Build special properties for Perl's internal case-folding needs: | |
1056 | ## \p{_CaseIgnorable} | |
1057 | ## \p{_CanonDCIJ} | |
1058 | ## \p{_CombAbove} | |
1059 | ## _CombAbove was built above. Others are built here.... | |
1060 | ## | |
1061 | ||
1062 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] | |
1063 | New_Prop(Is => '_CaseIgnorable', | |
1064 | Table->Merge($Cat{Mn}, | |
1065 | 0x00AD, #SOFT HYPHEN | |
1066 | 0x2010), #HYPHEN | |
1067 | Desc => '(for internal casefolding use)', | |
1068 | Fuzzy => 0); | |
1069 | ||
1070 | ||
1071 | ## \p{_CanonDCIJ} is fairly complex... | |
1072 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', | |
1073 | Desc => '(for internal casefolding use)', | |
1074 | Fuzzy => 0); | |
1075 | ## It contains the ASCII 'i' and 'j'.... | |
1076 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") | |
1077 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") | |
1078 | ## ...and any character with a decomposition that starts with either of | |
1079 | ## those code points, but only if the decomposition does not have any | |
1080 | ## combining character with the "ABOVE" canonical combining class. | |
1081 | for my $code (sort { $a <=> $b} keys %CodeToDeco) | |
1082 | { | |
1083 | ## Need to ensure that all decomposition characters do not have | |
1084 | ## a %HexCodeToComb in %AboveCombClasses. | |
1085 | my $want = 1; | |
1086 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) | |
1087 | { | |
1088 | if (exists $_Above_HexCodes{$deco_hexcode}) { | |
1089 | ## one of the decmposition chars has an ABOVE combination | |
1090 | ## class, so we're not interested in this one | |
1091 | $want = 0; | |
1092 | last; | |
1093 | } | |
1094 | } | |
1095 | if ($want) { | |
1096 | $CanonCDIJ->Append($code); | |
1097 | } | |
1098 | } | |
1099 | ||
1100 | ||
1101 | ||
1102 | ## | |
1103 | ## Now dump the files. | |
1104 | ## | |
1105 | $Name->Write("Name.pl"); | |
1106 | ||
1107 | { | |
1108 | my @PVA = $HEADER; | |
1109 | foreach my $name (qw (PropertyAlias PA_reverse PropValueAlias | |
1110 | PVA_reverse PVA_abbr_map)) { | |
1111 | # Should I really jump through typeglob hoops just to avoid a | |
1112 | # symbolic reference? (%{"utf8::$name}) | |
1113 | push @PVA, "\n", "\%utf8::$name = (\n", | |
1114 | simple_dumper (%{$utf8::{$name}}), ");\n"; | |
1115 | } | |
1116 | push @PVA, "1;\n"; | |
1117 | WriteIfChanged("PVA.pl", @PVA); | |
1118 | } | |
1119 | ||
1120 | # $Bidi->Write("Bidirectional.pl"); | |
1121 | for (keys %Bidi) { | |
1122 | $Bidi{$_}->Write( | |
1123 | ["lib","bc","$_.pl"], | |
1124 | "BidiClass category '$PropValueAlias{bc}{$_}'" | |
1125 | ); | |
1126 | } | |
1127 | ||
1128 | $Comb->Write("CombiningClass.pl"); | |
1129 | for (keys %{ $PropValueAlias{ccc} }) { | |
1130 | my ($code, $name) = @{ $PropValueAlias{ccc}{$_} }; | |
1131 | (my $c = Table->New())->Append($code); | |
1132 | $c->Write( | |
1133 | ["lib","ccc","$_.pl"], | |
1134 | "CombiningClass category '$name'" | |
1135 | ); | |
1136 | } | |
1137 | ||
1138 | $Deco->Write("Decomposition.pl"); | |
1139 | for (keys %DC) { | |
1140 | $DC{$_}->Write( | |
1141 | ["lib","dt","$_.pl"], | |
1142 | "DecompositionType category '$PropValueAlias{dt}{$_}'" | |
1143 | ); | |
1144 | } | |
1145 | ||
1146 | # $Number->Write("Number.pl"); | |
1147 | for (keys %Number) { | |
1148 | $Number{$_}->Write( | |
1149 | ["lib","nt","$_.pl"], | |
1150 | "NumericType category '$PropValueAlias{nt}{$_}'" | |
1151 | ); | |
1152 | } | |
1153 | ||
1154 | # $General->Write("Category.pl"); | |
1155 | ||
1156 | for my $to (sort keys %To) { | |
1157 | $To{$to}->Write(["To","$to.pl"]); | |
1158 | } | |
1159 | ||
1160 | for (keys %{ $PropValueAlias{gc} }) { | |
1161 | New_Alias(Is => $PropValueAlias{gc}{$_}, SameAs => $_, Fuzzy => 1); | |
1162 | } | |
1163 | } | |
1164 | ||
1165 | ## | |
1166 | ## Process LineBreak.txt | |
1167 | ## | |
1168 | sub LineBreak_Txt() | |
1169 | { | |
1170 | if (not open IN, "LineBreak.txt") { | |
1171 | die "$0: LineBreak.txt: $!\n"; | |
1172 | } | |
1173 | ||
1174 | my $Lbrk = Table->New(); | |
1175 | my %Lbrk; | |
1176 | ||
1177 | while (<IN>) | |
1178 | { | |
1179 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; | |
1180 | ||
1181 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); | |
1182 | ||
1183 | $Lbrk->Append($first, $lbrk); | |
1184 | ||
1185 | $Lbrk{$lbrk} ||= Table->New(); | |
1186 | $Lbrk{$lbrk}->Append($first); | |
1187 | ||
1188 | if ($last) { | |
1189 | $Lbrk->Extend($last); | |
1190 | $Lbrk{$lbrk}->Extend($last); | |
1191 | } | |
1192 | } | |
1193 | close IN; | |
1194 | ||
1195 | # $Lbrk->Write("Lbrk.pl"); | |
1196 | ||
1197 | ||
1198 | for (keys %Lbrk) { | |
1199 | $Lbrk{$_}->Write( | |
1200 | ["lib","lb","$_.pl"], | |
1201 | "Linebreak category '$PropValueAlias{lb}{$_}'" | |
1202 | ); | |
1203 | } | |
1204 | } | |
1205 | ||
1206 | ## | |
1207 | ## Process ArabicShaping.txt. | |
1208 | ## | |
1209 | sub ArabicShaping_txt() | |
1210 | { | |
1211 | if (not open IN, "ArabicShaping.txt") { | |
1212 | die "$0: ArabicShaping.txt: $!\n"; | |
1213 | } | |
1214 | ||
1215 | my $ArabLink = Table->New(); | |
1216 | my $ArabLinkGroup = Table->New(); | |
1217 | ||
1218 | my %JoinType; | |
1219 | ||
1220 | while (<IN>) | |
1221 | { | |
1222 | next unless /^[0-9A-Fa-f]+;/; | |
1223 | s/\s+$//; | |
1224 | ||
1225 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); | |
1226 | my $code = hex($hexcode); | |
1227 | $ArabLink->Append($code, $link); | |
1228 | $ArabLinkGroup->Append($code, $linkgroup); | |
1229 | ||
1230 | $JoinType{$link} ||= Table->New(Is => "JoinType$link"); | |
1231 | $JoinType{$link}->Append($code); | |
1232 | } | |
1233 | close IN; | |
1234 | ||
1235 | # $ArabLink->Write("ArabLink.pl"); | |
1236 | # $ArabLinkGroup->Write("ArabLnkGrp.pl"); | |
1237 | ||
1238 | ||
1239 | for (keys %JoinType) { | |
1240 | $JoinType{$_}->Write( | |
1241 | ["lib","jt","$_.pl"], | |
1242 | "JoiningType category '$PropValueAlias{jt}{$_}'" | |
1243 | ); | |
1244 | } | |
1245 | } | |
1246 | ||
1247 | ## | |
1248 | ## Process EastAsianWidth.txt. | |
1249 | ## | |
1250 | sub EastAsianWidth_txt() | |
1251 | { | |
1252 | if (not open IN, "EastAsianWidth.txt") { | |
1253 | die "$0: EastAsianWidth.txt: $!\n"; | |
1254 | } | |
1255 | ||
1256 | my %EAW; | |
1257 | ||
1258 | while (<IN>) | |
1259 | { | |
1260 | next unless /^[0-9A-Fa-f]+(\.\.[0-9A-Fa-f]+)?;/; | |
1261 | s/#.*//; | |
1262 | s/\s+$//; | |
1263 | ||
1264 | my ($hexcodes, $pv) = split(/\s*;\s*/); | |
1265 | $EAW{$pv} ||= Table->New(Is => "EastAsianWidth$pv"); | |
1266 | my ($start, $end) = split(/\.\./, $hexcodes); | |
1267 | if (defined $end) { | |
1268 | $EAW{$pv}->AppendRange(hex($start), hex($end)); | |
1269 | } else { | |
1270 | $EAW{$pv}->Append(hex($start)); | |
1271 | } | |
1272 | } | |
1273 | close IN; | |
1274 | ||
1275 | ||
1276 | for (keys %EAW) { | |
1277 | $EAW{$_}->Write( | |
1278 | ["lib","ea","$_.pl"], | |
1279 | "EastAsianWidth category '$PropValueAlias{ea}{$_}'" | |
1280 | ); | |
1281 | } | |
1282 | } | |
1283 | ||
1284 | ## | |
1285 | ## Process HangulSyllableType.txt. | |
1286 | ## | |
1287 | sub HangulSyllableType_txt() | |
1288 | { | |
1289 | if (not open IN, "HangulSyllableType.txt") { | |
1290 | die "$0: HangulSyllableType.txt: $!\n"; | |
1291 | } | |
1292 | ||
1293 | my %HST; | |
1294 | ||
1295 | while (<IN>) | |
1296 | { | |
1297 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; | |
1298 | my ($first, $last, $pv) = (hex($1), hex($2||""), $3); | |
1299 | ||
1300 | $HST{$pv} ||= Table->New(Is => "HangulSyllableType$pv"); | |
1301 | $HST{$pv}->Append($first); | |
1302 | ||
1303 | if ($last) { $HST{$pv}->Extend($last) } | |
1304 | } | |
1305 | close IN; | |
1306 | ||
1307 | for (keys %HST) { | |
1308 | $HST{$_}->Write( | |
1309 | ["lib","hst","$_.pl"], | |
1310 | "HangulSyllableType category '$PropValueAlias{hst}{$_}'" | |
1311 | ); | |
1312 | } | |
1313 | } | |
1314 | ||
1315 | ## | |
1316 | ## Process Jamo.txt. | |
1317 | ## | |
1318 | sub Jamo_txt() | |
1319 | { | |
1320 | if (not open IN, "Jamo.txt") { | |
1321 | die "$0: Jamo.txt: $!\n"; | |
1322 | } | |
1323 | my $Short = Table->New(); | |
1324 | ||
1325 | while (<IN>) | |
1326 | { | |
1327 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; | |
1328 | my ($code, $short) = (hex($1), $2); | |
1329 | ||
1330 | $Short->Append($code, $short); | |
1331 | } | |
1332 | close IN; | |
1333 | # $Short->Write("JamoShort.pl"); | |
1334 | } | |
1335 | ||
1336 | ## | |
1337 | ## Process Scripts.txt. | |
1338 | ## | |
1339 | sub Scripts_txt() | |
1340 | { | |
1341 | my @ScriptInfo; | |
1342 | ||
1343 | if (not open(IN, "Scripts.txt")) { | |
1344 | die "$0: Scripts.txt: $!\n"; | |
1345 | } | |
1346 | while (<IN>) { | |
1347 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
1348 | ||
1349 | # Wait until all the scripts have been read since | |
1350 | # they are not listed in numeric order. | |
1351 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; | |
1352 | } | |
1353 | close IN; | |
1354 | ||
1355 | # Now append the scripts properties in their code point order. | |
1356 | ||
1357 | my %Script; | |
1358 | my $Scripts = Table->New(); | |
1359 | ||
1360 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) | |
1361 | { | |
1362 | my ($first, $last, $name) = @$script; | |
1363 | $Scripts->Append($first, $name); | |
1364 | ||
1365 | $Script{$name} ||= Table->New(Is => $name, | |
1366 | Desc => "Script '$name'", | |
1367 | Fuzzy => 1); | |
1368 | $Script{$name}->Append($first, $name); | |
1369 | ||
1370 | if ($last) { | |
1371 | $Scripts->Extend($last); | |
1372 | $Script{$name}->Extend($last); | |
1373 | } | |
1374 | } | |
1375 | ||
1376 | # $Scripts->Write("Scripts.pl"); | |
1377 | ||
1378 | ## Common is everything not explicitly assigned to a Script | |
1379 | ## | |
1380 | ## ***shouldn't this be intersected with \p{Assigned}? ****** | |
1381 | ## | |
1382 | New_Prop(Is => 'Common', | |
1383 | $Scripts->Invert, | |
1384 | Desc => 'Pseudo-Script of codepoints not in other Unicode scripts', | |
1385 | Fuzzy => 1); | |
1386 | } | |
1387 | ||
1388 | ## | |
1389 | ## Given a name like "Close Punctuation", return a regex (that when applied | |
1390 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", | |
1391 | ## "Close-Punctuation", etc.) | |
1392 | ## | |
1393 | ## Accept any space, dash, or underbar where in the official name there is | |
1394 | ## space or a dash (or underbar, but there never is). | |
1395 | ## | |
1396 | ## | |
1397 | sub NameToRegex($) | |
1398 | { | |
1399 | my $Name = shift; | |
1400 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; | |
1401 | return $Name; | |
1402 | } | |
1403 | ||
1404 | ## | |
1405 | ## Process Blocks.txt. | |
1406 | ## | |
1407 | sub Blocks_txt() | |
1408 | { | |
1409 | my $Blocks = Table->New(); | |
1410 | my %Blocks; | |
1411 | ||
1412 | if (not open IN, "Blocks.txt") { | |
1413 | die "$0: Blocks.txt: $!\n"; | |
1414 | } | |
1415 | ||
1416 | while (<IN>) | |
1417 | { | |
1418 | #next if not /Private Use$/; | |
1419 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; | |
1420 | ||
1421 | my ($first, $last, $name) = (hex($1), hex($2), $3); | |
1422 | ||
1423 | $Blocks->Append($first, $name); | |
1424 | ||
1425 | $Blocks{$name} ||= Table->New(In => $name, | |
1426 | Desc => "Block '$name'", | |
1427 | Fuzzy => 1); | |
1428 | $Blocks{$name}->Append($first, $name); | |
1429 | ||
1430 | if ($last and $last != $first) { | |
1431 | $Blocks->Extend($last); | |
1432 | $Blocks{$name}->Extend($last); | |
1433 | } | |
1434 | } | |
1435 | close IN; | |
1436 | ||
1437 | # $Blocks->Write("Blocks.pl"); | |
1438 | } | |
1439 | ||
1440 | ## | |
1441 | ## Read in the PropList.txt. It contains extended properties not | |
1442 | ## listed in the UnicodeData.txt, such as 'Other_Alphabetic': | |
1443 | ## alphabetic but not of the general category L; many modifiers | |
1444 | ## belong to this extended property category: while they are not | |
1445 | ## alphabets, they are alphabetic in nature. | |
1446 | ## | |
1447 | sub PropList_txt() | |
1448 | { | |
1449 | my @PropInfo; | |
1450 | ||
1451 | if (not open IN, "PropList.txt") { | |
1452 | die "$0: PropList.txt: $!\n"; | |
1453 | } | |
1454 | ||
1455 | while (<IN>) | |
1456 | { | |
1457 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
1458 | ||
1459 | # Wait until all the extended properties have been read since | |
1460 | # they are not listed in numeric order. | |
1461 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; | |
1462 | } | |
1463 | close IN; | |
1464 | ||
1465 | # Now append the extended properties in their code point order. | |
1466 | my $Props = Table->New(); | |
1467 | my %Prop; | |
1468 | ||
1469 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) | |
1470 | { | |
1471 | my ($first, $last, $name) = @$prop; | |
1472 | $Props->Append($first, $name); | |
1473 | ||
1474 | $Prop{$name} ||= Table->New(Is => $name, | |
1475 | Desc => "Extended property '$name'", | |
1476 | Fuzzy => 1); | |
1477 | $Prop{$name}->Append($first, $name); | |
1478 | ||
1479 | if ($last) { | |
1480 | $Props->Extend($last); | |
1481 | $Prop{$name}->Extend($last); | |
1482 | } | |
1483 | } | |
1484 | ||
1485 | for (keys %Prop) { | |
1486 | (my $file = $PA_reverse{$_}) =~ tr/_//d; | |
1487 | # XXX I'm assuming that the names from %Prop don't suffer 8.3 clashes. | |
1488 | $BaseNames{lc $file}++; | |
1489 | $Prop{$_}->Write( | |
1490 | ["lib","gc_sc","$file.pl"], | |
1491 | "Binary property '$_'" | |
1492 | ); | |
1493 | } | |
1494 | ||
1495 | # Alphabetic is L and Other_Alphabetic. | |
1496 | New_Prop(Is => 'Alphabetic', | |
1497 | Table->Merge($Cat{L}, $Prop{Other_Alphabetic}), | |
1498 | Desc => '[\p{L}\p{OtherAlphabetic}]', # use canonical names here | |
1499 | Fuzzy => 1); | |
1500 | ||
1501 | # Lowercase is Ll and Other_Lowercase. | |
1502 | New_Prop(Is => 'Lowercase', | |
1503 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), | |
1504 | Desc => '[\p{Ll}\p{OtherLowercase}]', # use canonical names here | |
1505 | Fuzzy => 1); | |
1506 | ||
1507 | # Uppercase is Lu and Other_Uppercase. | |
1508 | New_Prop(Is => 'Uppercase', | |
1509 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), | |
1510 | Desc => '[\p{Lu}\p{Other_Uppercase}]', # use canonical names here | |
1511 | Fuzzy => 1); | |
1512 | ||
1513 | # Math is Sm and Other_Math. | |
1514 | New_Prop(Is => 'Math', | |
1515 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), | |
1516 | Desc => '[\p{Sm}\p{OtherMath}]', # use canonical names here | |
1517 | Fuzzy => 1); | |
1518 | ||
1519 | # ID_Start is Ll, Lu, Lt, Lm, Lo, and Nl. | |
1520 | New_Prop(Is => 'ID_Start', | |
1521 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}), | |
1522 | Desc => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}]', | |
1523 | Fuzzy => 1); | |
1524 | ||
1525 | # ID_Continue is ID_Start, Mn, Mc, Nd, and Pc. | |
1526 | New_Prop(Is => 'ID_Continue', | |
1527 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}), | |
1528 | Desc => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}]', | |
1529 | Fuzzy => 1); | |
1530 | } | |
1531 | ||
1532 | ||
1533 | ## | |
1534 | ## These are used in: | |
1535 | ## MakePropTestScript() | |
1536 | ## WriteAllMappings() | |
1537 | ## for making the test script. | |
1538 | ## | |
1539 | my %FuzzyNameToTest; | |
1540 | my %ExactNameToTest; | |
1541 | ||
1542 | ||
1543 | ## This used only for making the test script | |
1544 | sub GenTests($$$$) | |
1545 | { | |
1546 | my $FH = shift; | |
1547 | my $Prop = shift; | |
1548 | my $MatchCode = shift; | |
1549 | my $FailCode = shift; | |
1550 | ||
1551 | if (defined $MatchCode) { | |
1552 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode; | |
1553 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode; | |
1554 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode; | |
1555 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode; | |
1556 | } | |
1557 | if (defined $FailCode) { | |
1558 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode; | |
1559 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode; | |
1560 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode; | |
1561 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode; | |
1562 | } | |
1563 | } | |
1564 | ||
1565 | ## This used only for making the test script | |
1566 | sub ExpectError($$) | |
1567 | { | |
1568 | my $FH = shift; | |
1569 | my $prop = shift; | |
1570 | ||
1571 | print $FH qq/Error('\\p{$prop}');\n/; | |
1572 | print $FH qq/Error('\\P{$prop}');\n/; | |
1573 | } | |
1574 | ||
1575 | ## This used only for making the test script | |
1576 | my @GoodSeps = ( | |
1577 | " ", | |
1578 | "-", | |
1579 | " \t ", | |
1580 | "", | |
1581 | "", | |
1582 | "_", | |
1583 | ); | |
1584 | my @BadSeps = ( | |
1585 | "--", | |
1586 | "__", | |
1587 | " _", | |
1588 | "/" | |
1589 | ); | |
1590 | ||
1591 | ## This used only for making the test script | |
1592 | sub RandomlyFuzzifyName($;$) | |
1593 | { | |
1594 | my $Name = shift; | |
1595 | my $WantError = shift; ## if true, make an error | |
1596 | ||
1597 | my @parts; | |
1598 | for my $part (split /[-\s_]+/, $Name) | |
1599 | { | |
1600 | if (@parts) { | |
1601 | if ($WantError and rand() < 0.3) { | |
1602 | push @parts, $BadSeps[rand(@BadSeps)]; | |
1603 | $WantError = 0; | |
1604 | } else { | |
1605 | push @parts, $GoodSeps[rand(@GoodSeps)]; | |
1606 | } | |
1607 | } | |
1608 | my $switch = int rand(4); | |
1609 | if ($switch == 0) { | |
1610 | push @parts, uc $part; | |
1611 | } elsif ($switch == 1) { | |
1612 | push @parts, lc $part; | |
1613 | } elsif ($switch == 2) { | |
1614 | push @parts, ucfirst $part; | |
1615 | } else { | |
1616 | push @parts, $part; | |
1617 | } | |
1618 | } | |
1619 | my $new = join('', @parts); | |
1620 | ||
1621 | if ($WantError) { | |
1622 | if (rand() >= 0.5) { | |
1623 | $new .= $BadSeps[rand(@BadSeps)]; | |
1624 | } else { | |
1625 | $new = $BadSeps[rand(@BadSeps)] . $new; | |
1626 | } | |
1627 | } | |
1628 | return $new; | |
1629 | } | |
1630 | ||
1631 | ## This used only for making the test script | |
1632 | sub MakePropTestScript() | |
1633 | { | |
1634 | ## this written directly -- it's huge. | |
1635 | force_unlink ("TestProp.pl"); | |
1636 | if (not open OUT, ">TestProp.pl") { | |
1637 | die "$0: TestProp.pl: $!\n"; | |
1638 | } | |
1639 | print OUT <DATA>; | |
1640 | ||
1641 | while (my ($Name, $Table) = each %ExactNameToTest) | |
1642 | { | |
1643 | GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode); | |
1644 | ExpectError(*OUT, uc $Name) if uc $Name ne $Name; | |
1645 | ExpectError(*OUT, lc $Name) if lc $Name ne $Name; | |
1646 | } | |
1647 | ||
1648 | ||
1649 | while (my ($Name, $Table) = each %FuzzyNameToTest) | |
1650 | { | |
1651 | my $Orig = $CanonicalToOrig{$Name}; | |
1652 | my %Names = ( | |
1653 | $Name => 1, | |
1654 | $Orig => 1, | |
1655 | RandomlyFuzzifyName($Orig) => 1 | |
1656 | ); | |
1657 | ||
1658 | for my $N (keys %Names) { | |
1659 | GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode); | |
1660 | } | |
1661 | ||
1662 | ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR')); | |
1663 | } | |
1664 | ||
1665 | print OUT "Finished();\n"; | |
1666 | close OUT; | |
1667 | } | |
1668 | ||
1669 | ||
1670 | ## | |
1671 | ## These are used only in: | |
1672 | ## RegisterFileForName() | |
1673 | ## WriteAllMappings() | |
1674 | ## | |
1675 | my %Exact; ## will become %utf8::Exact; | |
1676 | my %Canonical; ## will become %utf8::Canonical; | |
1677 | my %CaComment; ## Comment for %Canonical entry of same key | |
1678 | ||
1679 | ## | |
1680 | ## Given info about a name and a datafile that it should be associated with, | |
1681 | ## register that assocation in %Exact and %Canonical. | |
1682 | sub RegisterFileForName($$$$) | |
1683 | { | |
1684 | my $Type = shift; | |
1685 | my $Name = shift; | |
1686 | my $IsFuzzy = shift; | |
1687 | my $filename = shift; | |
1688 | ||
1689 | ## | |
1690 | ## Now in details for the mapping. $Type eq 'Is' has the | |
1691 | ## Is removed, as it will be removed in utf8_heavy when this | |
1692 | ## data is being checked. In keeps its "In", but a second | |
1693 | ## sans-In record is written if it doesn't conflict with | |
1694 | ## anything already there. | |
1695 | ## | |
1696 | if (not $IsFuzzy) | |
1697 | { | |
1698 | if ($Type eq 'Is') { | |
1699 | die "oops[$Name]" if $Exact{$Name}; | |
1700 | $Exact{$Name} = $filename; | |
1701 | } else { | |
1702 | die "oops[$Type$Name]" if $Exact{"$Type$Name"}; | |
1703 | $Exact{"$Type$Name"} = $filename; | |
1704 | $Exact{$Name} = $filename if not $Exact{$Name}; | |
1705 | } | |
1706 | } | |
1707 | else | |
1708 | { | |
1709 | my $CName = lc $Name; | |
1710 | if ($Type eq 'Is') { | |
1711 | die "oops[$CName]" if $Canonical{$CName}; | |
1712 | $Canonical{$CName} = $filename; | |
1713 | $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2; | |
1714 | } else { | |
1715 | die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"}; | |
1716 | $Canonical{lc "$Type$CName"} = $filename; | |
1717 | $CaComment{lc "$Type$CName"} = "$Type$Name"; | |
1718 | if (not $Canonical{$CName}) { | |
1719 | $Canonical{$CName} = $filename; | |
1720 | $CaComment{$CName} = "$Type$Name"; | |
1721 | } | |
1722 | } | |
1723 | } | |
1724 | } | |
1725 | ||
1726 | ## | |
1727 | ## Writes the info accumulated in | |
1728 | ## | |
1729 | ## %TableInfo; | |
1730 | ## %FuzzyNames; | |
1731 | ## %AliasInfo; | |
1732 | ## | |
1733 | ## | |
1734 | sub WriteAllMappings() | |
1735 | { | |
1736 | my @MAP; | |
1737 | ||
1738 | ## 'Is' *MUST* come first, so its names have precidence over 'In's | |
1739 | for my $Type ('Is', 'In') | |
1740 | { | |
1741 | my %RawNameToFile; ## a per-$Type cache | |
1742 | ||
1743 | for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}}) | |
1744 | { | |
1745 | ## Note: $Name is already canonical | |
1746 | my $Table = $TableInfo{$Type}->{$Name}; | |
1747 | my $IsFuzzy = $FuzzyNames{$Type}->{$Name}; | |
1748 | ||
1749 | ## Need an 8.3 safe filename (which means "an 8 safe" $filename) | |
1750 | my $filename; | |
1751 | { | |
1752 | ## 'Is' items lose 'Is' from the basename. | |
1753 | $filename = $Type eq 'Is' ? | |
1754 | ($PVA_reverse{sc}{$Name} || $Name) : | |
1755 | "$Type$Name"; | |
1756 | ||
1757 | $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_" | |
1758 | substr($filename, 8) = '' if length($filename) > 8; | |
1759 | ||
1760 | ## | |
1761 | ## Make sure the basename doesn't conflict with something we | |
1762 | ## might have already written. If we have, say, | |
1763 | ## InGreekExtended1 | |
1764 | ## InGreekExtended2 | |
1765 | ## they become | |
1766 | ## InGreekE | |
1767 | ## InGreek2 | |
1768 | ## | |
1769 | while (my $num = $BaseNames{lc $filename}++) | |
1770 | { | |
1771 | $num++; ## so basenames with numbers start with '2', which | |
1772 | ## just looks more natural. | |
1773 | ## Want to append $num, but if it'll make the basename longer | |
1774 | ## than 8 characters, pre-truncate $filename so that the result | |
1775 | ## is acceptable. | |
1776 | my $delta = length($filename) + length($num) - 8; | |
1777 | if ($delta > 0) { | |
1778 | substr($filename, -$delta) = $num; | |
1779 | } else { | |
1780 | $filename .= $num; | |
1781 | } | |
1782 | } | |
1783 | }; | |
1784 | ||
1785 | ## | |
1786 | ## Construct a nice comment to add to the file, and build data | |
1787 | ## for the "./Properties" file along the way. | |
1788 | ## | |
1789 | my $Comment; | |
1790 | { | |
1791 | my $Desc = $TableDesc{$Type}->{$Name} || ""; | |
1792 | ## get list of names this table is reference by | |
1793 | my @Supported = $Name; | |
1794 | while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} }) | |
1795 | { | |
1796 | if ($Orig eq $Name) { | |
1797 | push @Supported, $Alias; | |
1798 | } | |
1799 | } | |
1800 | ||
1801 | my $TypeToShow = $Type eq 'Is' ? "" : $Type; | |
1802 | my $OrigProp; | |
1803 | ||
1804 | $Comment = "This file supports:\n"; | |
1805 | for my $N (@Supported) | |
1806 | { | |
1807 | my $IsFuzzy = $FuzzyNames{$Type}->{$N}; | |
1808 | my $Prop = "\\p{$TypeToShow$Name}"; | |
1809 | $OrigProp = $Prop if not $OrigProp; #cache for aliases | |
1810 | if ($IsFuzzy) { | |
1811 | $Comment .= "\t$Prop (and fuzzy permutations)\n"; | |
1812 | } else { | |
1813 | $Comment .= "\t$Prop\n"; | |
1814 | } | |
1815 | my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)"; | |
1816 | ||
1817 | push @MAP, sprintf("%s %-42s %s\n", | |
1818 | $IsFuzzy ? '*' : ' ', $Prop, $MyDesc); | |
1819 | } | |
1820 | if ($Desc) { | |
1821 | $Comment .= "\nMeaning: $Desc\n"; | |
1822 | } | |
1823 | ||
1824 | } | |
1825 | ## | |
1826 | ## Okay, write the file... | |
1827 | ## | |
1828 | $Table->Write(["lib","gc_sc","$filename.pl"], $Comment); | |
1829 | ||
1830 | ## and register it | |
1831 | $RawNameToFile{$Name} = $filename; | |
1832 | RegisterFileForName($Type => $Name, $IsFuzzy, $filename); | |
1833 | ||
1834 | if ($IsFuzzy) | |
1835 | { | |
1836 | my $CName = CanonicalName($Type . '_'. $Name); | |
1837 | $FuzzyNameToTest{$Name} = $Table if !$FuzzyNameToTest{$Name}; | |
1838 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; | |
1839 | } else { | |
1840 | $ExactNameToTest{$Name} = $Table; | |
1841 | } | |
1842 | ||
1843 | } | |
1844 | ||
1845 | ## Register aliase info | |
1846 | for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}}) | |
1847 | { | |
1848 | my $Alias = $AliasInfo{$Type}->{$Name}; | |
1849 | my $IsFuzzy = $FuzzyNames{$Type}->{$Alias}; | |
1850 | my $filename = $RawNameToFile{$Name}; | |
1851 | die "oops [$Alias]->[$Name]" if not $filename; | |
1852 | RegisterFileForName($Type => $Alias, $IsFuzzy, $filename); | |
1853 | ||
1854 | my $Table = $TableInfo{$Type}->{$Name}; | |
1855 | die "oops" if not $Table; | |
1856 | if ($IsFuzzy) | |
1857 | { | |
1858 | my $CName = CanonicalName($Type .'_'. $Alias); | |
1859 | $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias}; | |
1860 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; | |
1861 | } else { | |
1862 | $ExactNameToTest{$Alias} = $Table; | |
1863 | } | |
1864 | } | |
1865 | } | |
1866 | ||
1867 | ## | |
1868 | ## Write out the property list | |
1869 | ## | |
1870 | { | |
1871 | my @OUT = ( | |
1872 | "##\n", | |
1873 | "## This file created by $0\n", | |
1874 | "## List of built-in \\p{...}/\\P{...} properties.\n", | |
1875 | "##\n", | |
1876 | "## '*' means name may be 'fuzzy'\n", | |
1877 | "##\n\n", | |
1878 | sort { substr($a,2) cmp substr($b, 2) } @MAP, | |
1879 | ); | |
1880 | WriteIfChanged('Properties', @OUT); | |
1881 | } | |
1882 | ||
1883 | use Text::Tabs (); ## using this makes the files about half the size | |
1884 | ||
1885 | ## Write Exact.pl | |
1886 | { | |
1887 | my @OUT = ( | |
1888 | $HEADER, | |
1889 | "##\n", | |
1890 | "## Data in this file used by ../utf8_heavy.pl\n", | |
1891 | "##\n\n", | |
1892 | "## Mapping from name to filename in ./lib/gc_sc\n", | |
1893 | "%utf8::Exact = (\n", | |
1894 | ); | |
1895 | ||
1896 | $Exact{InGreek} = 'InGreekA'; # this is evil kludge | |
1897 | for my $Name (sort keys %Exact) | |
1898 | { | |
1899 | my $File = $Exact{$Name}; | |
1900 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; | |
1901 | my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/); | |
1902 | push @OUT, Text::Tabs::unexpand($Text); | |
1903 | } | |
1904 | push @OUT, ");\n1;\n"; | |
1905 | ||
1906 | WriteIfChanged('Exact.pl', @OUT); | |
1907 | } | |
1908 | ||
1909 | ## Write Canonical.pl | |
1910 | { | |
1911 | my @OUT = ( | |
1912 | $HEADER, | |
1913 | "##\n", | |
1914 | "## Data in this file used by ../utf8_heavy.pl\n", | |
1915 | "##\n\n", | |
1916 | "## Mapping from lc(canonical name) to filename in ./lib\n", | |
1917 | "%utf8::Canonical = (\n", | |
1918 | ); | |
1919 | my $Trail = ""; ## used just to keep the spacing pretty | |
1920 | for my $Name (sort keys %Canonical) | |
1921 | { | |
1922 | my $File = $Canonical{$Name}; | |
1923 | if ($CaComment{$Name}) { | |
1924 | push @OUT, "\n" if not $Trail; | |
1925 | push @OUT, " # $CaComment{$Name}\n"; | |
1926 | $Trail = "\n"; | |
1927 | } else { | |
1928 | $Trail = ""; | |
1929 | } | |
1930 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; | |
1931 | my $Text = sprintf(" %-41s => %s,\n$Trail", $Name, qq/'$File'/); | |
1932 | push @OUT, Text::Tabs::unexpand($Text); | |
1933 | } | |
1934 | push @OUT, ");\n1\n"; | |
1935 | WriteIfChanged('Canonical.pl', @OUT); | |
1936 | } | |
1937 | ||
1938 | MakePropTestScript() if $MakeTestScript; | |
1939 | } | |
1940 | ||
1941 | ||
1942 | sub SpecialCasing_txt() | |
1943 | { | |
1944 | # | |
1945 | # Read in the special cases. | |
1946 | # | |
1947 | ||
1948 | my %CaseInfo; | |
1949 | ||
1950 | if (not open IN, "SpecialCasing.txt") { | |
1951 | die "$0: SpecialCasing.txt: $!\n"; | |
1952 | } | |
1953 | while (<IN>) { | |
1954 | next unless /^[0-9A-Fa-f]+;/; | |
1955 | s/\#.*//; | |
1956 | s/\s+$//; | |
1957 | ||
1958 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); | |
1959 | ||
1960 | if ($condition) { # not implemented yet | |
1961 | print "# SKIPPING $_\n" if $Verbose; | |
1962 | next; | |
1963 | } | |
1964 | ||
1965 | # Wait until all the special cases have been read since | |
1966 | # they are not listed in numeric order. | |
1967 | my $ix = hex($code); | |
1968 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ] | |
1969 | unless $code eq $lower; | |
1970 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ] | |
1971 | unless $code eq $title; | |
1972 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ] | |
1973 | unless $code eq $upper; | |
1974 | } | |
1975 | close IN; | |
1976 | ||
1977 | # Now write out the special cases properties in their code point order. | |
1978 | # Prepend them to the To/{Upper,Lower,Title}.pl. | |
1979 | ||
1980 | for my $case (qw(Lower Title Upper)) | |
1981 | { | |
1982 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; | |
1983 | ||
1984 | my @OUT = | |
1985 | ( | |
1986 | $HEADER, "\n", | |
1987 | "# The key UTF-8 _bytes_, the value UTF-8 (speed hack)\n", | |
1988 | "%utf8::ToSpec$case =\n(\n", | |
1989 | ); | |
1990 | ||
1991 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { | |
1992 | my ($ix, $code, $to) = @$prop; | |
1993 | my $tostr = | |
1994 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; | |
1995 | push @OUT, sprintf qq["%s" => "$tostr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $ix))); | |
1996 | # Remove any single-character mappings for | |
1997 | # the same character since we are going for | |
1998 | # the special casing rules. | |
1999 | $NormalCase =~ s/^$code\t\t\w+\n//m; | |
2000 | } | |
2001 | push @OUT, ( | |
2002 | ");\n\n", | |
2003 | "return <<'END';\n", | |
2004 | $NormalCase, | |
2005 | "END\n" | |
2006 | ); | |
2007 | WriteIfChanged(["To","$case.pl"], @OUT); | |
2008 | } | |
2009 | } | |
2010 | ||
2011 | # | |
2012 | # Read in the case foldings. | |
2013 | # | |
2014 | # We will do full case folding, C + F + I (see CaseFolding.txt). | |
2015 | # | |
2016 | sub CaseFolding_txt() | |
2017 | { | |
2018 | if (not open IN, "CaseFolding.txt") { | |
2019 | die "$0: CaseFolding.txt: $!\n"; | |
2020 | } | |
2021 | ||
2022 | my $Fold = Table->New(); | |
2023 | my %Fold; | |
2024 | ||
2025 | while (<IN>) { | |
2026 | # Skip status 'S', simple case folding | |
2027 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; | |
2028 | ||
2029 | my ($code, $status, $fold) = (hex($1), $2, $3); | |
2030 | ||
2031 | if ($status eq 'C') { # Common: one-to-one folding | |
2032 | # No append() since several codes may fold into one. | |
2033 | $Fold->RawAppendRange($code, $code, $fold); | |
2034 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I | |
2035 | $Fold{$code} = $fold; | |
2036 | } | |
2037 | } | |
2038 | close IN; | |
2039 | ||
2040 | $Fold->Write("To/Fold.pl"); | |
2041 | ||
2042 | # | |
2043 | # Prepend the special foldings to the common foldings. | |
2044 | # | |
2045 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; | |
2046 | ||
2047 | my @OUT = | |
2048 | ( | |
2049 | $HEADER, "\n", | |
2050 | "# The ke UTF-8 _bytes_, the value UTF-8 (speed hack)\n", | |
2051 | "%utf8::ToSpecFold =\n(\n", | |
2052 | ); | |
2053 | for my $code (sort { $a <=> $b } keys %Fold) { | |
2054 | my $foldstr = | |
2055 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; | |
2056 | push @OUT, sprintf qq["%s" => "$foldstr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $code))); | |
2057 | } | |
2058 | push @OUT, ( | |
2059 | ");\n\n", | |
2060 | "return <<'END';\n", | |
2061 | $CommonFold, | |
2062 | "END\n", | |
2063 | ); | |
2064 | ||
2065 | WriteIfChanged(["To","Fold.pl"], @OUT); | |
2066 | } | |
2067 | ||
2068 | ## Do it.... | |
2069 | ||
2070 | Build_Aliases(); | |
2071 | UnicodeData_Txt(); | |
2072 | PropList_txt(); | |
2073 | ||
2074 | Scripts_txt(); | |
2075 | Blocks_txt(); | |
2076 | ||
2077 | WriteAllMappings(); | |
2078 | ||
2079 | LineBreak_Txt(); | |
2080 | ArabicShaping_txt(); | |
2081 | EastAsianWidth_txt(); | |
2082 | HangulSyllableType_txt(); | |
2083 | Jamo_txt(); | |
2084 | SpecialCasing_txt(); | |
2085 | CaseFolding_txt(); | |
2086 | ||
2087 | if ( $FileList and $MakeList ) { | |
2088 | ||
2089 | print "Updating '$FileList'\n" | |
2090 | if ($Verbose); | |
2091 | ||
2092 | open my $ofh,">",$FileList | |
2093 | or die "Can't write to '$FileList':$!"; | |
2094 | print $ofh <<"EOFHEADER"; | |
2095 | # | |
2096 | # mktables.lst -- File list for mktables. | |
2097 | # | |
2098 | # Autogenerated on @{[scalar localtime]} | |
2099 | # | |
2100 | # - First section is input files | |
2101 | # (mktables itself is automatically included) | |
2102 | # - Section seperator is /^=+\$/ | |
2103 | # - Second section is a list of output files. | |
2104 | # - Lines matching /^\\s*#/ are treated as comments | |
2105 | # which along with blank lines are ignored. | |
2106 | # | |
2107 | ||
2108 | # Input files: | |
2109 | ||
2110 | EOFHEADER | |
2111 | my @input=("version",glob('*.txt')); | |
2112 | print $ofh "$_\n" for | |
2113 | @input, | |
2114 | "\n=================================\n", | |
2115 | "# Output files:\n", | |
2116 | # special files | |
2117 | "Properties"; | |
2118 | ||
2119 | ||
2120 | require File::Find; | |
2121 | my $count=0; | |
2122 | File::Find::find({ | |
2123 | no_chdir=>1, | |
2124 | wanted=>sub { | |
2125 | if (/\.pl$/) { | |
2126 | s!^\./!!; | |
2127 | print $ofh "$_\n"; | |
2128 | $count++; | |
2129 | } | |
2130 | }, | |
2131 | },"."); | |
2132 | ||
2133 | print $ofh "\n# ",scalar(@input)," input files\n", | |
2134 | "# ",scalar($count+1)," output files\n\n", | |
2135 | "# End list\n"; | |
2136 | close $ofh | |
2137 | or warn "Failed to close $ofh: $!"; | |
2138 | ||
2139 | print "Filelist has ",scalar(@input)," input files and ", | |
2140 | scalar($count+1)," output files\n" | |
2141 | if $Verbose; | |
2142 | } | |
2143 | print "All done\n" if $Verbose; | |
2144 | exit(0); | |
2145 | ||
2146 | ## TRAILING CODE IS USED BY MakePropTestScript() | |
2147 | __DATA__ | |
2148 | use strict; | |
2149 | use warnings; | |
2150 | ||
2151 | my $Tests = 0; | |
2152 | my $Fails = 0; | |
2153 | ||
2154 | sub Expect($$$) | |
2155 | { | |
2156 | my $Expect = shift; | |
2157 | my $String = shift; | |
2158 | my $Regex = shift; | |
2159 | my $Line = (caller)[2]; | |
2160 | ||
2161 | $Tests++; | |
2162 | my $RegObj; | |
2163 | my $result = eval { | |
2164 | $RegObj = qr/$Regex/; | |
2165 | $String =~ $RegObj ? 1 : 0 | |
2166 | }; | |
2167 | ||
2168 | if (not defined $result) { | |
2169 | print "couldn't compile /$Regex/ on $0 line $Line: $@\n"; | |
2170 | $Fails++; | |
2171 | } elsif ($result ^ $Expect) { | |
2172 | print "bad result (expected $Expect) on $0 line $Line: $@\n"; | |
2173 | $Fails++; | |
2174 | } | |
2175 | } | |
2176 | ||
2177 | sub Error($) | |
2178 | { | |
2179 | my $Regex = shift; | |
2180 | $Tests++; | |
2181 | if (eval { 'x' =~ qr/$Regex/; 1 }) { | |
2182 | $Fails++; | |
2183 | my $Line = (caller)[2]; | |
2184 | print "expected error for /$Regex/ on $0 line $Line: $@\n"; | |
2185 | } | |
2186 | } | |
2187 | ||
2188 | sub Finished() | |
2189 | { | |
2190 | if ($Fails == 0) { | |
2191 | print "All $Tests tests passed.\n"; | |
2192 | exit(0); | |
2193 | } else { | |
2194 | print "$Tests tests, $Fails failed!\n"; | |
2195 | exit(-1); | |
2196 | } | |
2197 | } |