Commit | Line | Data |
00a8df5c |
1 | ## !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!! |
2 | ## Any files created or read by this program should be listed in 'mktables.lst' |
3 | |
d73e5302 |
4 | #!/usr/bin/perl -w |
4392c1cd |
5 | require 5.008; # Needs pack "U". Probably safest to run on 5.8.x |
d73e5302 |
6 | use strict; |
cf25bb62 |
7 | use Carp; |
d07a55ed |
8 | use File::Spec; |
5beb625e |
9 | |
cf25bb62 |
10 | ## |
11 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) |
12 | ## from the Unicode database files (lib/unicore/*.txt). |
13 | ## |
d2d499f5 |
14 | |
cb3d361a |
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 | |
8e4b7420 |
25 | ## Base names already used in lib/gc_sc (for avoiding 8.3 conflicts) |
26 | my %BaseNames; |
27 | |
cf25bb62 |
28 | ## |
29 | ## Process any args. |
30 | ## |
5beb625e |
31 | my $Verbose = 0; |
32 | my $MakeTestScript = 0; |
00da7942 |
33 | my $AlwaysWrite = 0; |
02b1aeec |
34 | my $UseDir = ""; |
35 | my $FileList = "$0.lst"; |
97050450 |
36 | my $MakeList = 0; |
cf25bb62 |
37 | |
38 | while (@ARGV) |
39 | { |
40 | my $arg = shift @ARGV; |
41 | if ($arg eq '-v') { |
42 | $Verbose = 1; |
43 | } elsif ($arg eq '-q') { |
44 | $Verbose = 0; |
00da7942 |
45 | } elsif ($arg eq '-w') { |
46 | $AlwaysWrite = 1; # update the files even if they havent changed |
02b1aeec |
47 | $FileList = ""; |
6ae7e459 |
48 | } elsif ($arg eq '-check') { |
49 | my $this = shift @ARGV; |
50 | my $ok = shift @ARGV; |
51 | if ($this ne $ok) { |
52 | print "Skipping as check params are not the same.\n"; |
53 | exit(0); |
54 | } |
5beb625e |
55 | } elsif ($arg eq '-maketest') { |
56 | $MakeTestScript = 1; |
97050450 |
57 | } elsif ($arg eq '-makelist') { |
58 | $MakeList = 1; |
02b1aeec |
59 | } elsif ($arg eq '-C' && defined ($UseDir = shift)) { |
60 | -d $UseDir or die "Unknown directory '$UseDir'"; |
61 | } elsif ($arg eq '-L' && defined ($FileList = shift)) { |
62 | -e $FileList or die "Filelist '$FileList' doesn't appear to exist!"; |
d73e5302 |
63 | } else { |
97050450 |
64 | die "usage: $0 [-v|-q|-w|-C dir|-L filelist] [-maketest] [-makelist]\n", |
65 | " -v : Verbose Mode\n", |
66 | " -q : Quiet Mode\n", |
67 | " -w : Write files regardless\n", |
68 | " -maketest : Make test script\n", |
69 | " -makelist : Rewrite the file list based on current setup\n", |
b6922eda |
70 | " -L filelist : Use this file list, (defaults to $0.lst)\n", |
6ae7e459 |
71 | " -C dir : Change to this directory before proceeding\n", |
72 | " -check A B : Executes only if A and B are the same\n"; |
d73e5302 |
73 | } |
74 | } |
75 | |
02b1aeec |
76 | if ($FileList) { |
77 | print "Reading file list '$FileList'\n" |
00a8df5c |
78 | if $Verbose; |
02b1aeec |
79 | open my $fh,"<",$FileList or die "Failed to read '$FileList':$!"; |
80 | my @input; |
00a8df5c |
81 | my @output; |
82 | for my $list ( \@input, \@output ) { |
83 | while (<$fh>) { |
84 | s/^ \s+ | \s+ $//xg; |
85 | next if /^ \s* (?: \# .* )? $/x; |
86 | last if /^ =+ $/x; |
87 | my ( $file ) = split /\t/, $_; |
88 | push @$list, $file; |
89 | } |
90 | my %dupe; |
91 | @$list = grep !$dupe{ $_ }++, @$list; |
92 | } |
93 | close $fh; |
02b1aeec |
94 | die "No input or output files in '$FileList'!" |
00a8df5c |
95 | if !@input or !@output; |
97050450 |
96 | if ( $MakeList ) { |
97 | foreach my $file (@output) { |
98 | unlink $file; |
99 | } |
100 | } |
00a8df5c |
101 | if ( $Verbose ) { |
102 | print "Expecting ".scalar( @input )." input files. ", |
103 | "Checking ".scalar( @output )." output files.\n"; |
104 | } |
02b1aeec |
105 | # we set maxtime to be the youngest input file, including $0 itself. |
106 | my $maxtime = -M $0; # do this before the chdir! |
107 | if ($UseDir) { |
108 | chdir $UseDir or die "Failed to chdir to '$UseDir':$!"; |
109 | } |
00a8df5c |
110 | foreach my $in (@input) { |
111 | my $time = -M $in; |
112 | die "Missing input file '$in'" unless defined $time; |
113 | $maxtime = $time if $maxtime < $time; |
114 | } |
115 | |
02b1aeec |
116 | # now we check to see if any output files are older than maxtime, if |
117 | # they are we need to continue on, otherwise we can presumably bail. |
00a8df5c |
118 | my $ok = 1; |
119 | foreach my $out (@output) { |
120 | if ( ! -e $out ) { |
121 | print "'$out' is missing.\n" |
122 | if $Verbose; |
123 | $ok = 0; |
124 | last; |
125 | } |
126 | if ( -M $out > $maxtime ) { |
127 | print "'$out' is too old.\n" |
128 | if $Verbose; |
129 | $ok = 0; |
130 | last; |
131 | } |
132 | } |
133 | if ($ok) { |
134 | print "Files seem to be ok, not bothering to rebuild.\n"; |
135 | exit(0); |
136 | } |
137 | print "Must rebuild tables.\n" |
138 | if $Verbose; |
02b1aeec |
139 | } else { |
140 | if ($Verbose) { |
141 | print "Not checking filelist.\n"; |
142 | } |
143 | if ($UseDir) { |
144 | chdir $UseDir or die "Failed to chdir to '$UseDir':$!"; |
145 | } |
00a8df5c |
146 | } |
147 | |
ca12659b |
148 | foreach my $lib ('To', 'lib', |
149 | map {File::Spec->catdir("lib",$_)} |
150 | qw(gc_sc dt bc hst ea jt lb nt ccc)) { |
151 | next if -d $lib; |
152 | mkdir $lib, 0755 or die "mkdir '$lib': $!"; |
153 | } |
154 | |
ae6979a8 |
155 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 5.1. |
d73e5302 |
156 | |
cf25bb62 |
157 | my $HEADER=<<"EOF"; |
d73e5302 |
158 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! |
551b6b6f |
159 | # This file is built by $0 from e.g. UnicodeData.txt. |
d73e5302 |
160 | # Any changes made here will be lost! |
cf25bb62 |
161 | |
162 | EOF |
163 | |
b6922eda |
164 | my $INTERNAL_ONLY=<<"EOF"; |
165 | # This file is for internal use by the Perl program only. The format and even |
166 | # name or existence of this file are subject to change without notice. Don't |
167 | # use it directly. |
168 | |
169 | EOF |
170 | |
d07a55ed |
171 | sub force_unlink { |
172 | my $filename = shift; |
173 | return unless -e $filename; |
174 | return if CORE::unlink($filename); |
175 | # We might need write permission |
176 | chmod 0777, $filename; |
177 | CORE::unlink($filename) or die "Couldn't unlink $filename: $!\n"; |
178 | } |
5beb625e |
179 | |
180 | ## |
181 | ## Given a filename and a reference to an array of lines, |
182 | ## write the lines to the file only if the contents have not changed. |
d07a55ed |
183 | ## Filename can be given as an arrayref of directory names |
5beb625e |
184 | ## |
185 | sub WriteIfChanged($\@) |
186 | { |
187 | my $file = shift; |
188 | my $lines = shift; |
189 | |
d07a55ed |
190 | $file = File::Spec->catfile(@$file) if ref $file; |
191 | |
5beb625e |
192 | my $TextToWrite = join '', @$lines; |
193 | if (open IN, $file) { |
194 | local($/) = undef; |
195 | my $PreviousText = <IN>; |
196 | close IN; |
197 | if ($PreviousText eq $TextToWrite) { |
198 | print "$file unchanged.\n" if $Verbose; |
00da7942 |
199 | return unless $AlwaysWrite; |
5beb625e |
200 | } |
201 | } |
d07a55ed |
202 | force_unlink ($file); |
5beb625e |
203 | if (not open OUT, ">$file") { |
204 | die "$0: can't open $file for output: $!\n"; |
205 | } |
206 | print "$file written.\n" if $Verbose; |
207 | |
208 | print OUT $TextToWrite; |
209 | close OUT; |
210 | } |
211 | |
cf25bb62 |
212 | ## |
213 | ## The main datastructure (a "Table") represents a set of code points that |
214 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, |
215 | ## etc.). They are kept as ranges of code points (starting and ending of |
216 | ## each range). |
217 | ## |
218 | ## For example, a range ASCII LETTERS would be represented as: |
219 | ## [ [ 0x41 => 0x5A, 'UPPER' ], |
220 | ## [ 0x61 => 0x7A, 'LOWER, ] ] |
221 | ## |
222 | sub RANGE_START() { 0 } ## index into range element |
223 | sub RANGE_END() { 1 } ## index into range element |
224 | sub RANGE_NAME() { 2 } ## index into range element |
225 | |
99598c8c |
226 | ## Conceptually, these should really be folded into the 'Table' objects |
cf25bb62 |
227 | my %TableInfo; |
99598c8c |
228 | my %TableDesc; |
cf25bb62 |
229 | my %FuzzyNames; |
230 | my %AliasInfo; |
5beb625e |
231 | my %CanonicalToOrig; |
cf25bb62 |
232 | |
233 | ## |
99598c8c |
234 | ## Turn something like |
235 | ## OLD-ITALIC |
5beb625e |
236 | ## into |
99598c8c |
237 | ## OldItalic |
238 | ## |
239 | sub CanonicalName($) |
240 | { |
5beb625e |
241 | my $orig = shift; |
242 | my $name = lc $orig; |
99598c8c |
243 | $name =~ s/(?<![a-z])(\w)/\u$1/g; |
5beb625e |
244 | $name =~ s/[-_\s]+//g; |
99598c8c |
245 | |
5beb625e |
246 | $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name}; |
99598c8c |
247 | return $name; |
248 | } |
249 | |
12ac2576 |
250 | |
251 | ## |
252 | ## Store the alias definitions for later use. |
253 | ## |
254 | my %PropertyAlias; |
255 | my %PropValueAlias; |
256 | |
257 | my %PA_reverse; |
258 | my %PVA_reverse; |
259 | |
260 | sub Build_Aliases() |
261 | { |
262 | ## |
263 | ## Most of the work with aliases doesn't occur here, |
7ebf06b3 |
264 | ## but rather in utf8_heavy.pl, which uses PVA.pl, |
265 | |
266 | # Placate the warnings about used only once. (They are used again, but |
267 | # via a typeglob lookup) |
268 | %utf8::PropertyAlias = (); |
269 | %utf8::PA_reverse = (); |
270 | %utf8::PropValueAlias = (); |
271 | %utf8::PVA_reverse = (); |
272 | %utf8::PVA_abbr_map = (); |
12ac2576 |
273 | |
274 | open PA, "< PropertyAliases.txt" |
275 | or confess "Can't open PropertyAliases.txt: $!"; |
276 | while (<PA>) { |
277 | s/#.*//; |
278 | s/\s+$//; |
279 | next if /^$/; |
280 | |
281 | my ($abbrev, $name) = split /\s*;\s*/; |
282 | next if $abbrev eq "n/a"; |
283 | $PropertyAlias{$abbrev} = $name; |
284 | $PA_reverse{$name} = $abbrev; |
7ebf06b3 |
285 | |
286 | # The %utf8::... versions use japhy's code originally from utf8_pva.pl |
287 | # However, it's moved here so that we build the tables at runtime. |
288 | tr/ _-//d for $abbrev, $name; |
289 | $utf8::PropertyAlias{lc $abbrev} = $name; |
290 | $utf8::PA_reverse{lc $name} = $abbrev; |
12ac2576 |
291 | } |
292 | close PA; |
293 | |
294 | open PVA, "< PropValueAliases.txt" |
295 | or confess "Can't open PropValueAliases.txt: $!"; |
296 | while (<PVA>) { |
297 | s/#.*//; |
298 | s/\s+$//; |
299 | next if /^$/; |
300 | |
301 | my ($prop, @data) = split /\s*;\s*/; |
302 | |
303 | if ($prop eq 'ccc') { |
304 | $PropValueAlias{$prop}{$data[1]} = [ @data[0,2] ]; |
305 | $PVA_reverse{$prop}{$data[2]} = [ @data[0,1] ]; |
306 | } |
307 | else { |
308 | next if $data[0] eq "n/a"; |
309 | $PropValueAlias{$prop}{$data[0]} = $data[1]; |
310 | $PVA_reverse{$prop}{$data[1]} = $data[0]; |
311 | } |
7ebf06b3 |
312 | |
313 | shift @data if $prop eq 'ccc'; |
314 | next if $data[0] eq "n/a"; |
315 | |
316 | $data[1] =~ tr/ _-//d; |
317 | $utf8::PropValueAlias{$prop}{lc $data[0]} = $data[1]; |
318 | $utf8::PVA_reverse{$prop}{lc $data[1]} = $data[0]; |
319 | |
320 | my $abbr_class = ($prop eq 'gc' or $prop eq 'sc') ? 'gc_sc' : $prop; |
321 | $utf8::PVA_abbr_map{$abbr_class}{lc $data[0]} = $data[0]; |
12ac2576 |
322 | } |
323 | close PVA; |
7ebf06b3 |
324 | |
325 | # backwards compatibility for L& -> LC |
326 | $utf8::PropValueAlias{gc}{'l&'} = $utf8::PropValueAlias{gc}{lc}; |
327 | $utf8::PVA_abbr_map{gc_sc}{'l&'} = $utf8::PVA_abbr_map{gc_sc}{lc}; |
328 | |
12ac2576 |
329 | } |
330 | |
331 | |
99598c8c |
332 | ## |
cf25bb62 |
333 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. |
334 | ## |
335 | ## Called like: |
99598c8c |
336 | ## New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1); |
cf25bb62 |
337 | ## |
338 | ## Normally, these parameters are set when the Table is created (when the |
339 | ## Table->New constructor is called), but there are times when it needs to |
340 | ## be done after-the-fact...) |
341 | ## |
342 | sub New_Prop($$$@) |
343 | { |
344 | my $Type = shift; ## "Is" or "In"; |
345 | my $Name = shift; |
346 | my $Table = shift; |
347 | |
348 | ## remaining args are optional key/val |
349 | my %Args = @_; |
350 | |
99598c8c |
351 | my $Fuzzy = delete $Args{Fuzzy}; |
352 | my $Desc = delete $Args{Desc}; # description |
353 | |
5beb625e |
354 | $Name = CanonicalName($Name) if $Fuzzy; |
cf25bb62 |
355 | |
356 | ## sanity check a few args |
357 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { |
358 | confess "$0: bad args to New_Prop" |
359 | } |
360 | |
361 | if (not $TableInfo{$Type}->{$Name}) |
362 | { |
363 | $TableInfo{$Type}->{$Name} = $Table; |
99598c8c |
364 | $TableDesc{$Type}->{$Name} = $Desc; |
365 | if ($Fuzzy) { |
cf25bb62 |
366 | $FuzzyNames{$Type}->{$Name} = $Name; |
367 | } |
368 | } |
d73e5302 |
369 | } |
370 | |
d73e5302 |
371 | |
cf25bb62 |
372 | ## |
373 | ## Creates a new Table object. |
374 | ## |
375 | ## Args are key/value pairs: |
99598c8c |
376 | ## In => Name -- Name of "In" property to be associated with |
377 | ## Is => Name -- Name of "Is" property to be associated with |
378 | ## Fuzzy => Boolean -- True if name can be accessed "fuzzily" |
379 | ## Desc => String -- Description of the property |
cf25bb62 |
380 | ## |
381 | ## No args are required. |
382 | ## |
383 | sub Table::New |
384 | { |
385 | my $class = shift; |
386 | my %Args = @_; |
387 | |
388 | my $Table = bless [], $class; |
389 | |
99598c8c |
390 | my $Fuzzy = delete $Args{Fuzzy}; |
391 | my $Desc = delete $Args{Desc}; |
cf25bb62 |
392 | |
393 | for my $Type ('Is', 'In') |
394 | { |
395 | if (my $Name = delete $Args{$Type}) { |
99598c8c |
396 | New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy); |
cf25bb62 |
397 | } |
398 | } |
399 | |
400 | ## shouldn't have any left over |
401 | if (%Args) { |
402 | confess "$0: bad args to Table->New" |
403 | } |
404 | |
405 | return $Table; |
d73e5302 |
406 | } |
407 | |
d73e5302 |
408 | |
cf25bb62 |
409 | ## |
410 | ## Returns the maximum code point currently in the table. |
411 | ## |
412 | sub Table::Max |
413 | { |
683a4240 |
414 | my $last = $_[0]->[-1]; ## last code point |
415 | confess "oops" unless $last; ## must have code points to have a max |
416 | return $last->[RANGE_END]; |
cf25bb62 |
417 | } |
d73e5302 |
418 | |
cf25bb62 |
419 | ## |
420 | ## Replaces the codepoints in the Table with those in the Table given |
421 | ## as an arg. (NOTE: this is not a "deep copy"). |
422 | ## |
423 | sub Table::Replace($$) |
424 | { |
425 | my $Table = shift; #self |
426 | my $New = shift; |
d73e5302 |
427 | |
cf25bb62 |
428 | @$Table = @$New; |
429 | } |
71d929cb |
430 | |
cf25bb62 |
431 | ## |
432 | ## Given a new code point, make the last range of the Table extend to |
433 | ## include the new (and all intervening) code points. |
434 | ## |
683a4240 |
435 | ## Takes the time to make sure that the extension is valid. |
436 | ## |
cf25bb62 |
437 | sub Table::Extend |
438 | { |
439 | my $Table = shift; #self |
440 | my $codepoint = shift; |
d73e5302 |
441 | |
cf25bb62 |
442 | my $PrevMax = $Table->Max; |
e904f995 |
443 | |
cf25bb62 |
444 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; |
e904f995 |
445 | |
683a4240 |
446 | $Table->ExtendNoCheck($codepoint); |
447 | } |
448 | |
449 | |
450 | ## |
451 | ## Given a new code point, make the last range of the Table extend to |
452 | ## include the new (and all intervening) code points. |
453 | ## |
454 | ## Does NOT check that the extension is valid. Assumes that the caller |
455 | ## has already made this check. |
456 | ## |
457 | sub Table::ExtendNoCheck |
458 | { |
459 | ## Optmized adding: Assumes $Table and $codepoint as parms |
460 | $_[0]->[-1]->[RANGE_END] = $_[1]; |
cf25bb62 |
461 | } |
c3a8a2b8 |
462 | |
cf25bb62 |
463 | ## |
464 | ## Given a code point range start and end (and optional name), blindly |
465 | ## append them to the list of ranges for the Table. |
466 | ## |
467 | ## NOTE: Code points must be added in strictly ascending numeric order. |
468 | ## |
469 | sub Table::RawAppendRange |
470 | { |
471 | my $Table = shift; #self |
472 | my $start = shift; |
473 | my $end = shift; |
474 | my $name = shift; |
475 | $name = "" if not defined $name; ## warning: $name can be "0" |
476 | |
477 | push @$Table, [ $start, # RANGE_START |
478 | $end, # RANGE_END |
479 | $name ]; # RANGE_NAME |
480 | } |
c3a8a2b8 |
481 | |
cf25bb62 |
482 | ## |
483 | ## Given a code point (and optional name), add it to the Table. |
484 | ## |
485 | ## NOTE: Code points must be added in strictly ascending numeric order. |
486 | ## |
487 | sub Table::Append |
488 | { |
489 | my $Table = shift; #self |
490 | my $codepoint = shift; |
491 | my $name = shift; |
492 | $name = "" if not defined $name; ## warning: $name can be "0" |
493 | |
494 | ## |
495 | ## If we've already got a range working, and this code point is the next |
496 | ## one in line, and if the name is the same, just extend the current range. |
497 | ## |
683a4240 |
498 | my $last = $Table->[-1]; |
499 | if ($last |
cf25bb62 |
500 | and |
683a4240 |
501 | $last->[RANGE_END] == $codepoint - 1 |
cf25bb62 |
502 | and |
683a4240 |
503 | $last->[RANGE_NAME] eq $name) |
cf25bb62 |
504 | { |
683a4240 |
505 | $Table->ExtendNoCheck($codepoint); |
c3a8a2b8 |
506 | } |
cf25bb62 |
507 | else |
508 | { |
509 | $Table->RawAppendRange($codepoint, $codepoint, $name); |
d73e5302 |
510 | } |
cf25bb62 |
511 | } |
d73e5302 |
512 | |
cf25bb62 |
513 | ## |
514 | ## Given a code point range starting value and ending value (and name), |
b6922eda |
515 | ## Add the range to the Table. |
cf25bb62 |
516 | ## |
517 | ## NOTE: Code points must be added in strictly ascending numeric order. |
518 | ## |
519 | sub Table::AppendRange |
520 | { |
521 | my $Table = shift; #self |
522 | my $start = shift; |
523 | my $end = shift; |
524 | my $name = shift; |
525 | $name = "" if not defined $name; ## warning: $name can be "0" |
526 | |
527 | $Table->Append($start, $name); |
528 | $Table->Extend($end) if $end > $start; |
529 | } |
d73e5302 |
530 | |
cf25bb62 |
531 | ## |
532 | ## Return a new Table that represents all code points not in the Table. |
533 | ## |
534 | sub Table::Invert |
535 | { |
536 | my $Table = shift; #self |
537 | |
538 | my $New = Table->New(); |
539 | my $max = -1; |
540 | for my $range (@$Table) |
541 | { |
542 | my $start = $range->[RANGE_START]; |
543 | my $end = $range->[RANGE_END]; |
544 | if ($start-1 >= $max+1) { |
545 | $New->AppendRange($max+1, $start-1, ""); |
546 | } |
547 | $max = $end; |
d73e5302 |
548 | } |
cf25bb62 |
549 | if ($max+1 < $LastUnicodeCodepoint) { |
550 | $New->AppendRange($max+1, $LastUnicodeCodepoint); |
d73e5302 |
551 | } |
cf25bb62 |
552 | return $New; |
553 | } |
d73e5302 |
554 | |
cf25bb62 |
555 | ## |
556 | ## Merges any number of other tables with $self, returning the new table. |
557 | ## (existing tables are not modified) |
558 | ## |
a3a8c5f0 |
559 | ## |
560 | ## Args may be Tables, or individual code points (as integers). |
561 | ## |
cf25bb62 |
562 | ## Can be called as either a constructor or a method. |
563 | ## |
564 | sub Table::Merge |
565 | { |
566 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class |
567 | my @Tables = @_; |
568 | |
569 | ## Accumulate all records from all tables |
570 | my @Records; |
a3a8c5f0 |
571 | for my $Arg (@Tables) |
572 | { |
573 | if (ref $Arg) { |
574 | ## arg is a table -- get its ranges |
575 | push @Records, @$Arg; |
576 | } else { |
577 | ## arg is a codepoint, make a range |
578 | push @Records, [ $Arg, $Arg ] |
579 | } |
d73e5302 |
580 | } |
581 | |
cf25bb62 |
582 | ## sort by range start, with longer ranges coming first. |
583 | my ($first, @Rest) = sort { |
584 | ($a->[RANGE_START] <=> $b->[RANGE_START]) |
585 | or |
586 | ($b->[RANGE_END] <=> $b->[RANGE_END]) |
587 | } @Records; |
588 | |
589 | my $New = Table->New(); |
590 | |
591 | ## Ensuring the first range is there makes the subsequent loop easier |
592 | $New->AppendRange($first->[RANGE_START], |
593 | $first->[RANGE_END]); |
594 | |
595 | ## Fold in records so long as they add new information. |
596 | for my $set (@Rest) |
597 | { |
598 | my $start = $set->[RANGE_START]; |
599 | my $end = $set->[RANGE_END]; |
600 | if ($start > $New->Max) { |
601 | $New->AppendRange($start, $end); |
602 | } elsif ($end > $New->Max) { |
683a4240 |
603 | $New->ExtendNoCheck($end); |
cf25bb62 |
604 | } |
d73e5302 |
605 | } |
d73e5302 |
606 | |
cf25bb62 |
607 | return $New; |
d73e5302 |
608 | } |
609 | |
cf25bb62 |
610 | ## |
611 | ## Given a filename, write a representation of the Table to a file. |
99598c8c |
612 | ## May have an optional comment as a 2nd arg. |
d07a55ed |
613 | ## Filename may actually be an arrayref of directories |
cf25bb62 |
614 | ## |
615 | sub Table::Write |
616 | { |
99598c8c |
617 | my $Table = shift; #self |
cf25bb62 |
618 | my $filename = shift; |
99598c8c |
619 | my $comment = shift; |
d73e5302 |
620 | |
5beb625e |
621 | my @OUT = $HEADER; |
b6922eda |
622 | |
623 | # files in subdirectories are internal-use-only |
624 | push @OUT, $INTERNAL_ONLY if ref $filename; |
625 | |
99598c8c |
626 | if (defined $comment) { |
627 | $comment =~ s/\s+\Z//; |
628 | $comment =~ s/^/# /gm; |
5beb625e |
629 | push @OUT, "#\n$comment\n#\n"; |
99598c8c |
630 | } |
5beb625e |
631 | push @OUT, "return <<'END';\n"; |
d73e5302 |
632 | |
cf25bb62 |
633 | for my $set (@$Table) |
634 | { |
635 | my $start = $set->[RANGE_START]; |
636 | my $end = $set->[RANGE_END]; |
637 | my $name = $set->[RANGE_NAME]; |
d73e5302 |
638 | |
cf25bb62 |
639 | if ($start == $end) { |
5beb625e |
640 | push @OUT, sprintf "%04X\t\t%s\n", $start, $name; |
cf25bb62 |
641 | } else { |
5beb625e |
642 | push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name; |
cf25bb62 |
643 | } |
644 | } |
d73e5302 |
645 | |
5beb625e |
646 | push @OUT, "END\n"; |
647 | |
648 | WriteIfChanged($filename, @OUT); |
649 | } |
650 | |
651 | ## This used only for making the test script. |
652 | ## helper function |
653 | sub IsUsable($) |
654 | { |
655 | my $code = shift; |
656 | return 0 if $code <= 0x0000; ## don't use null |
657 | return 0 if $code >= $LastUnicodeCodepoint; ## keep in range |
658 | return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates |
659 | return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good |
660 | return 0 if (($code & 0xFFFF) == 0xFFFE); ## utf8.c says no good |
661 | return 0 if (($code & 0xFFFF) == 0xFFFF); ## utf8.c says no good |
662 | return 1; |
663 | } |
664 | |
665 | ## Return a code point that's part of the table. |
666 | ## Returns nothing if the table is empty (or covers only surrogates). |
667 | ## This used only for making the test script. |
668 | sub Table::ValidCode |
669 | { |
670 | my $Table = shift; #self |
671 | for my $set (@$Table) { |
672 | return $set->[RANGE_END] if IsUsable($set->[RANGE_END]); |
673 | } |
674 | return (); |
675 | } |
676 | |
677 | ## Return a code point that's not part of the table |
678 | ## Returns nothing if the table covers all code points. |
679 | ## This used only for making the test script. |
680 | sub Table::InvalidCode |
681 | { |
682 | my $Table = shift; #self |
683 | |
683a4240 |
684 | return 0x1234 if not @$Table; |
5beb625e |
685 | |
686 | for my $set (@$Table) |
687 | { |
688 | if (IsUsable($set->[RANGE_END] + 1)) |
689 | { |
690 | return $set->[RANGE_END] + 1; |
691 | } |
692 | |
693 | if (IsUsable($set->[RANGE_START] - 1)) |
694 | { |
695 | return $set->[RANGE_START] - 1; |
696 | } |
697 | } |
698 | return (); |
cf25bb62 |
699 | } |
700 | |
701 | ########################################################################### |
702 | ########################################################################### |
703 | ########################################################################### |
704 | |
705 | |
706 | ## |
707 | ## Called like: |
99598c8c |
708 | ## New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1); |
cf25bb62 |
709 | ## |
99598c8c |
710 | ## The args must be in that order, although the Fuzzy pair may be omitted. |
cf25bb62 |
711 | ## |
712 | ## This creates 'IsAll' as an alias for 'IsAny' |
713 | ## |
714 | sub New_Alias($$$@) |
715 | { |
716 | my $Type = shift; ## "Is" or "In" |
717 | my $Alias = shift; |
99598c8c |
718 | my $SameAs = shift; # expecting "SameAs" -- just ignored |
cf25bb62 |
719 | my $Name = shift; |
720 | |
721 | ## remaining args are optional key/val |
722 | my %Args = @_; |
723 | |
99598c8c |
724 | my $Fuzzy = delete $Args{Fuzzy}; |
cf25bb62 |
725 | |
726 | ## sanity check a few args |
727 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { |
728 | confess "$0: bad args to New_Alias" |
d73e5302 |
729 | } |
730 | |
5beb625e |
731 | $Alias = CanonicalName($Alias) if $Fuzzy; |
732 | |
733 | if (not $TableInfo{$Type}->{$Name}) |
734 | { |
735 | my $CName = CanonicalName($Name); |
736 | if ($TableInfo{$Type}->{$CName}) { |
737 | confess "$0: Use canonical form '$CName' instead of '$Name' for alias."; |
738 | } else { |
12ac2576 |
739 | confess "$0: don't have original $Type => $Name to make alias\n"; |
5beb625e |
740 | } |
cf25bb62 |
741 | } |
742 | if ($TableInfo{$Alias}) { |
743 | confess "$0: already have original $Type => $Alias; can't make alias"; |
d73e5302 |
744 | } |
cf25bb62 |
745 | $AliasInfo{$Type}->{$Name} = $Alias; |
99598c8c |
746 | if ($Fuzzy) { |
cf25bb62 |
747 | $FuzzyNames{$Type}->{$Alias} = $Name; |
748 | } |
749 | |
d73e5302 |
750 | } |
751 | |
d73e5302 |
752 | |
cf25bb62 |
753 | ## All assigned code points |
99598c8c |
754 | my $Assigned = Table->New(Is => 'Assigned', |
755 | Desc => "All assigned code points", |
5beb625e |
756 | Fuzzy => 0); |
d2d499f5 |
757 | |
cf25bb62 |
758 | my $Name = Table->New(); ## all characters, individually by name |
759 | my $General = Table->New(); ## all characters, grouped by category |
760 | my %General; |
761 | my %Cat; |
d73e5302 |
762 | |
ae6979a8 |
763 | ## Simple Data::Dumper like. Good enough for our needs. We can't use the real |
7ebf06b3 |
764 | ## thing as we have to run under miniperl |
765 | sub simple_dumper { |
766 | my @lines; |
767 | my $item; |
768 | foreach $item (@_) { |
769 | if (ref $item) { |
770 | if (ref $item eq 'ARRAY') { |
771 | push @lines, "[\n", simple_dumper (@$item), "],\n"; |
772 | } elsif (ref $item eq 'HASH') { |
773 | push @lines, "{\n", simple_dumper (%$item), "},\n"; |
774 | } else { |
775 | die "Can't cope with $item"; |
776 | } |
777 | } else { |
778 | if (defined $item) { |
779 | my $copy = $item; |
780 | $copy =~ s/([\'\\])/\\$1/gs; |
781 | push @lines, "'$copy',\n"; |
782 | } else { |
783 | push @lines, "undef,\n"; |
784 | } |
785 | } |
786 | } |
787 | @lines; |
788 | } |
789 | |
cf25bb62 |
790 | ## |
551b6b6f |
791 | ## Process UnicodeData.txt (Categories, etc.) |
cf25bb62 |
792 | ## |
da7fcca4 |
793 | # These are the character mappings as defined in the POSIX standard |
794 | # and in the case of PerlSpace and PerlWord as is defined in the test macros |
795 | # for binary strings. IOW, PerlWord is [A-Za-z_] and PerlSpace is [\f\r\n\t ] |
796 | # This differs from Word and the existing SpacePerl (note the prefix/suffix difference) |
797 | # which is basically the Unicode WhiteSpace without the vertical tab included |
798 | # |
799 | my %TRUE_POSIX_PERL_CC= ( |
800 | PosixAlnum => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x005a, 0x0061..0x007a )}, |
801 | PosixAlpha => { map { $_ => 1 } ( 0x0041..0x005a, 0x0061..0x007a )}, |
802 | # Not Needed: Ascii => { map { $_ => 1 } ( 0x0000..0x007f )}, |
803 | PosixBlank => { map { $_ => 1 } ( 0x0009, 0x0020 )}, |
804 | PosixCntrl => { map { $_ => 1 } ( 0x0000..0x001f, 0x007f )}, |
805 | PosixGraph => { map { $_ => 1 } ( 0x0021..0x007e )}, |
806 | PosixLower => { map { $_ => 1 } ( 0x0061..0x007a )}, |
807 | PosixPrint => { map { $_ => 1 } ( 0x0020..0x007e )}, |
808 | PosixPunct => { map { $_ => 1 } ( 0x0021..0x002f, 0x003a..0x0040, 0x005b..0x0060, 0x007b..0x007e )}, |
809 | PosixSpace => { map { $_ => 1 } ( 0x0009..0x000d, 0x0020 )}, |
810 | PosixUpper => { map { $_ => 1 } ( 0x0041..0x005a )}, |
811 | # Not needed: PosixXdigit => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x0046, 0x0061..0x0066 )}, |
812 | PosixDigit => { map { $_ => 1 } ( 0x0030..0x0039 )}, |
813 | |
814 | PerlSpace => { map { $_ => 1 } ( 0x0009..0x000a, 0x000c..0x000d, 0x0020 )}, |
815 | PerlWord => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x005a, 0x005f, 0x0061..0x007a )}, |
816 | ); |
817 | |
44da8cae |
818 | sub UnicodeData_Txt() |
cf25bb62 |
819 | { |
820 | my $Bidi = Table->New(); |
821 | my $Deco = Table->New(); |
822 | my $Comb = Table->New(); |
823 | my $Number = Table->New(); |
12ac2576 |
824 | my $Mirrored = Table->New();#Is => 'Mirrored', |
825 | #Desc => "Mirrored in bidirectional text", |
826 | #Fuzzy => 0); |
d73e5302 |
827 | |
cf25bb62 |
828 | my %DC; |
829 | my %Bidi; |
12ac2576 |
830 | my %Number; |
b6922eda |
831 | $DC{Can} = Table->New(); |
832 | $DC{Com} = Table->New(); |
cf25bb62 |
833 | |
da7fcca4 |
834 | ## Initialize Broken Perl-generated categories |
551b6b6f |
835 | ## (Categories from UnicodeData.txt are auto-initialized in gencat) |
44da8cae |
836 | $Cat{Alnum} = |
837 | Table->New(Is => 'Alnum', Desc => "[[:Alnum:]]", Fuzzy => 0); |
838 | $Cat{Alpha} = |
839 | Table->New(Is => 'Alpha', Desc => "[[:Alpha:]]", Fuzzy => 0); |
840 | $Cat{ASCII} = |
841 | Table->New(Is => 'ASCII', Desc => "[[:ASCII:]]", Fuzzy => 0); |
842 | $Cat{Blank} = |
843 | Table->New(Is => 'Blank', Desc => "[[:Blank:]]", Fuzzy => 0); |
844 | $Cat{Cntrl} = |
845 | Table->New(Is => 'Cntrl', Desc => "[[:Cntrl:]]", Fuzzy => 0); |
846 | $Cat{Digit} = |
847 | Table->New(Is => 'Digit', Desc => "[[:Digit:]]", Fuzzy => 0); |
848 | $Cat{Graph} = |
849 | Table->New(Is => 'Graph', Desc => "[[:Graph:]]", Fuzzy => 0); |
850 | $Cat{Lower} = |
851 | Table->New(Is => 'Lower', Desc => "[[:Lower:]]", Fuzzy => 0); |
852 | $Cat{Print} = |
853 | Table->New(Is => 'Print', Desc => "[[:Print:]]", Fuzzy => 0); |
854 | $Cat{Punct} = |
855 | Table->New(Is => 'Punct', Desc => "[[:Punct:]]", Fuzzy => 0); |
856 | $Cat{Space} = |
857 | Table->New(Is => 'Space', Desc => "[[:Space:]]", Fuzzy => 0); |
858 | $Cat{Title} = |
859 | Table->New(Is => 'Title', Desc => "[[:Title:]]", Fuzzy => 0); |
860 | $Cat{Upper} = |
861 | Table->New(Is => 'Upper', Desc => "[[:Upper:]]", Fuzzy => 0); |
862 | $Cat{XDigit} = |
863 | Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0); |
864 | $Cat{Word} = |
865 | Table->New(Is => 'Word', Desc => "[[:Word:]]", Fuzzy => 0); |
866 | $Cat{SpacePerl} = |
867 | Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0); |
e1d1eefb |
868 | $Cat{VertSpace} = |
869 | Table->New(Is => 'VertSpace', Desc => '\v', Fuzzy => 0); |
870 | $Cat{HorizSpace} = |
871 | Table->New(Is => 'HorizSpace', Desc => '\h', Fuzzy => 0); |
cf25bb62 |
872 | my %To; |
873 | $To{Upper} = Table->New(); |
874 | $To{Lower} = Table->New(); |
875 | $To{Title} = Table->New(); |
876 | $To{Digit} = Table->New(); |
877 | |
da7fcca4 |
878 | foreach my $cat (keys %TRUE_POSIX_PERL_CC) { |
879 | $Cat{$cat} = Table->New(Is=>$cat, Fuzzy => 0); |
880 | } |
881 | |
cf25bb62 |
882 | sub gencat($$$$) |
883 | { |
884 | my ($name, ## Name ("LATIN CAPITAL LETTER A") |
885 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) |
886 | $code, ## Code point (as an integer) |
887 | $op) = @_; |
888 | |
889 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc |
890 | |
891 | $Assigned->$op($code); |
892 | $Name->$op($code, $name); |
893 | $General->$op($code, $cat); |
894 | |
895 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) |
99598c8c |
896 | $Cat{$cat} ||= Table->New(Is => $cat, |
897 | Desc => "General Category '$cat'", |
898 | Fuzzy => 0); |
cf25bb62 |
899 | $Cat{$cat}->$op($code); |
900 | |
901 | ## add to the major category (e.g. "L", "N", "C", ...) |
99598c8c |
902 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, |
903 | Desc => "Major Category '$MajorCat'", |
904 | Fuzzy => 0); |
cf25bb62 |
905 | $Cat{$MajorCat}->$op($code); |
906 | |
907 | ($General{$name} ||= Table->New)->$op($code, $name); |
908 | |
909 | # 005F: SPACING UNDERSCORE |
d75d706f |
910 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]|Pc/; |
c65e4d19 |
911 | $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/; |
cf25bb62 |
912 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; |
913 | |
d75d706f |
914 | my $isspace = |
915 | ($cat =~ /Zs|Zl|Zp/ && |
b6922eda |
916 | $code != 0x200B) # 200B is ZWSP which is for line break control |
917 | # and therefore it is not part of "space" even |
918 | # while it is "Zs" in some versions of Unicode. |
919 | # In 5.1 it is Cf, so this line is no longer |
920 | # necessary. |
cf25bb62 |
921 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
922 | || $code == 0x000A # 000A: LINE FEED |
923 | || $code == 0x000B # 000B: VERTICAL TAB |
924 | || $code == 0x000C # 000C: FORM FEED |
e6f58734 |
925 | || $code == 0x000D # 000D: CARRIAGE RETURN |
d75d706f |
926 | || $code == 0x0085 # 0085: NEL |
927 | |
928 | ; |
cf25bb62 |
929 | |
d75d706f |
930 | $Cat{Space}->$op($code) if $isspace; |
cf25bb62 |
931 | |
d75d706f |
932 | $Cat{SpacePerl}->$op($code) if $isspace |
933 | && $code != 0x000B; # Backward compat. |
cf25bb62 |
934 | |
e1d1eefb |
935 | $Cat{VertSpace}->$op($code) if grep {$code == $_} |
936 | ( 0x0A..0x0D,0x85,0x2028,0x2029 ); |
937 | |
938 | $Cat{HorizSpace}->$op($code) if grep {$code == $_} ( |
939 | 0x09, 0x20, 0xa0, 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, |
940 | 0x2003, 0x2004, 0x2005, 0x2006, 0x2007, 0x2008, 0x2009, 0x200a, |
941 | 0x202f, 0x205f, 0x3000 |
942 | ); |
943 | |
d75d706f |
944 | $Cat{Blank}->$op($code) if $isspace |
945 | && !($code == 0x000A || |
946 | $code == 0x000B || |
947 | $code == 0x000C || |
948 | $code == 0x000D || |
949 | $code == 0x0085 || |
950 | $cat =~ /^Z[lp]/); |
cf25bb62 |
951 | |
952 | $Cat{Digit}->$op($code) if $cat eq "Nd"; |
953 | $Cat{Upper}->$op($code) if $cat eq "Lu"; |
954 | $Cat{Lower}->$op($code) if $cat eq "Ll"; |
955 | $Cat{Title}->$op($code) if $cat eq "Lt"; |
956 | $Cat{ASCII}->$op($code) if $code <= 0x007F; |
957 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; |
d75d706f |
958 | my $isgraph = !$isspace && $cat !~ /Cc|Cs|Cn/; |
959 | $Cat{Graph}->$op($code) if $isgraph; |
960 | $Cat{Print}->$op($code) if $isgraph || $isspace; |
cf25bb62 |
961 | $Cat{Punct}->$op($code) if $cat =~ /^P/; |
962 | |
963 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 |
964 | || ($code >= 0x41 && $code <= 0x46) ## A..F |
965 | || ($code >= 0x61 && $code <= 0x66); ## a..f |
da7fcca4 |
966 | if ($code<=0x7F) { |
967 | foreach my $cat (keys %TRUE_POSIX_PERL_CC) { |
968 | if ($TRUE_POSIX_PERL_CC{$cat}{$code}) { |
969 | $Cat{$cat}->$op($code); |
970 | } |
971 | } |
972 | } |
cf25bb62 |
973 | } |
d73e5302 |
974 | |
ae6979a8 |
975 | ## open and read file..... |
551b6b6f |
976 | if (not open IN, "UnicodeData.txt") { |
977 | die "$0: UnicodeData.txt: $!\n"; |
cf25bb62 |
978 | } |
d73e5302 |
979 | |
a3a8c5f0 |
980 | ## |
981 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} |
982 | ## |
983 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") |
984 | |
985 | my %CodeToDeco; ## Maps code to decomp. list for chars with first |
986 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) |
987 | |
988 | ## This is filled in as we go.... |
99598c8c |
989 | my $CombAbove = Table->New(Is => '_CombAbove', |
990 | Desc => '(for internal casefolding use)', |
991 | Fuzzy => 0); |
a3a8c5f0 |
992 | |
cf25bb62 |
993 | while (<IN>) |
994 | { |
995 | next unless /^[0-9A-Fa-f]+;/; |
996 | s/\s+$//; |
997 | |
998 | my ($hexcode, ## code point in hex (e.g. "0041") |
999 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") |
1000 | $cat, ## category (e.g. "Lu") |
b6922eda |
1001 | $comb, ## Canonical combining class (e.g. "230") |
cf25bb62 |
1002 | $bidi, ## directional category (e.g. "L") |
1003 | $deco, ## decomposition mapping |
1004 | $decimal, ## decimal digit value |
1005 | $digit, ## digit value |
1006 | $number, ## numeric value |
1007 | $mirrored, ## mirrored |
1008 | $unicode10, ## name in Unicode 1.0 |
1009 | $comment, ## comment field |
1010 | $upper, ## uppercase mapping |
1011 | $lower, ## lowercase mapping |
1012 | $title, ## titlecase mapping |
1013 | ) = split(/\s*;\s*/); |
1014 | |
2eb5892f |
1015 | # Note that in Unicode 3.2 there will be names like |
1016 | # LINE FEED (LF), which probably means that \N{} needs |
1017 | # to cope also with LINE FEED and LF. |
1018 | $name = $unicode10 if $name eq '<control>' && $unicode10 ne ''; |
1019 | |
cf25bb62 |
1020 | my $code = hex($hexcode); |
1021 | |
a3a8c5f0 |
1022 | if ($comb and $comb == 230) { |
1023 | $CombAbove->Append($code); |
1024 | $_Above_HexCodes{$hexcode} = 1; |
1025 | } |
1026 | |
1027 | ## Used in building \p{_CanonDCIJ} |
1028 | if ($deco and $deco =~ m/^006[9A]\b/) { |
1029 | $CodeToDeco{$code} = $deco; |
1030 | } |
1031 | |
cf25bb62 |
1032 | ## |
1033 | ## There are a few pairs of lines like: |
1034 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; |
1035 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; |
1036 | ## that define ranges. |
1037 | ## |
1038 | if ($name =~ /^<(.+), (First|Last)>$/) |
1039 | { |
1040 | $name = $1; |
1041 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); |
99598c8c |
1042 | #New_Prop(In => $name, $General{$name}, Fuzzy => 1); |
cf25bb62 |
1043 | } |
1044 | else |
1045 | { |
1046 | ## normal (single-character) lines |
1047 | gencat($name, $cat, $code, 'Append'); |
1048 | |
1049 | # No Append() here since since several codes may map into one. |
1050 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; |
1051 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; |
1052 | $To{Title}->RawAppendRange($code, $code, $title) if $title; |
1053 | $To{Digit}->Append($code, $decimal) if length $decimal; |
1054 | |
1055 | $Bidi->Append($code, $bidi); |
1056 | $Comb->Append($code, $comb) if $comb; |
1057 | $Number->Append($code, $number) if length $number; |
1058 | |
12ac2576 |
1059 | length($decimal) and ($Number{De} ||= Table->New())->Append($code) |
1060 | or |
1061 | length($digit) and ($Number{Di} ||= Table->New())->Append($code) |
1062 | or |
1063 | length($number) and ($Number{Nu} ||= Table->New())->Append($code); |
1064 | |
cf25bb62 |
1065 | $Mirrored->Append($code) if $mirrored eq "Y"; |
1066 | |
12ac2576 |
1067 | $Bidi{$bidi} ||= Table->New();#Is => "bt/$bidi", |
1068 | #Desc => "Bi-directional category '$bidi'", |
1069 | #Fuzzy => 0); |
cf25bb62 |
1070 | $Bidi{$bidi}->Append($code); |
1071 | |
1072 | if ($deco) |
1073 | { |
1074 | $Deco->Append($code, $deco); |
1075 | if ($deco =~/^<(\w+)>/) |
1076 | { |
12ac2576 |
1077 | my $dshort = $PVA_reverse{dt}{ucfirst lc $1}; |
b6922eda |
1078 | $DC{Com}->Append($code); |
1079 | $dshort = $PVA_reverse{dt}{lc $1} unless $dshort ne ""; |
1080 | die "No reverse for $1'" unless $dshort ne ""; |
1081 | #$dshort = lc $dshort; # use lower case only |
1082 | $DC{$dshort} ||= Table->New(); |
1083 | $DC{$dshort}->Append($code); |
cf25bb62 |
1084 | } |
1085 | else |
1086 | { |
b6922eda |
1087 | $DC{Can}->Append($code); |
cf25bb62 |
1088 | } |
1089 | } |
1090 | } |
1091 | } |
1092 | close IN; |
d2d499f5 |
1093 | |
ae6979a8 |
1094 | ## Read in the NameAliases.txt. It contains other normative names of code |
1095 | ## points not listed in UnicodeData.txt. This happens when there is an |
1096 | ## error in the name found after the data base was published, but instead of |
1097 | ## changing it, to avoid breaking any code that came to rely on the |
1098 | ## erroneous version, the correct name is added as an alias. |
1099 | |
1100 | my $NameAliases = Table->New(); |
1101 | |
1102 | if (not open IN, "NameAliases.txt") { |
1103 | die "$0: NameAliases.txt: $!\n"; |
1104 | } |
1105 | |
1106 | while (<IN>) |
1107 | { |
1108 | next unless /^[0-9A-Fa-f]+;/; |
1109 | s/\s+$//; |
1110 | |
1111 | my ($hexcode, ## code point in hex (e.g. "0041") |
1112 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") |
1113 | ) = split(/\s*;\s*/); |
1114 | |
1115 | my $code = hex($hexcode); |
1116 | |
1117 | ## One is supposed to enter elements into tables in strictly increasing |
1118 | ## order, but this in fact works to append duplicate code points at |
1119 | ## the end of the table. The table is intended to be indexed by name |
1120 | ## anyway. |
1121 | |
1122 | $Name->RawAppendRange($code, $code, $name); |
1123 | } |
1124 | close IN; |
1125 | |
1126 | |
cf25bb62 |
1127 | ## |
1128 | ## Tidy up a few special cases.... |
1129 | ## |
d73e5302 |
1130 | |
cf25bb62 |
1131 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist |
99598c8c |
1132 | New_Prop(Is => 'Cn', |
1133 | $Cat{Cn}, |
1134 | Desc => "General Category 'Cn' [not functional in Perl]", |
1135 | Fuzzy => 0); |
d73e5302 |
1136 | |
cf25bb62 |
1137 | ## Unassigned is the same as 'Cn' |
5beb625e |
1138 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0); |
d73e5302 |
1139 | |
cf25bb62 |
1140 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C |
d73e5302 |
1141 | |
d73e5302 |
1142 | |
12ac2576 |
1143 | # LC is Ll, Lu, and Lt. |
1144 | # (used to be L& or L_, but PropValueAliases.txt defines it as LC) |
1145 | New_Prop(Is => 'LC', |
cf25bb62 |
1146 | Table->Merge(@Cat{qw[Ll Lu Lt]}), |
99598c8c |
1147 | Desc => '[\p{Ll}\p{Lu}\p{Lt}]', |
1148 | Fuzzy => 0); |
d73e5302 |
1149 | |
cf25bb62 |
1150 | ## Any and All are all code points. |
99598c8c |
1151 | my $Any = Table->New(Is => 'Any', |
1152 | Desc => sprintf("[\\x{0000}-\\x{%X}]", |
1153 | $LastUnicodeCodepoint), |
5beb625e |
1154 | Fuzzy => 0); |
cf25bb62 |
1155 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); |
d73e5302 |
1156 | |
5beb625e |
1157 | New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0); |
d73e5302 |
1158 | |
a3a8c5f0 |
1159 | ## |
1160 | ## Build special properties for Perl's internal case-folding needs: |
1161 | ## \p{_CaseIgnorable} |
1162 | ## \p{_CanonDCIJ} |
1163 | ## \p{_CombAbove} |
1164 | ## _CombAbove was built above. Others are built here.... |
1165 | ## |
1166 | |
1167 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] |
1168 | New_Prop(Is => '_CaseIgnorable', |
1169 | Table->Merge($Cat{Mn}, |
1170 | 0x00AD, #SOFT HYPHEN |
1171 | 0x2010), #HYPHEN |
99598c8c |
1172 | Desc => '(for internal casefolding use)', |
1173 | Fuzzy => 0); |
a3a8c5f0 |
1174 | |
1175 | |
1176 | ## \p{_CanonDCIJ} is fairly complex... |
99598c8c |
1177 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', |
1178 | Desc => '(for internal casefolding use)', |
1179 | Fuzzy => 0); |
a3a8c5f0 |
1180 | ## It contains the ASCII 'i' and 'j'.... |
1181 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") |
1182 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") |
1183 | ## ...and any character with a decomposition that starts with either of |
1184 | ## those code points, but only if the decomposition does not have any |
1185 | ## combining character with the "ABOVE" canonical combining class. |
1186 | for my $code (sort { $a <=> $b} keys %CodeToDeco) |
1187 | { |
1188 | ## Need to ensure that all decomposition characters do not have |
1189 | ## a %HexCodeToComb in %AboveCombClasses. |
1190 | my $want = 1; |
1191 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) |
1192 | { |
1193 | if (exists $_Above_HexCodes{$deco_hexcode}) { |
1194 | ## one of the decmposition chars has an ABOVE combination |
1195 | ## class, so we're not interested in this one |
1196 | $want = 0; |
1197 | last; |
1198 | } |
1199 | } |
1200 | if ($want) { |
1201 | $CanonCDIJ->Append($code); |
1202 | } |
1203 | } |
1204 | |
1205 | |
d73e5302 |
1206 | |
cf25bb62 |
1207 | ## |
1208 | ## Now dump the files. |
1209 | ## |
1210 | $Name->Write("Name.pl"); |
12ac2576 |
1211 | |
7ebf06b3 |
1212 | { |
1213 | my @PVA = $HEADER; |
1214 | foreach my $name (qw (PropertyAlias PA_reverse PropValueAlias |
1215 | PVA_reverse PVA_abbr_map)) { |
1216 | # Should I really jump through typeglob hoops just to avoid a |
1217 | # symbolic reference? (%{"utf8::$name}) |
1218 | push @PVA, "\n", "\%utf8::$name = (\n", |
1219 | simple_dumper (%{$utf8::{$name}}), ");\n"; |
1220 | } |
d403599d |
1221 | push @PVA, "1;\n"; |
7ebf06b3 |
1222 | WriteIfChanged("PVA.pl", @PVA); |
1223 | } |
1224 | |
12ac2576 |
1225 | # $Bidi->Write("Bidirectional.pl"); |
12ac2576 |
1226 | for (keys %Bidi) { |
1227 | $Bidi{$_}->Write( |
d07a55ed |
1228 | ["lib","bc","$_.pl"], |
12ac2576 |
1229 | "BidiClass category '$PropValueAlias{bc}{$_}'" |
1230 | ); |
1231 | } |
1232 | |
cf25bb62 |
1233 | $Comb->Write("CombiningClass.pl"); |
12ac2576 |
1234 | for (keys %{ $PropValueAlias{ccc} }) { |
1235 | my ($code, $name) = @{ $PropValueAlias{ccc}{$_} }; |
1236 | (my $c = Table->New())->Append($code); |
1237 | $c->Write( |
d07a55ed |
1238 | ["lib","ccc","$_.pl"], |
12ac2576 |
1239 | "CombiningClass category '$name'" |
1240 | ); |
1241 | } |
1242 | |
cf25bb62 |
1243 | $Deco->Write("Decomposition.pl"); |
12ac2576 |
1244 | for (keys %DC) { |
1245 | $DC{$_}->Write( |
d07a55ed |
1246 | ["lib","dt","$_.pl"], |
12ac2576 |
1247 | "DecompositionType category '$PropValueAlias{dt}{$_}'" |
1248 | ); |
1249 | } |
1250 | |
1251 | # $Number->Write("Number.pl"); |
12ac2576 |
1252 | for (keys %Number) { |
1253 | $Number{$_}->Write( |
d07a55ed |
1254 | ["lib","nt","$_.pl"], |
12ac2576 |
1255 | "NumericType category '$PropValueAlias{nt}{$_}'" |
1256 | ); |
1257 | } |
1258 | |
1259 | # $General->Write("Category.pl"); |
cf25bb62 |
1260 | |
1261 | for my $to (sort keys %To) { |
d07a55ed |
1262 | $To{$to}->Write(["To","$to.pl"]); |
d73e5302 |
1263 | } |
12ac2576 |
1264 | |
1265 | for (keys %{ $PropValueAlias{gc} }) { |
1266 | New_Alias(Is => $PropValueAlias{gc}{$_}, SameAs => $_, Fuzzy => 1); |
1267 | } |
d73e5302 |
1268 | } |
1269 | |
cf25bb62 |
1270 | ## |
551b6b6f |
1271 | ## Process LineBreak.txt |
cf25bb62 |
1272 | ## |
551b6b6f |
1273 | sub LineBreak_Txt() |
cf25bb62 |
1274 | { |
551b6b6f |
1275 | if (not open IN, "LineBreak.txt") { |
1276 | die "$0: LineBreak.txt: $!\n"; |
cf25bb62 |
1277 | } |
d73e5302 |
1278 | |
cf25bb62 |
1279 | my $Lbrk = Table->New(); |
1280 | my %Lbrk; |
d73e5302 |
1281 | |
cf25bb62 |
1282 | while (<IN>) |
1283 | { |
1284 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; |
d73e5302 |
1285 | |
cf25bb62 |
1286 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); |
d73e5302 |
1287 | |
cf25bb62 |
1288 | $Lbrk->Append($first, $lbrk); |
d73e5302 |
1289 | |
12ac2576 |
1290 | $Lbrk{$lbrk} ||= Table->New(); |
cf25bb62 |
1291 | $Lbrk{$lbrk}->Append($first); |
d73e5302 |
1292 | |
cf25bb62 |
1293 | if ($last) { |
1294 | $Lbrk->Extend($last); |
1295 | $Lbrk{$lbrk}->Extend($last); |
d73e5302 |
1296 | } |
1297 | } |
cf25bb62 |
1298 | close IN; |
d73e5302 |
1299 | |
12ac2576 |
1300 | # $Lbrk->Write("Lbrk.pl"); |
1301 | |
12ac2576 |
1302 | |
1303 | for (keys %Lbrk) { |
1304 | $Lbrk{$_}->Write( |
d07a55ed |
1305 | ["lib","lb","$_.pl"], |
12ac2576 |
1306 | "Linebreak category '$PropValueAlias{lb}{$_}'" |
1307 | ); |
1308 | } |
cf25bb62 |
1309 | } |
d73e5302 |
1310 | |
cf25bb62 |
1311 | ## |
551b6b6f |
1312 | ## Process ArabicShaping.txt. |
cf25bb62 |
1313 | ## |
551b6b6f |
1314 | sub ArabicShaping_txt() |
cf25bb62 |
1315 | { |
551b6b6f |
1316 | if (not open IN, "ArabicShaping.txt") { |
1317 | die "$0: ArabicShaping.txt: $!\n"; |
cf25bb62 |
1318 | } |
d73e5302 |
1319 | |
cf25bb62 |
1320 | my $ArabLink = Table->New(); |
1321 | my $ArabLinkGroup = Table->New(); |
d73e5302 |
1322 | |
12ac2576 |
1323 | my %JoinType; |
1324 | |
cf25bb62 |
1325 | while (<IN>) |
1326 | { |
1327 | next unless /^[0-9A-Fa-f]+;/; |
1328 | s/\s+$//; |
d73e5302 |
1329 | |
cf25bb62 |
1330 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); |
1331 | my $code = hex($hexcode); |
1332 | $ArabLink->Append($code, $link); |
1333 | $ArabLinkGroup->Append($code, $linkgroup); |
12ac2576 |
1334 | |
1335 | $JoinType{$link} ||= Table->New(Is => "JoinType$link"); |
1336 | $JoinType{$link}->Append($code); |
d73e5302 |
1337 | } |
cf25bb62 |
1338 | close IN; |
1339 | |
12ac2576 |
1340 | # $ArabLink->Write("ArabLink.pl"); |
1341 | # $ArabLinkGroup->Write("ArabLnkGrp.pl"); |
1342 | |
12ac2576 |
1343 | |
1344 | for (keys %JoinType) { |
1345 | $JoinType{$_}->Write( |
d07a55ed |
1346 | ["lib","jt","$_.pl"], |
12ac2576 |
1347 | "JoiningType category '$PropValueAlias{jt}{$_}'" |
1348 | ); |
1349 | } |
1350 | } |
1351 | |
1352 | ## |
1353 | ## Process EastAsianWidth.txt. |
1354 | ## |
1355 | sub EastAsianWidth_txt() |
1356 | { |
1357 | if (not open IN, "EastAsianWidth.txt") { |
1358 | die "$0: EastAsianWidth.txt: $!\n"; |
1359 | } |
1360 | |
1361 | my %EAW; |
1362 | |
1363 | while (<IN>) |
1364 | { |
adc6a8bd |
1365 | next unless /^[0-9A-Fa-f]+(\.\.[0-9A-Fa-f]+)?;/; |
12ac2576 |
1366 | s/#.*//; |
1367 | s/\s+$//; |
1368 | |
adc6a8bd |
1369 | my ($hexcodes, $pv) = split(/\s*;\s*/); |
12ac2576 |
1370 | $EAW{$pv} ||= Table->New(Is => "EastAsianWidth$pv"); |
adc6a8bd |
1371 | my ($start, $end) = split(/\.\./, $hexcodes); |
1372 | if (defined $end) { |
1373 | $EAW{$pv}->AppendRange(hex($start), hex($end)); |
1374 | } else { |
1375 | $EAW{$pv}->Append(hex($start)); |
1376 | } |
12ac2576 |
1377 | } |
1378 | close IN; |
1379 | |
12ac2576 |
1380 | |
1381 | for (keys %EAW) { |
1382 | $EAW{$_}->Write( |
d07a55ed |
1383 | ["lib","ea","$_.pl"], |
12ac2576 |
1384 | "EastAsianWidth category '$PropValueAlias{ea}{$_}'" |
1385 | ); |
1386 | } |
1387 | } |
1388 | |
1389 | ## |
1390 | ## Process HangulSyllableType.txt. |
1391 | ## |
1392 | sub HangulSyllableType_txt() |
1393 | { |
1394 | if (not open IN, "HangulSyllableType.txt") { |
1395 | die "$0: HangulSyllableType.txt: $!\n"; |
1396 | } |
1397 | |
1398 | my %HST; |
1399 | |
1400 | while (<IN>) |
1401 | { |
1402 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; |
1403 | my ($first, $last, $pv) = (hex($1), hex($2||""), $3); |
1404 | |
1405 | $HST{$pv} ||= Table->New(Is => "HangulSyllableType$pv"); |
1406 | $HST{$pv}->Append($first); |
1407 | |
1408 | if ($last) { $HST{$pv}->Extend($last) } |
1409 | } |
1410 | close IN; |
1411 | |
12ac2576 |
1412 | for (keys %HST) { |
1413 | $HST{$_}->Write( |
d07a55ed |
1414 | ["lib","hst","$_.pl"], |
12ac2576 |
1415 | "HangulSyllableType category '$PropValueAlias{hst}{$_}'" |
1416 | ); |
1417 | } |
d73e5302 |
1418 | } |
1419 | |
cf25bb62 |
1420 | ## |
1421 | ## Process Jamo.txt. |
1422 | ## |
1423 | sub Jamo_txt() |
1424 | { |
1425 | if (not open IN, "Jamo.txt") { |
1426 | die "$0: Jamo.txt: $!\n"; |
1427 | } |
1428 | my $Short = Table->New(); |
d73e5302 |
1429 | |
cf25bb62 |
1430 | while (<IN>) |
1431 | { |
1432 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; |
1433 | my ($code, $short) = (hex($1), $2); |
d73e5302 |
1434 | |
cf25bb62 |
1435 | $Short->Append($code, $short); |
d73e5302 |
1436 | } |
cf25bb62 |
1437 | close IN; |
12ac2576 |
1438 | # $Short->Write("JamoShort.pl"); |
d73e5302 |
1439 | } |
1440 | |
cf25bb62 |
1441 | ## |
1442 | ## Process Scripts.txt. |
1443 | ## |
1444 | sub Scripts_txt() |
1445 | { |
1446 | my @ScriptInfo; |
d73e5302 |
1447 | |
cf25bb62 |
1448 | if (not open(IN, "Scripts.txt")) { |
1449 | die "$0: Scripts.txt: $!\n"; |
1450 | } |
1451 | while (<IN>) { |
1452 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
1453 | |
cf25bb62 |
1454 | # Wait until all the scripts have been read since |
1455 | # they are not listed in numeric order. |
1456 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; |
1457 | } |
1458 | close IN; |
d73e5302 |
1459 | |
cf25bb62 |
1460 | # Now append the scripts properties in their code point order. |
d73e5302 |
1461 | |
cf25bb62 |
1462 | my %Script; |
1463 | my $Scripts = Table->New(); |
d73e5302 |
1464 | |
cf25bb62 |
1465 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) |
1466 | { |
1467 | my ($first, $last, $name) = @$script; |
1468 | $Scripts->Append($first, $name); |
d73e5302 |
1469 | |
99598c8c |
1470 | $Script{$name} ||= Table->New(Is => $name, |
1471 | Desc => "Script '$name'", |
1472 | Fuzzy => 1); |
cf25bb62 |
1473 | $Script{$name}->Append($first, $name); |
d73e5302 |
1474 | |
cf25bb62 |
1475 | if ($last) { |
1476 | $Scripts->Extend($last); |
1477 | $Script{$name}->Extend($last); |
1478 | } |
1479 | } |
d73e5302 |
1480 | |
12ac2576 |
1481 | # $Scripts->Write("Scripts.pl"); |
d73e5302 |
1482 | |
cf25bb62 |
1483 | ## Common is everything not explicitly assigned to a Script |
1484 | ## |
1485 | ## ***shouldn't this be intersected with \p{Assigned}? ****** |
1486 | ## |
99598c8c |
1487 | New_Prop(Is => 'Common', |
1488 | $Scripts->Invert, |
1489 | Desc => 'Pseudo-Script of codepoints not in other Unicode scripts', |
1490 | Fuzzy => 1); |
cf25bb62 |
1491 | } |
d73e5302 |
1492 | |
cf25bb62 |
1493 | ## |
1494 | ## Given a name like "Close Punctuation", return a regex (that when applied |
1495 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", |
1496 | ## "Close-Punctuation", etc.) |
1497 | ## |
1498 | ## Accept any space, dash, or underbar where in the official name there is |
1499 | ## space or a dash (or underbar, but there never is). |
1500 | ## |
1501 | ## |
1502 | sub NameToRegex($) |
1503 | { |
1504 | my $Name = shift; |
1505 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; |
1506 | return $Name; |
1507 | } |
d73e5302 |
1508 | |
cf25bb62 |
1509 | ## |
1510 | ## Process Blocks.txt. |
1511 | ## |
1512 | sub Blocks_txt() |
1513 | { |
1514 | my $Blocks = Table->New(); |
1515 | my %Blocks; |
d73e5302 |
1516 | |
cf25bb62 |
1517 | if (not open IN, "Blocks.txt") { |
1518 | die "$0: Blocks.txt: $!\n"; |
1519 | } |
d73e5302 |
1520 | |
cf25bb62 |
1521 | while (<IN>) |
1522 | { |
1523 | #next if not /Private Use$/; |
1524 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; |
d73e5302 |
1525 | |
cf25bb62 |
1526 | my ($first, $last, $name) = (hex($1), hex($2), $3); |
d73e5302 |
1527 | |
cf25bb62 |
1528 | $Blocks->Append($first, $name); |
76ccdbe2 |
1529 | |
99598c8c |
1530 | $Blocks{$name} ||= Table->New(In => $name, |
1531 | Desc => "Block '$name'", |
1532 | Fuzzy => 1); |
cf25bb62 |
1533 | $Blocks{$name}->Append($first, $name); |
76ccdbe2 |
1534 | |
cf25bb62 |
1535 | if ($last and $last != $first) { |
1536 | $Blocks->Extend($last); |
1537 | $Blocks{$name}->Extend($last); |
d73e5302 |
1538 | } |
d73e5302 |
1539 | } |
cf25bb62 |
1540 | close IN; |
1541 | |
12ac2576 |
1542 | # $Blocks->Write("Blocks.pl"); |
d73e5302 |
1543 | } |
1544 | |
cf25bb62 |
1545 | ## |
1546 | ## Read in the PropList.txt. It contains extended properties not |
551b6b6f |
1547 | ## listed in the UnicodeData.txt, such as 'Other_Alphabetic': |
cf25bb62 |
1548 | ## alphabetic but not of the general category L; many modifiers |
1549 | ## belong to this extended property category: while they are not |
1550 | ## alphabets, they are alphabetic in nature. |
1551 | ## |
1552 | sub PropList_txt() |
1553 | { |
1554 | my @PropInfo; |
1555 | |
1556 | if (not open IN, "PropList.txt") { |
1557 | die "$0: PropList.txt: $!\n"; |
1558 | } |
d73e5302 |
1559 | |
cf25bb62 |
1560 | while (<IN>) |
1561 | { |
1562 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; |
d73e5302 |
1563 | |
cf25bb62 |
1564 | # Wait until all the extended properties have been read since |
1565 | # they are not listed in numeric order. |
1566 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; |
1567 | } |
1568 | close IN; |
71d929cb |
1569 | |
cf25bb62 |
1570 | # Now append the extended properties in their code point order. |
1571 | my $Props = Table->New(); |
1572 | my %Prop; |
71d929cb |
1573 | |
cf25bb62 |
1574 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) |
1575 | { |
1576 | my ($first, $last, $name) = @$prop; |
1577 | $Props->Append($first, $name); |
71d929cb |
1578 | |
99598c8c |
1579 | $Prop{$name} ||= Table->New(Is => $name, |
1580 | Desc => "Extended property '$name'", |
1581 | Fuzzy => 1); |
cf25bb62 |
1582 | $Prop{$name}->Append($first, $name); |
71d929cb |
1583 | |
cf25bb62 |
1584 | if ($last) { |
1585 | $Props->Extend($last); |
1586 | $Prop{$name}->Extend($last); |
1587 | } |
71d929cb |
1588 | } |
d73e5302 |
1589 | |
12ac2576 |
1590 | for (keys %Prop) { |
1591 | (my $file = $PA_reverse{$_}) =~ tr/_//d; |
8e4b7420 |
1592 | # XXX I'm assuming that the names from %Prop don't suffer 8.3 clashes. |
1593 | $BaseNames{lc $file}++; |
12ac2576 |
1594 | $Prop{$_}->Write( |
d07a55ed |
1595 | ["lib","gc_sc","$file.pl"], |
12ac2576 |
1596 | "Binary property '$_'" |
1597 | ); |
1598 | } |
1599 | |
8158862b |
1600 | # Alphabetic is L, Nl, and Other_Alphabetic. |
99598c8c |
1601 | New_Prop(Is => 'Alphabetic', |
8158862b |
1602 | Table->Merge($Cat{L}, $Cat{Nl}, $Prop{Other_Alphabetic}), |
1603 | Desc => '[\p{L}\p{Nl}\p{OtherAlphabetic}]', # canonical names |
99598c8c |
1604 | Fuzzy => 1); |
cf25bb62 |
1605 | |
1606 | # Lowercase is Ll and Other_Lowercase. |
99598c8c |
1607 | New_Prop(Is => 'Lowercase', |
cf25bb62 |
1608 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), |
8158862b |
1609 | Desc => '[\p{Ll}\p{OtherLowercase}]', # canonical names |
99598c8c |
1610 | Fuzzy => 1); |
cf25bb62 |
1611 | |
1612 | # Uppercase is Lu and Other_Uppercase. |
1613 | New_Prop(Is => 'Uppercase', |
1614 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), |
8158862b |
1615 | Desc => '[\p{Lu}\p{OtherUppercase}]', # canonical names |
99598c8c |
1616 | Fuzzy => 1); |
cf25bb62 |
1617 | |
1618 | # Math is Sm and Other_Math. |
1619 | New_Prop(Is => 'Math', |
1620 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), |
8158862b |
1621 | Desc => '[\p{Sm}\p{OtherMath}]', # canonical names |
99598c8c |
1622 | Fuzzy => 1); |
cf25bb62 |
1623 | |
8158862b |
1624 | # ID_Start is Ll, Lu, Lt, Lm, Lo, Nl, and Other_ID_Start. |
cf25bb62 |
1625 | New_Prop(Is => 'ID_Start', |
8158862b |
1626 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}, $Prop{Other_ID_Start}), |
1627 | Desc => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}\p{OtherIDStart}]', |
99598c8c |
1628 | Fuzzy => 1); |
cf25bb62 |
1629 | |
8158862b |
1630 | # ID_Continue is ID_Start, Mn, Mc, Nd, Pc, and Other_ID_Continue. |
cf25bb62 |
1631 | New_Prop(Is => 'ID_Continue', |
8158862b |
1632 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}, |
1633 | @Prop{qw[Other_ID_Start Other_ID_Continue]}), |
1634 | Desc => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}\p{OtherIDContinue}]', |
99598c8c |
1635 | Fuzzy => 1); |
8158862b |
1636 | |
1637 | # Default_Ignorable_Code_Point = Other_Default_Ignorable_Code_Point |
1638 | # + Cf + Cc + Cs + Noncharacter + Variation_Selector |
1639 | # - WhiteSpace - FFF9..FFFB (Annotation Characters) |
1640 | |
1641 | my $Annotation = Table->New(); |
1642 | $Annotation->RawAppendRange(0xFFF9, 0xFFFB); |
1643 | |
1644 | New_Prop(Is => 'Default_Ignorable_Code_Point', |
1645 | Table->Merge(@Cat{qw[Cf Cc Cs]}, |
1646 | $Prop{Noncharacter_Code_Point}, |
1647 | $Prop{Variation_Selector}, |
1648 | $Prop{Other_Default_Ignorable_Code_Point}) |
1649 | ->Invert |
1650 | ->Merge($Prop{White_Space}, $Annotation) |
1651 | ->Invert, |
1652 | Desc => '(?![\p{WhiteSpace}\x{FFF9}-\x{FFFB}])[\p{Cf}\p{Cc}'. |
1653 | '\p{Cs}\p{NoncharacterCodePoint}\p{VariationSelector}'. |
1654 | '\p{OtherDefaultIgnorableCodePoint}]', |
1655 | Fuzzy => 1); |
1656 | |
d73e5302 |
1657 | } |
1658 | |
5beb625e |
1659 | |
1660 | ## |
1661 | ## These are used in: |
1662 | ## MakePropTestScript() |
1663 | ## WriteAllMappings() |
1664 | ## for making the test script. |
1665 | ## |
1666 | my %FuzzyNameToTest; |
1667 | my %ExactNameToTest; |
1668 | |
1669 | |
1670 | ## This used only for making the test script |
1671 | sub GenTests($$$$) |
1672 | { |
1673 | my $FH = shift; |
1674 | my $Prop = shift; |
1675 | my $MatchCode = shift; |
1676 | my $FailCode = shift; |
1677 | |
1678 | if (defined $MatchCode) { |
1679 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode; |
1680 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode; |
1681 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode; |
1682 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode; |
1683 | } |
1684 | if (defined $FailCode) { |
1685 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode; |
1686 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode; |
1687 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode; |
1688 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode; |
1689 | } |
1690 | } |
1691 | |
1692 | ## This used only for making the test script |
1693 | sub ExpectError($$) |
1694 | { |
1695 | my $FH = shift; |
1696 | my $prop = shift; |
1697 | |
1698 | print $FH qq/Error('\\p{$prop}');\n/; |
1699 | print $FH qq/Error('\\P{$prop}');\n/; |
1700 | } |
1701 | |
1702 | ## This used only for making the test script |
1703 | my @GoodSeps = ( |
1704 | " ", |
1705 | "-", |
1706 | " \t ", |
1707 | "", |
1708 | "", |
1709 | "_", |
1710 | ); |
1711 | my @BadSeps = ( |
1712 | "--", |
1713 | "__", |
1714 | " _", |
1715 | "/" |
1716 | ); |
1717 | |
1718 | ## This used only for making the test script |
1719 | sub RandomlyFuzzifyName($;$) |
1720 | { |
1721 | my $Name = shift; |
1722 | my $WantError = shift; ## if true, make an error |
1723 | |
1724 | my @parts; |
1725 | for my $part (split /[-\s_]+/, $Name) |
1726 | { |
1727 | if (@parts) { |
1728 | if ($WantError and rand() < 0.3) { |
1729 | push @parts, $BadSeps[rand(@BadSeps)]; |
1730 | $WantError = 0; |
1731 | } else { |
1732 | push @parts, $GoodSeps[rand(@GoodSeps)]; |
1733 | } |
1734 | } |
1735 | my $switch = int rand(4); |
1736 | if ($switch == 0) { |
1737 | push @parts, uc $part; |
1738 | } elsif ($switch == 1) { |
1739 | push @parts, lc $part; |
1740 | } elsif ($switch == 2) { |
1741 | push @parts, ucfirst $part; |
1742 | } else { |
1743 | push @parts, $part; |
1744 | } |
1745 | } |
1746 | my $new = join('', @parts); |
1747 | |
1748 | if ($WantError) { |
1749 | if (rand() >= 0.5) { |
1750 | $new .= $BadSeps[rand(@BadSeps)]; |
1751 | } else { |
1752 | $new = $BadSeps[rand(@BadSeps)] . $new; |
1753 | } |
1754 | } |
1755 | return $new; |
1756 | } |
1757 | |
1758 | ## This used only for making the test script |
1759 | sub MakePropTestScript() |
1760 | { |
1761 | ## this written directly -- it's huge. |
d07a55ed |
1762 | force_unlink ("TestProp.pl"); |
5beb625e |
1763 | if (not open OUT, ">TestProp.pl") { |
1764 | die "$0: TestProp.pl: $!\n"; |
1765 | } |
1766 | print OUT <DATA>; |
1767 | |
1768 | while (my ($Name, $Table) = each %ExactNameToTest) |
1769 | { |
1770 | GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode); |
1771 | ExpectError(*OUT, uc $Name) if uc $Name ne $Name; |
1772 | ExpectError(*OUT, lc $Name) if lc $Name ne $Name; |
1773 | } |
1774 | |
1775 | |
1776 | while (my ($Name, $Table) = each %FuzzyNameToTest) |
1777 | { |
1778 | my $Orig = $CanonicalToOrig{$Name}; |
1779 | my %Names = ( |
1780 | $Name => 1, |
1781 | $Orig => 1, |
1782 | RandomlyFuzzifyName($Orig) => 1 |
1783 | ); |
1784 | |
1785 | for my $N (keys %Names) { |
1786 | GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode); |
1787 | } |
1788 | |
1789 | ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR')); |
1790 | } |
1791 | |
1792 | print OUT "Finished();\n"; |
1793 | close OUT; |
1794 | } |
1795 | |
1796 | |
1797 | ## |
1798 | ## These are used only in: |
1799 | ## RegisterFileForName() |
1800 | ## WriteAllMappings() |
1801 | ## |
1802 | my %Exact; ## will become %utf8::Exact; |
1803 | my %Canonical; ## will become %utf8::Canonical; |
1804 | my %CaComment; ## Comment for %Canonical entry of same key |
1805 | |
1806 | ## |
1807 | ## Given info about a name and a datafile that it should be associated with, |
1808 | ## register that assocation in %Exact and %Canonical. |
1809 | sub RegisterFileForName($$$$) |
1810 | { |
1811 | my $Type = shift; |
1812 | my $Name = shift; |
1813 | my $IsFuzzy = shift; |
1814 | my $filename = shift; |
1815 | |
1816 | ## |
1817 | ## Now in details for the mapping. $Type eq 'Is' has the |
1818 | ## Is removed, as it will be removed in utf8_heavy when this |
1819 | ## data is being checked. In keeps its "In", but a second |
1820 | ## sans-In record is written if it doesn't conflict with |
1821 | ## anything already there. |
1822 | ## |
1823 | if (not $IsFuzzy) |
1824 | { |
1825 | if ($Type eq 'Is') { |
1826 | die "oops[$Name]" if $Exact{$Name}; |
1827 | $Exact{$Name} = $filename; |
1828 | } else { |
1829 | die "oops[$Type$Name]" if $Exact{"$Type$Name"}; |
1830 | $Exact{"$Type$Name"} = $filename; |
1831 | $Exact{$Name} = $filename if not $Exact{$Name}; |
1832 | } |
1833 | } |
1834 | else |
1835 | { |
1836 | my $CName = lc $Name; |
1837 | if ($Type eq 'Is') { |
1838 | die "oops[$CName]" if $Canonical{$CName}; |
1839 | $Canonical{$CName} = $filename; |
1840 | $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2; |
1841 | } else { |
1842 | die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"}; |
1843 | $Canonical{lc "$Type$CName"} = $filename; |
1844 | $CaComment{lc "$Type$CName"} = "$Type$Name"; |
1845 | if (not $Canonical{$CName}) { |
1846 | $Canonical{$CName} = $filename; |
1847 | $CaComment{$CName} = "$Type$Name"; |
1848 | } |
1849 | } |
1850 | } |
1851 | } |
1852 | |
cf25bb62 |
1853 | ## |
1854 | ## Writes the info accumulated in |
1855 | ## |
1856 | ## %TableInfo; |
1857 | ## %FuzzyNames; |
1858 | ## %AliasInfo; |
1859 | ## |
1860 | ## |
1861 | sub WriteAllMappings() |
1862 | { |
99598c8c |
1863 | my @MAP; |
1864 | |
5beb625e |
1865 | ## 'Is' *MUST* come first, so its names have precidence over 'In's |
1866 | for my $Type ('Is', 'In') |
1867 | { |
1868 | my %RawNameToFile; ## a per-$Type cache |
cf25bb62 |
1869 | |
5beb625e |
1870 | for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}}) |
cf25bb62 |
1871 | { |
5beb625e |
1872 | ## Note: $Name is already canonical |
99598c8c |
1873 | my $Table = $TableInfo{$Type}->{$Name}; |
5beb625e |
1874 | my $IsFuzzy = $FuzzyNames{$Type}->{$Name}; |
99598c8c |
1875 | |
1876 | ## Need an 8.3 safe filename (which means "an 8 safe" $filename) |
5beb625e |
1877 | my $filename; |
cf25bb62 |
1878 | { |
5beb625e |
1879 | ## 'Is' items lose 'Is' from the basename. |
12ac2576 |
1880 | $filename = $Type eq 'Is' ? |
1881 | ($PVA_reverse{sc}{$Name} || $Name) : |
1882 | "$Type$Name"; |
5beb625e |
1883 | |
1884 | $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_" |
1885 | substr($filename, 8) = '' if length($filename) > 8; |
1886 | |
1887 | ## |
1888 | ## Make sure the basename doesn't conflict with something we |
1889 | ## might have already written. If we have, say, |
1890 | ## InGreekExtended1 |
1891 | ## InGreekExtended2 |
1892 | ## they become |
1893 | ## InGreekE |
1894 | ## InGreek2 |
1895 | ## |
1896 | while (my $num = $BaseNames{lc $filename}++) |
1897 | { |
1898 | $num++; ## so basenames with numbers start with '2', which |
1899 | ## just looks more natural. |
1900 | ## Want to append $num, but if it'll make the basename longer |
1901 | ## than 8 characters, pre-truncate $filename so that the result |
1902 | ## is acceptable. |
1903 | my $delta = length($filename) + length($num) - 8; |
1904 | if ($delta > 0) { |
1905 | substr($filename, -$delta) = $num; |
1906 | } else { |
1907 | $filename .= $num; |
1908 | } |
99598c8c |
1909 | } |
5beb625e |
1910 | }; |
99598c8c |
1911 | |
1912 | ## |
1913 | ## Construct a nice comment to add to the file, and build data |
1914 | ## for the "./Properties" file along the way. |
1915 | ## |
1916 | my $Comment; |
1917 | { |
1918 | my $Desc = $TableDesc{$Type}->{$Name} || ""; |
1919 | ## get list of names this table is reference by |
1920 | my @Supported = $Name; |
1921 | while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} }) |
1922 | { |
1923 | if ($Orig eq $Name) { |
1924 | push @Supported, $Alias; |
1925 | } |
1926 | } |
1927 | |
1928 | my $TypeToShow = $Type eq 'Is' ? "" : $Type; |
1929 | my $OrigProp; |
1930 | |
1931 | $Comment = "This file supports:\n"; |
1932 | for my $N (@Supported) |
1933 | { |
1934 | my $IsFuzzy = $FuzzyNames{$Type}->{$N}; |
5beb625e |
1935 | my $Prop = "\\p{$TypeToShow$Name}"; |
99598c8c |
1936 | $OrigProp = $Prop if not $OrigProp; #cache for aliases |
1937 | if ($IsFuzzy) { |
1938 | $Comment .= "\t$Prop (and fuzzy permutations)\n"; |
1939 | } else { |
1940 | $Comment .= "\t$Prop\n"; |
1941 | } |
1942 | my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)"; |
1943 | |
1944 | push @MAP, sprintf("%s %-42s %s\n", |
1945 | $IsFuzzy ? '*' : ' ', $Prop, $MyDesc); |
1946 | } |
1947 | if ($Desc) { |
1948 | $Comment .= "\nMeaning: $Desc\n"; |
1949 | } |
1950 | |
1951 | } |
cf25bb62 |
1952 | ## |
1953 | ## Okay, write the file... |
1954 | ## |
d07a55ed |
1955 | $Table->Write(["lib","gc_sc","$filename.pl"], $Comment); |
99598c8c |
1956 | |
5beb625e |
1957 | ## and register it |
1958 | $RawNameToFile{$Name} = $filename; |
1959 | RegisterFileForName($Type => $Name, $IsFuzzy, $filename); |
cf25bb62 |
1960 | |
5beb625e |
1961 | if ($IsFuzzy) |
1962 | { |
1963 | my $CName = CanonicalName($Type . '_'. $Name); |
1964 | $FuzzyNameToTest{$Name} = $Table if !$FuzzyNameToTest{$Name}; |
1965 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1966 | } else { |
1967 | $ExactNameToTest{$Name} = $Table; |
cf25bb62 |
1968 | } |
1969 | |
cf25bb62 |
1970 | } |
1971 | |
5beb625e |
1972 | ## Register aliase info |
1973 | for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}}) |
cf25bb62 |
1974 | { |
5beb625e |
1975 | my $Alias = $AliasInfo{$Type}->{$Name}; |
1976 | my $IsFuzzy = $FuzzyNames{$Type}->{$Alias}; |
1977 | my $filename = $RawNameToFile{$Name}; |
1978 | die "oops [$Alias]->[$Name]" if not $filename; |
1979 | RegisterFileForName($Type => $Alias, $IsFuzzy, $filename); |
1980 | |
1981 | my $Table = $TableInfo{$Type}->{$Name}; |
1982 | die "oops" if not $Table; |
1983 | if ($IsFuzzy) |
1984 | { |
1985 | my $CName = CanonicalName($Type .'_'. $Alias); |
1986 | $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias}; |
1987 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; |
1988 | } else { |
1989 | $ExactNameToTest{$Alias} = $Table; |
1990 | } |
cf25bb62 |
1991 | } |
5beb625e |
1992 | } |
cf25bb62 |
1993 | |
5beb625e |
1994 | ## |
1995 | ## Write out the property list |
1996 | ## |
1997 | { |
1998 | my @OUT = ( |
1999 | "##\n", |
2000 | "## This file created by $0\n", |
2001 | "## List of built-in \\p{...}/\\P{...} properties.\n", |
2002 | "##\n", |
2003 | "## '*' means name may be 'fuzzy'\n", |
2004 | "##\n\n", |
2005 | sort { substr($a,2) cmp substr($b, 2) } @MAP, |
2006 | ); |
2007 | WriteIfChanged('Properties', @OUT); |
2008 | } |
cf25bb62 |
2009 | |
5beb625e |
2010 | use Text::Tabs (); ## using this makes the files about half the size |
2011 | |
2012 | ## Write Exact.pl |
2013 | { |
2014 | my @OUT = ( |
2015 | $HEADER, |
2016 | "##\n", |
2017 | "## Data in this file used by ../utf8_heavy.pl\n", |
2018 | "##\n\n", |
12ac2576 |
2019 | "## Mapping from name to filename in ./lib/gc_sc\n", |
5beb625e |
2020 | "%utf8::Exact = (\n", |
2021 | ); |
cf25bb62 |
2022 | |
12ac2576 |
2023 | $Exact{InGreek} = 'InGreekA'; # this is evil kludge |
cf25bb62 |
2024 | for my $Name (sort keys %Exact) |
2025 | { |
2026 | my $File = $Exact{$Name}; |
5beb625e |
2027 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
2028 | my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/); |
2029 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
2030 | } |
5beb625e |
2031 | push @OUT, ");\n1;\n"; |
2032 | |
2033 | WriteIfChanged('Exact.pl', @OUT); |
2034 | } |
cf25bb62 |
2035 | |
5beb625e |
2036 | ## Write Canonical.pl |
2037 | { |
2038 | my @OUT = ( |
2039 | $HEADER, |
2040 | "##\n", |
2041 | "## Data in this file used by ../utf8_heavy.pl\n", |
2042 | "##\n\n", |
2043 | "## Mapping from lc(canonical name) to filename in ./lib\n", |
2044 | "%utf8::Canonical = (\n", |
2045 | ); |
2046 | my $Trail = ""; ## used just to keep the spacing pretty |
2047 | for my $Name (sort keys %Canonical) |
cf25bb62 |
2048 | { |
5beb625e |
2049 | my $File = $Canonical{$Name}; |
2050 | if ($CaComment{$Name}) { |
2051 | push @OUT, "\n" if not $Trail; |
2052 | push @OUT, " # $CaComment{$Name}\n"; |
2053 | $Trail = "\n"; |
2054 | } else { |
2055 | $Trail = ""; |
cf25bb62 |
2056 | } |
5beb625e |
2057 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
2058 | my $Text = sprintf(" %-41s => %s,\n$Trail", $Name, qq/'$File'/); |
2059 | push @OUT, Text::Tabs::unexpand($Text); |
cf25bb62 |
2060 | } |
5beb625e |
2061 | push @OUT, ");\n1\n"; |
2062 | WriteIfChanged('Canonical.pl', @OUT); |
d2d499f5 |
2063 | } |
5beb625e |
2064 | |
2065 | MakePropTestScript() if $MakeTestScript; |
d2d499f5 |
2066 | } |
2067 | |
5beb625e |
2068 | |
551b6b6f |
2069 | sub SpecialCasing_txt() |
cf25bb62 |
2070 | { |
2071 | # |
2072 | # Read in the special cases. |
2073 | # |
983ffd37 |
2074 | |
cf25bb62 |
2075 | my %CaseInfo; |
2076 | |
551b6b6f |
2077 | if (not open IN, "SpecialCasing.txt") { |
2078 | die "$0: SpecialCasing.txt: $!\n"; |
cf25bb62 |
2079 | } |
2080 | while (<IN>) { |
2081 | next unless /^[0-9A-Fa-f]+;/; |
2082 | s/\#.*//; |
2083 | s/\s+$//; |
2084 | |
2085 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); |
2086 | |
2087 | if ($condition) { # not implemented yet |
2088 | print "# SKIPPING $_\n" if $Verbose; |
2089 | next; |
2090 | } |
2091 | |
2092 | # Wait until all the special cases have been read since |
2093 | # they are not listed in numeric order. |
2094 | my $ix = hex($code); |
a6da3000 |
2095 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ] |
2096 | unless $code eq $lower; |
2097 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ] |
2098 | unless $code eq $title; |
2099 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ] |
2100 | unless $code eq $upper; |
cf25bb62 |
2101 | } |
2102 | close IN; |
2103 | |
2104 | # Now write out the special cases properties in their code point order. |
2105 | # Prepend them to the To/{Upper,Lower,Title}.pl. |
2106 | |
2107 | for my $case (qw(Lower Title Upper)) |
2108 | { |
2109 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; |
cf25bb62 |
2110 | |
b08cf34e |
2111 | my @OUT = |
2112 | ( |
b6922eda |
2113 | $HEADER, $INTERNAL_ONLY, "\n", |
2114 | "# The key: UTF-8 _bytes_, the value: UTF-8 (speed hack)\n", |
b08cf34e |
2115 | "%utf8::ToSpec$case =\n(\n", |
2116 | ); |
cf25bb62 |
2117 | |
2118 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { |
2119 | my ($ix, $code, $to) = @$prop; |
2120 | my $tostr = |
2121 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; |
b08cf34e |
2122 | push @OUT, sprintf qq["%s" => "$tostr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $ix))); |
5cb851a6 |
2123 | # Remove any single-character mappings for |
2124 | # the same character since we are going for |
2125 | # the special casing rules. |
2126 | $NormalCase =~ s/^$code\t\t\w+\n//m; |
cf25bb62 |
2127 | } |
5beb625e |
2128 | push @OUT, ( |
2129 | ");\n\n", |
2130 | "return <<'END';\n", |
2131 | $NormalCase, |
2132 | "END\n" |
2133 | ); |
d07a55ed |
2134 | WriteIfChanged(["To","$case.pl"], @OUT); |
d2d499f5 |
2135 | } |
d2d499f5 |
2136 | } |
2137 | |
c4051cc5 |
2138 | # |
2139 | # Read in the case foldings. |
2140 | # |
ae6979a8 |
2141 | # We will do full case folding, C + F + I (see CaseFolding.txt). Note that |
2142 | # there are no I entries starting with Unicode 3.2, but leaving it in allows |
2143 | # for backward compatibility. |
c4051cc5 |
2144 | # |
551b6b6f |
2145 | sub CaseFolding_txt() |
cf25bb62 |
2146 | { |
551b6b6f |
2147 | if (not open IN, "CaseFolding.txt") { |
2148 | die "$0: CaseFolding.txt: $!\n"; |
cf25bb62 |
2149 | } |
c4051cc5 |
2150 | |
cf25bb62 |
2151 | my $Fold = Table->New(); |
c4051cc5 |
2152 | my %Fold; |
2153 | |
cf25bb62 |
2154 | while (<IN>) { |
254ba52a |
2155 | # Skip status 'S', simple case folding |
c4051cc5 |
2156 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; |
2157 | |
cf25bb62 |
2158 | my ($code, $status, $fold) = (hex($1), $2, $3); |
c4051cc5 |
2159 | |
2160 | if ($status eq 'C') { # Common: one-to-one folding |
254ba52a |
2161 | # No append() since several codes may fold into one. |
cf25bb62 |
2162 | $Fold->RawAppendRange($code, $code, $fold); |
c4051cc5 |
2163 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I |
cf25bb62 |
2164 | $Fold{$code} = $fold; |
c4051cc5 |
2165 | } |
2166 | } |
cf25bb62 |
2167 | close IN; |
c4051cc5 |
2168 | |
cf25bb62 |
2169 | $Fold->Write("To/Fold.pl"); |
c4051cc5 |
2170 | |
2171 | # |
2172 | # Prepend the special foldings to the common foldings. |
2173 | # |
c4051cc5 |
2174 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; |
5beb625e |
2175 | |
b08cf34e |
2176 | my @OUT = |
2177 | ( |
b6922eda |
2178 | $HEADER, $INTERNAL_ONLY, "\n", |
2179 | "# The key: UTF-8 _bytes_, the value: UTF-8 (speed hack)\n", |
b08cf34e |
2180 | "%utf8::ToSpecFold =\n(\n", |
2181 | ); |
cf25bb62 |
2182 | for my $code (sort { $a <=> $b } keys %Fold) { |
2183 | my $foldstr = |
2184 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; |
b08cf34e |
2185 | push @OUT, sprintf qq["%s" => "$foldstr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $code))); |
c4051cc5 |
2186 | } |
5beb625e |
2187 | push @OUT, ( |
2188 | ");\n\n", |
2189 | "return <<'END';\n", |
2190 | $CommonFold, |
2191 | "END\n", |
2192 | ); |
2193 | |
d07a55ed |
2194 | WriteIfChanged(["To","Fold.pl"], @OUT); |
c4051cc5 |
2195 | } |
2196 | |
cf25bb62 |
2197 | ## Do it.... |
2198 | |
12ac2576 |
2199 | Build_Aliases(); |
44da8cae |
2200 | UnicodeData_Txt(); |
cf25bb62 |
2201 | PropList_txt(); |
2202 | |
2203 | Scripts_txt(); |
2204 | Blocks_txt(); |
2205 | |
5beb625e |
2206 | WriteAllMappings(); |
2207 | |
551b6b6f |
2208 | LineBreak_Txt(); |
2209 | ArabicShaping_txt(); |
12ac2576 |
2210 | EastAsianWidth_txt(); |
2211 | HangulSyllableType_txt(); |
cf25bb62 |
2212 | Jamo_txt(); |
551b6b6f |
2213 | SpecialCasing_txt(); |
2214 | CaseFolding_txt(); |
cf25bb62 |
2215 | |
97050450 |
2216 | if ( $FileList and $MakeList ) { |
2217 | |
2218 | print "Updating '$FileList'\n" |
2219 | if ($Verbose); |
2220 | |
2221 | open my $ofh,">",$FileList |
2222 | or die "Can't write to '$FileList':$!"; |
2223 | print $ofh <<"EOFHEADER"; |
2224 | # |
2225 | # mktables.lst -- File list for mktables. |
2226 | # |
2227 | # Autogenerated on @{[scalar localtime]} |
2228 | # |
2229 | # - First section is input files |
2230 | # (mktables itself is automatically included) |
2231 | # - Section seperator is /^=+\$/ |
2232 | # - Second section is a list of output files. |
2233 | # - Lines matching /^\\s*#/ are treated as comments |
2234 | # which along with blank lines are ignored. |
2235 | # |
2236 | |
2237 | # Input files: |
2238 | |
2239 | EOFHEADER |
2240 | my @input=("version",glob('*.txt')); |
2241 | print $ofh "$_\n" for |
ae6979a8 |
2242 | sort(@input), |
97050450 |
2243 | "\n=================================\n", |
2244 | "# Output files:\n", |
2245 | # special files |
2246 | "Properties"; |
2247 | |
2248 | |
2249 | require File::Find; |
ae6979a8 |
2250 | my @output_files; |
97050450 |
2251 | File::Find::find({ |
2252 | no_chdir=>1, |
2253 | wanted=>sub { |
2254 | if (/\.pl$/) { |
2255 | s!^\./!!; |
ae6979a8 |
2256 | push @output_files, "$_\n"; |
97050450 |
2257 | } |
2258 | }, |
2259 | },"."); |
2260 | |
ae6979a8 |
2261 | print $ofh sort @output_files; |
97050450 |
2262 | print $ofh "\n# ",scalar(@input)," input files\n", |
ae6979a8 |
2263 | "# ",scalar(@output_files)+1," output files\n\n", |
97050450 |
2264 | "# End list\n"; |
2265 | close $ofh |
2266 | or warn "Failed to close $ofh: $!"; |
2267 | |
2268 | print "Filelist has ",scalar(@input)," input files and ", |
ae6979a8 |
2269 | scalar(@output_files)+1," output files\n" |
97050450 |
2270 | if $Verbose; |
2271 | } |
2272 | print "All done\n" if $Verbose; |
5beb625e |
2273 | exit(0); |
cf25bb62 |
2274 | |
5beb625e |
2275 | ## TRAILING CODE IS USED BY MakePropTestScript() |
2276 | __DATA__ |
2277 | use strict; |
2278 | use warnings; |
2279 | |
2280 | my $Tests = 0; |
2281 | my $Fails = 0; |
cf25bb62 |
2282 | |
5beb625e |
2283 | sub Expect($$$) |
2284 | { |
2285 | my $Expect = shift; |
2286 | my $String = shift; |
2287 | my $Regex = shift; |
2288 | my $Line = (caller)[2]; |
2289 | |
2290 | $Tests++; |
2291 | my $RegObj; |
2292 | my $result = eval { |
2293 | $RegObj = qr/$Regex/; |
2294 | $String =~ $RegObj ? 1 : 0 |
2295 | }; |
2296 | |
2297 | if (not defined $result) { |
2298 | print "couldn't compile /$Regex/ on $0 line $Line: $@\n"; |
2299 | $Fails++; |
2300 | } elsif ($result ^ $Expect) { |
2301 | print "bad result (expected $Expect) on $0 line $Line: $@\n"; |
2302 | $Fails++; |
2303 | } |
2304 | } |
d73e5302 |
2305 | |
5beb625e |
2306 | sub Error($) |
2307 | { |
2308 | my $Regex = shift; |
2309 | $Tests++; |
2310 | if (eval { 'x' =~ qr/$Regex/; 1 }) { |
2311 | $Fails++; |
2312 | my $Line = (caller)[2]; |
2313 | print "expected error for /$Regex/ on $0 line $Line: $@\n"; |
2314 | } |
2315 | } |
2316 | |
2317 | sub Finished() |
2318 | { |
2319 | if ($Fails == 0) { |
2320 | print "All $Tests tests passed.\n"; |
2321 | exit(0); |
2322 | } else { |
2323 | print "$Tests tests, $Fails failed!\n"; |
2324 | exit(-1); |
2325 | } |
2326 | } |