Commit | Line | Data |
---|---|---|
d73e5302 | 1 | #!/usr/bin/perl -w |
d73e5302 | 2 | use strict; |
cf25bb62 | 3 | use Carp; |
5beb625e | 4 | |
c65e4d19 JH |
5 | die "$0: Please run me as ./mktables to avoid unnecessary differences\n" |
6 | unless $0 eq "./mktables"; | |
7 | ||
cf25bb62 JH |
8 | ## |
9 | ## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl) | |
10 | ## from the Unicode database files (lib/unicore/*.txt). | |
11 | ## | |
d2d499f5 | 12 | |
5beb625e JH |
13 | mkdir("lib", 0755); |
14 | mkdir("To", 0755); | |
d73e5302 | 15 | |
cf25bb62 JH |
16 | ## |
17 | ## Process any args. | |
18 | ## | |
5beb625e JH |
19 | my $Verbose = 0; |
20 | my $MakeTestScript = 0; | |
cf25bb62 JH |
21 | |
22 | while (@ARGV) | |
23 | { | |
24 | my $arg = shift @ARGV; | |
25 | if ($arg eq '-v') { | |
26 | $Verbose = 1; | |
27 | } elsif ($arg eq '-q') { | |
28 | $Verbose = 0; | |
5beb625e JH |
29 | } elsif ($arg eq '-maketest') { |
30 | $MakeTestScript = 1; | |
d73e5302 | 31 | } else { |
5beb625e | 32 | die "usage: $0 [-v|-q] [-maketest]"; |
d73e5302 JH |
33 | } |
34 | } | |
35 | ||
cf25bb62 | 36 | my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 3.1.1. |
d73e5302 | 37 | |
cf25bb62 | 38 | my $HEADER=<<"EOF"; |
d73e5302 | 39 | # !!!!!!! DO NOT EDIT THIS FILE !!!!!!! |
551b6b6f | 40 | # This file is built by $0 from e.g. UnicodeData.txt. |
d73e5302 | 41 | # Any changes made here will be lost! |
cf25bb62 JH |
42 | |
43 | EOF | |
44 | ||
5beb625e JH |
45 | |
46 | ## | |
47 | ## Given a filename and a reference to an array of lines, | |
48 | ## write the lines to the file only if the contents have not changed. | |
49 | ## | |
50 | sub WriteIfChanged($\@) | |
51 | { | |
52 | my $file = shift; | |
53 | my $lines = shift; | |
54 | ||
55 | my $TextToWrite = join '', @$lines; | |
56 | if (open IN, $file) { | |
57 | local($/) = undef; | |
58 | my $PreviousText = <IN>; | |
59 | close IN; | |
60 | if ($PreviousText eq $TextToWrite) { | |
61 | print "$file unchanged.\n" if $Verbose; | |
62 | return; | |
63 | } | |
64 | } | |
65 | if (not open OUT, ">$file") { | |
66 | die "$0: can't open $file for output: $!\n"; | |
67 | } | |
68 | print "$file written.\n" if $Verbose; | |
69 | ||
70 | print OUT $TextToWrite; | |
71 | close OUT; | |
72 | } | |
73 | ||
cf25bb62 JH |
74 | ## |
75 | ## The main datastructure (a "Table") represents a set of code points that | |
76 | ## are part of a particular quality (that are part of \pL, \p{InGreek}, | |
77 | ## etc.). They are kept as ranges of code points (starting and ending of | |
78 | ## each range). | |
79 | ## | |
80 | ## For example, a range ASCII LETTERS would be represented as: | |
81 | ## [ [ 0x41 => 0x5A, 'UPPER' ], | |
82 | ## [ 0x61 => 0x7A, 'LOWER, ] ] | |
83 | ## | |
84 | sub RANGE_START() { 0 } ## index into range element | |
85 | sub RANGE_END() { 1 } ## index into range element | |
86 | sub RANGE_NAME() { 2 } ## index into range element | |
87 | ||
99598c8c | 88 | ## Conceptually, these should really be folded into the 'Table' objects |
cf25bb62 | 89 | my %TableInfo; |
99598c8c | 90 | my %TableDesc; |
cf25bb62 JH |
91 | my %FuzzyNames; |
92 | my %AliasInfo; | |
5beb625e | 93 | my %CanonicalToOrig; |
cf25bb62 JH |
94 | |
95 | ## | |
99598c8c JH |
96 | ## Turn something like |
97 | ## OLD-ITALIC | |
5beb625e | 98 | ## into |
99598c8c JH |
99 | ## OldItalic |
100 | ## | |
101 | sub CanonicalName($) | |
102 | { | |
5beb625e JH |
103 | my $orig = shift; |
104 | my $name = lc $orig; | |
99598c8c | 105 | $name =~ s/(?<![a-z])(\w)/\u$1/g; |
5beb625e | 106 | $name =~ s/[-_\s]+//g; |
99598c8c | 107 | |
5beb625e | 108 | $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name}; |
99598c8c JH |
109 | return $name; |
110 | } | |
111 | ||
99598c8c | 112 | ## |
cf25bb62 JH |
113 | ## Associates a property ("Greek", "Lu", "Assigned",...) with a Table. |
114 | ## | |
115 | ## Called like: | |
99598c8c | 116 | ## New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1); |
cf25bb62 JH |
117 | ## |
118 | ## Normally, these parameters are set when the Table is created (when the | |
119 | ## Table->New constructor is called), but there are times when it needs to | |
120 | ## be done after-the-fact...) | |
121 | ## | |
122 | sub New_Prop($$$@) | |
123 | { | |
124 | my $Type = shift; ## "Is" or "In"; | |
125 | my $Name = shift; | |
126 | my $Table = shift; | |
127 | ||
128 | ## remaining args are optional key/val | |
129 | my %Args = @_; | |
130 | ||
99598c8c JH |
131 | my $Fuzzy = delete $Args{Fuzzy}; |
132 | my $Desc = delete $Args{Desc}; # description | |
133 | ||
5beb625e | 134 | $Name = CanonicalName($Name) if $Fuzzy; |
cf25bb62 JH |
135 | |
136 | ## sanity check a few args | |
137 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) { | |
138 | confess "$0: bad args to New_Prop" | |
139 | } | |
140 | ||
141 | if (not $TableInfo{$Type}->{$Name}) | |
142 | { | |
143 | $TableInfo{$Type}->{$Name} = $Table; | |
99598c8c JH |
144 | $TableDesc{$Type}->{$Name} = $Desc; |
145 | if ($Fuzzy) { | |
cf25bb62 JH |
146 | $FuzzyNames{$Type}->{$Name} = $Name; |
147 | } | |
148 | } | |
d73e5302 JH |
149 | } |
150 | ||
d73e5302 | 151 | |
cf25bb62 JH |
152 | ## |
153 | ## Creates a new Table object. | |
154 | ## | |
155 | ## Args are key/value pairs: | |
99598c8c JH |
156 | ## In => Name -- Name of "In" property to be associated with |
157 | ## Is => Name -- Name of "Is" property to be associated with | |
158 | ## Fuzzy => Boolean -- True if name can be accessed "fuzzily" | |
159 | ## Desc => String -- Description of the property | |
cf25bb62 JH |
160 | ## |
161 | ## No args are required. | |
162 | ## | |
163 | sub Table::New | |
164 | { | |
165 | my $class = shift; | |
166 | my %Args = @_; | |
167 | ||
168 | my $Table = bless [], $class; | |
169 | ||
99598c8c JH |
170 | my $Fuzzy = delete $Args{Fuzzy}; |
171 | my $Desc = delete $Args{Desc}; | |
cf25bb62 JH |
172 | |
173 | for my $Type ('Is', 'In') | |
174 | { | |
175 | if (my $Name = delete $Args{$Type}) { | |
99598c8c | 176 | New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy); |
cf25bb62 JH |
177 | } |
178 | } | |
179 | ||
180 | ## shouldn't have any left over | |
181 | if (%Args) { | |
182 | confess "$0: bad args to Table->New" | |
183 | } | |
184 | ||
185 | return $Table; | |
d73e5302 JH |
186 | } |
187 | ||
cf25bb62 JH |
188 | ## |
189 | ## Returns true if the Table has no code points | |
190 | ## | |
191 | sub Table::IsEmpty | |
192 | { | |
193 | my $Table = shift; #self | |
194 | return not @$Table; | |
d73e5302 JH |
195 | } |
196 | ||
cf25bb62 JH |
197 | ## |
198 | ## Returns true if the Table has code points | |
199 | ## | |
200 | sub Table::NotEmpty | |
201 | { | |
202 | my $Table = shift; #self | |
203 | return @$Table; | |
d73e5302 JH |
204 | } |
205 | ||
cf25bb62 JH |
206 | ## |
207 | ## Returns the maximum code point currently in the table. | |
208 | ## | |
209 | sub Table::Max | |
210 | { | |
211 | my $Table = shift; #self | |
212 | confess "oops" if $Table->IsEmpty; ## must have code points to have a max | |
213 | return $Table->[-1]->[RANGE_END]; | |
214 | } | |
d73e5302 | 215 | |
cf25bb62 JH |
216 | ## |
217 | ## Replaces the codepoints in the Table with those in the Table given | |
218 | ## as an arg. (NOTE: this is not a "deep copy"). | |
219 | ## | |
220 | sub Table::Replace($$) | |
221 | { | |
222 | my $Table = shift; #self | |
223 | my $New = shift; | |
d73e5302 | 224 | |
cf25bb62 JH |
225 | @$Table = @$New; |
226 | } | |
71d929cb | 227 | |
cf25bb62 JH |
228 | ## |
229 | ## Given a new code point, make the last range of the Table extend to | |
230 | ## include the new (and all intervening) code points. | |
231 | ## | |
232 | sub Table::Extend | |
233 | { | |
234 | my $Table = shift; #self | |
235 | my $codepoint = shift; | |
d73e5302 | 236 | |
cf25bb62 | 237 | my $PrevMax = $Table->Max; |
e904f995 | 238 | |
cf25bb62 | 239 | confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax; |
e904f995 | 240 | |
cf25bb62 JH |
241 | $Table->[-1]->[RANGE_END] = $codepoint; |
242 | } | |
c3a8a2b8 | 243 | |
cf25bb62 JH |
244 | ## |
245 | ## Given a code point range start and end (and optional name), blindly | |
246 | ## append them to the list of ranges for the Table. | |
247 | ## | |
248 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
249 | ## | |
250 | sub Table::RawAppendRange | |
251 | { | |
252 | my $Table = shift; #self | |
253 | my $start = shift; | |
254 | my $end = shift; | |
255 | my $name = shift; | |
256 | $name = "" if not defined $name; ## warning: $name can be "0" | |
257 | ||
258 | push @$Table, [ $start, # RANGE_START | |
259 | $end, # RANGE_END | |
260 | $name ]; # RANGE_NAME | |
261 | } | |
c3a8a2b8 | 262 | |
cf25bb62 JH |
263 | ## |
264 | ## Given a code point (and optional name), add it to the Table. | |
265 | ## | |
266 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
267 | ## | |
268 | sub Table::Append | |
269 | { | |
270 | my $Table = shift; #self | |
271 | my $codepoint = shift; | |
272 | my $name = shift; | |
273 | $name = "" if not defined $name; ## warning: $name can be "0" | |
274 | ||
275 | ## | |
276 | ## If we've already got a range working, and this code point is the next | |
277 | ## one in line, and if the name is the same, just extend the current range. | |
278 | ## | |
279 | if ($Table->NotEmpty | |
280 | and | |
281 | $Table->Max == $codepoint - 1 | |
282 | and | |
283 | $Table->[-1]->[RANGE_NAME] eq $name) | |
284 | { | |
285 | $Table->Extend($codepoint); | |
c3a8a2b8 | 286 | } |
cf25bb62 JH |
287 | else |
288 | { | |
289 | $Table->RawAppendRange($codepoint, $codepoint, $name); | |
d73e5302 | 290 | } |
cf25bb62 | 291 | } |
d73e5302 | 292 | |
cf25bb62 JH |
293 | ## |
294 | ## Given a code point range starting value and ending value (and name), | |
295 | ## Add the range to teh Table. | |
296 | ## | |
297 | ## NOTE: Code points must be added in strictly ascending numeric order. | |
298 | ## | |
299 | sub Table::AppendRange | |
300 | { | |
301 | my $Table = shift; #self | |
302 | my $start = shift; | |
303 | my $end = shift; | |
304 | my $name = shift; | |
305 | $name = "" if not defined $name; ## warning: $name can be "0" | |
306 | ||
307 | $Table->Append($start, $name); | |
308 | $Table->Extend($end) if $end > $start; | |
309 | } | |
d73e5302 | 310 | |
cf25bb62 JH |
311 | ## |
312 | ## Return a new Table that represents all code points not in the Table. | |
313 | ## | |
314 | sub Table::Invert | |
315 | { | |
316 | my $Table = shift; #self | |
317 | ||
318 | my $New = Table->New(); | |
319 | my $max = -1; | |
320 | for my $range (@$Table) | |
321 | { | |
322 | my $start = $range->[RANGE_START]; | |
323 | my $end = $range->[RANGE_END]; | |
324 | if ($start-1 >= $max+1) { | |
325 | $New->AppendRange($max+1, $start-1, ""); | |
326 | } | |
327 | $max = $end; | |
d73e5302 | 328 | } |
cf25bb62 JH |
329 | if ($max+1 < $LastUnicodeCodepoint) { |
330 | $New->AppendRange($max+1, $LastUnicodeCodepoint); | |
d73e5302 | 331 | } |
cf25bb62 JH |
332 | return $New; |
333 | } | |
d73e5302 | 334 | |
cf25bb62 JH |
335 | ## |
336 | ## Merges any number of other tables with $self, returning the new table. | |
337 | ## (existing tables are not modified) | |
338 | ## | |
a3a8c5f0 JH |
339 | ## |
340 | ## Args may be Tables, or individual code points (as integers). | |
341 | ## | |
cf25bb62 JH |
342 | ## Can be called as either a constructor or a method. |
343 | ## | |
344 | sub Table::Merge | |
345 | { | |
346 | shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class | |
347 | my @Tables = @_; | |
348 | ||
349 | ## Accumulate all records from all tables | |
350 | my @Records; | |
a3a8c5f0 JH |
351 | for my $Arg (@Tables) |
352 | { | |
353 | if (ref $Arg) { | |
354 | ## arg is a table -- get its ranges | |
355 | push @Records, @$Arg; | |
356 | } else { | |
357 | ## arg is a codepoint, make a range | |
358 | push @Records, [ $Arg, $Arg ] | |
359 | } | |
d73e5302 JH |
360 | } |
361 | ||
cf25bb62 JH |
362 | ## sort by range start, with longer ranges coming first. |
363 | my ($first, @Rest) = sort { | |
364 | ($a->[RANGE_START] <=> $b->[RANGE_START]) | |
365 | or | |
366 | ($b->[RANGE_END] <=> $b->[RANGE_END]) | |
367 | } @Records; | |
368 | ||
369 | my $New = Table->New(); | |
370 | ||
371 | ## Ensuring the first range is there makes the subsequent loop easier | |
372 | $New->AppendRange($first->[RANGE_START], | |
373 | $first->[RANGE_END]); | |
374 | ||
375 | ## Fold in records so long as they add new information. | |
376 | for my $set (@Rest) | |
377 | { | |
378 | my $start = $set->[RANGE_START]; | |
379 | my $end = $set->[RANGE_END]; | |
380 | if ($start > $New->Max) { | |
381 | $New->AppendRange($start, $end); | |
382 | } elsif ($end > $New->Max) { | |
383 | $New->Extend($end); | |
384 | } | |
d73e5302 | 385 | } |
d73e5302 | 386 | |
cf25bb62 | 387 | return $New; |
d73e5302 JH |
388 | } |
389 | ||
cf25bb62 JH |
390 | ## |
391 | ## Given a filename, write a representation of the Table to a file. | |
99598c8c | 392 | ## May have an optional comment as a 2nd arg. |
cf25bb62 JH |
393 | ## |
394 | sub Table::Write | |
395 | { | |
99598c8c | 396 | my $Table = shift; #self |
cf25bb62 | 397 | my $filename = shift; |
99598c8c | 398 | my $comment = shift; |
d73e5302 | 399 | |
5beb625e | 400 | my @OUT = $HEADER; |
99598c8c JH |
401 | if (defined $comment) { |
402 | $comment =~ s/\s+\Z//; | |
403 | $comment =~ s/^/# /gm; | |
5beb625e | 404 | push @OUT, "#\n$comment\n#\n"; |
99598c8c | 405 | } |
5beb625e | 406 | push @OUT, "return <<'END';\n"; |
d73e5302 | 407 | |
cf25bb62 JH |
408 | for my $set (@$Table) |
409 | { | |
410 | my $start = $set->[RANGE_START]; | |
411 | my $end = $set->[RANGE_END]; | |
412 | my $name = $set->[RANGE_NAME]; | |
d73e5302 | 413 | |
cf25bb62 | 414 | if ($start == $end) { |
5beb625e | 415 | push @OUT, sprintf "%04X\t\t%s\n", $start, $name; |
cf25bb62 | 416 | } else { |
5beb625e | 417 | push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name; |
cf25bb62 JH |
418 | } |
419 | } | |
d73e5302 | 420 | |
5beb625e JH |
421 | push @OUT, "END\n"; |
422 | ||
423 | WriteIfChanged($filename, @OUT); | |
424 | } | |
425 | ||
426 | ## This used only for making the test script. | |
427 | ## helper function | |
428 | sub IsUsable($) | |
429 | { | |
430 | my $code = shift; | |
431 | return 0 if $code <= 0x0000; ## don't use null | |
432 | return 0 if $code >= $LastUnicodeCodepoint; ## keep in range | |
433 | return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates | |
434 | return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good | |
435 | return 0 if (($code & 0xFFFF) == 0xFFFE); ## utf8.c says no good | |
436 | return 0 if (($code & 0xFFFF) == 0xFFFF); ## utf8.c says no good | |
437 | return 1; | |
438 | } | |
439 | ||
440 | ## Return a code point that's part of the table. | |
441 | ## Returns nothing if the table is empty (or covers only surrogates). | |
442 | ## This used only for making the test script. | |
443 | sub Table::ValidCode | |
444 | { | |
445 | my $Table = shift; #self | |
446 | for my $set (@$Table) { | |
447 | return $set->[RANGE_END] if IsUsable($set->[RANGE_END]); | |
448 | } | |
449 | return (); | |
450 | } | |
451 | ||
452 | ## Return a code point that's not part of the table | |
453 | ## Returns nothing if the table covers all code points. | |
454 | ## This used only for making the test script. | |
455 | sub Table::InvalidCode | |
456 | { | |
457 | my $Table = shift; #self | |
458 | ||
459 | return 0x1234 if $Table->IsEmpty(); | |
460 | ||
461 | for my $set (@$Table) | |
462 | { | |
463 | if (IsUsable($set->[RANGE_END] + 1)) | |
464 | { | |
465 | return $set->[RANGE_END] + 1; | |
466 | } | |
467 | ||
468 | if (IsUsable($set->[RANGE_START] - 1)) | |
469 | { | |
470 | return $set->[RANGE_START] - 1; | |
471 | } | |
472 | } | |
473 | return (); | |
cf25bb62 JH |
474 | } |
475 | ||
476 | ########################################################################### | |
477 | ########################################################################### | |
478 | ########################################################################### | |
479 | ||
480 | ||
481 | ## | |
482 | ## Called like: | |
99598c8c | 483 | ## New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1); |
cf25bb62 | 484 | ## |
99598c8c | 485 | ## The args must be in that order, although the Fuzzy pair may be omitted. |
cf25bb62 JH |
486 | ## |
487 | ## This creates 'IsAll' as an alias for 'IsAny' | |
488 | ## | |
489 | sub New_Alias($$$@) | |
490 | { | |
491 | my $Type = shift; ## "Is" or "In" | |
492 | my $Alias = shift; | |
99598c8c | 493 | my $SameAs = shift; # expecting "SameAs" -- just ignored |
cf25bb62 JH |
494 | my $Name = shift; |
495 | ||
496 | ## remaining args are optional key/val | |
497 | my %Args = @_; | |
498 | ||
99598c8c | 499 | my $Fuzzy = delete $Args{Fuzzy}; |
cf25bb62 JH |
500 | |
501 | ## sanity check a few args | |
502 | if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') { | |
503 | confess "$0: bad args to New_Alias" | |
d73e5302 JH |
504 | } |
505 | ||
5beb625e JH |
506 | $Alias = CanonicalName($Alias) if $Fuzzy; |
507 | ||
508 | if (not $TableInfo{$Type}->{$Name}) | |
509 | { | |
510 | my $CName = CanonicalName($Name); | |
511 | if ($TableInfo{$Type}->{$CName}) { | |
512 | confess "$0: Use canonical form '$CName' instead of '$Name' for alias."; | |
513 | } else { | |
514 | confess "$0: don't have orignial $Type => $Name to make alias"; | |
515 | } | |
cf25bb62 JH |
516 | } |
517 | if ($TableInfo{$Alias}) { | |
518 | confess "$0: already have original $Type => $Alias; can't make alias"; | |
d73e5302 | 519 | } |
cf25bb62 | 520 | $AliasInfo{$Type}->{$Name} = $Alias; |
99598c8c | 521 | if ($Fuzzy) { |
cf25bb62 JH |
522 | $FuzzyNames{$Type}->{$Alias} = $Name; |
523 | } | |
524 | ||
d73e5302 JH |
525 | } |
526 | ||
d73e5302 | 527 | |
cf25bb62 | 528 | ## All assigned code points |
99598c8c JH |
529 | my $Assigned = Table->New(Is => 'Assigned', |
530 | Desc => "All assigned code points", | |
5beb625e | 531 | Fuzzy => 0); |
d2d499f5 | 532 | |
cf25bb62 JH |
533 | my $Name = Table->New(); ## all characters, individually by name |
534 | my $General = Table->New(); ## all characters, grouped by category | |
535 | my %General; | |
536 | my %Cat; | |
d73e5302 | 537 | |
cf25bb62 | 538 | ## |
551b6b6f | 539 | ## Process UnicodeData.txt (Categories, etc.) |
cf25bb62 | 540 | ## |
44da8cae | 541 | sub UnicodeData_Txt() |
cf25bb62 JH |
542 | { |
543 | my $Bidi = Table->New(); | |
544 | my $Deco = Table->New(); | |
545 | my $Comb = Table->New(); | |
546 | my $Number = Table->New(); | |
99598c8c JH |
547 | my $Mirrored = Table->New(Is => 'Mirrored', |
548 | Desc => "Mirrored in bidirectional text", | |
549 | Fuzzy => 0); | |
d73e5302 | 550 | |
cf25bb62 JH |
551 | my %DC; |
552 | my %Bidi; | |
553 | my %Deco; | |
99598c8c JH |
554 | $Deco{Canon} = Table->New(Is => 'Canon', |
555 | Desc => 'Decomposes to multiple characters', | |
556 | Fuzzy => 0); | |
557 | $Deco{Compat} = Table->New(Is => 'Compat', | |
558 | Desc => 'Compatible with a more-basic character', | |
559 | Fuzzy => 0); | |
cf25bb62 JH |
560 | |
561 | ## Initialize Perl-generated categories | |
551b6b6f | 562 | ## (Categories from UnicodeData.txt are auto-initialized in gencat) |
44da8cae JH |
563 | $Cat{Alnum} = |
564 | Table->New(Is => 'Alnum', Desc => "[[:Alnum:]]", Fuzzy => 0); | |
565 | $Cat{Alpha} = | |
566 | Table->New(Is => 'Alpha', Desc => "[[:Alpha:]]", Fuzzy => 0); | |
567 | $Cat{ASCII} = | |
568 | Table->New(Is => 'ASCII', Desc => "[[:ASCII:]]", Fuzzy => 0); | |
569 | $Cat{Blank} = | |
570 | Table->New(Is => 'Blank', Desc => "[[:Blank:]]", Fuzzy => 0); | |
571 | $Cat{Cntrl} = | |
572 | Table->New(Is => 'Cntrl', Desc => "[[:Cntrl:]]", Fuzzy => 0); | |
573 | $Cat{Digit} = | |
574 | Table->New(Is => 'Digit', Desc => "[[:Digit:]]", Fuzzy => 0); | |
575 | $Cat{Graph} = | |
576 | Table->New(Is => 'Graph', Desc => "[[:Graph:]]", Fuzzy => 0); | |
577 | $Cat{Lower} = | |
578 | Table->New(Is => 'Lower', Desc => "[[:Lower:]]", Fuzzy => 0); | |
579 | $Cat{Print} = | |
580 | Table->New(Is => 'Print', Desc => "[[:Print:]]", Fuzzy => 0); | |
581 | $Cat{Punct} = | |
582 | Table->New(Is => 'Punct', Desc => "[[:Punct:]]", Fuzzy => 0); | |
583 | $Cat{Space} = | |
584 | Table->New(Is => 'Space', Desc => "[[:Space:]]", Fuzzy => 0); | |
585 | $Cat{Title} = | |
586 | Table->New(Is => 'Title', Desc => "[[:Title:]]", Fuzzy => 0); | |
587 | $Cat{Upper} = | |
588 | Table->New(Is => 'Upper', Desc => "[[:Upper:]]", Fuzzy => 0); | |
589 | $Cat{XDigit} = | |
590 | Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0); | |
591 | $Cat{Word} = | |
592 | Table->New(Is => 'Word', Desc => "[[:Word:]]", Fuzzy => 0); | |
593 | $Cat{SpacePerl} = | |
594 | Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0); | |
d73e5302 | 595 | |
cf25bb62 JH |
596 | my %To; |
597 | $To{Upper} = Table->New(); | |
598 | $To{Lower} = Table->New(); | |
599 | $To{Title} = Table->New(); | |
600 | $To{Digit} = Table->New(); | |
601 | ||
602 | sub gencat($$$$) | |
603 | { | |
604 | my ($name, ## Name ("LATIN CAPITAL LETTER A") | |
605 | $cat, ## Category ("Lu", "Zp", "Nd", etc.) | |
606 | $code, ## Code point (as an integer) | |
607 | $op) = @_; | |
608 | ||
609 | my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc | |
610 | ||
611 | $Assigned->$op($code); | |
612 | $Name->$op($code, $name); | |
613 | $General->$op($code, $cat); | |
614 | ||
615 | ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..) | |
99598c8c JH |
616 | $Cat{$cat} ||= Table->New(Is => $cat, |
617 | Desc => "General Category '$cat'", | |
618 | Fuzzy => 0); | |
cf25bb62 JH |
619 | $Cat{$cat}->$op($code); |
620 | ||
621 | ## add to the major category (e.g. "L", "N", "C", ...) | |
99598c8c JH |
622 | $Cat{$MajorCat} ||= Table->New(Is => $MajorCat, |
623 | Desc => "Major Category '$MajorCat'", | |
624 | Fuzzy => 0); | |
cf25bb62 JH |
625 | $Cat{$MajorCat}->$op($code); |
626 | ||
627 | ($General{$name} ||= Table->New)->$op($code, $name); | |
628 | ||
629 | # 005F: SPACING UNDERSCORE | |
d75d706f | 630 | $Cat{Word}->$op($code) if $cat =~ /^[LMN]|Pc/; |
c65e4d19 | 631 | $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/; |
cf25bb62 JH |
632 | $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/; |
633 | ||
d75d706f JH |
634 | my $isspace = |
635 | ($cat =~ /Zs|Zl|Zp/ && | |
636 | $code != 0x200B) # 200B is ZWSP which is for line break control | |
637 | # and therefore it is not part of "space" even while it is "Zs". | |
cf25bb62 JH |
638 | || $code == 0x0009 # 0009: HORIZONTAL TAB |
639 | || $code == 0x000A # 000A: LINE FEED | |
640 | || $code == 0x000B # 000B: VERTICAL TAB | |
641 | || $code == 0x000C # 000C: FORM FEED | |
e6f58734 | 642 | || $code == 0x000D # 000D: CARRIAGE RETURN |
d75d706f JH |
643 | || $code == 0x0085 # 0085: NEL |
644 | ||
645 | ; | |
cf25bb62 | 646 | |
d75d706f | 647 | $Cat{Space}->$op($code) if $isspace; |
cf25bb62 | 648 | |
d75d706f JH |
649 | $Cat{SpacePerl}->$op($code) if $isspace |
650 | && $code != 0x000B; # Backward compat. | |
cf25bb62 | 651 | |
d75d706f JH |
652 | $Cat{Blank}->$op($code) if $isspace |
653 | && !($code == 0x000A || | |
654 | $code == 0x000B || | |
655 | $code == 0x000C || | |
656 | $code == 0x000D || | |
657 | $code == 0x0085 || | |
658 | $cat =~ /^Z[lp]/); | |
cf25bb62 JH |
659 | |
660 | $Cat{Digit}->$op($code) if $cat eq "Nd"; | |
661 | $Cat{Upper}->$op($code) if $cat eq "Lu"; | |
662 | $Cat{Lower}->$op($code) if $cat eq "Ll"; | |
663 | $Cat{Title}->$op($code) if $cat eq "Lt"; | |
664 | $Cat{ASCII}->$op($code) if $code <= 0x007F; | |
665 | $Cat{Cntrl}->$op($code) if $cat =~ /^C/; | |
d75d706f JH |
666 | my $isgraph = !$isspace && $cat !~ /Cc|Cs|Cn/; |
667 | $Cat{Graph}->$op($code) if $isgraph; | |
668 | $Cat{Print}->$op($code) if $isgraph || $isspace; | |
cf25bb62 JH |
669 | $Cat{Punct}->$op($code) if $cat =~ /^P/; |
670 | ||
671 | $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39) ## 0..9 | |
672 | || ($code >= 0x41 && $code <= 0x46) ## A..F | |
673 | || ($code >= 0x61 && $code <= 0x66); ## a..f | |
674 | } | |
d73e5302 | 675 | |
cf25bb62 | 676 | ## open ane read file..... |
551b6b6f JH |
677 | if (not open IN, "UnicodeData.txt") { |
678 | die "$0: UnicodeData.txt: $!\n"; | |
cf25bb62 | 679 | } |
d73e5302 | 680 | |
a3a8c5f0 JH |
681 | ## |
682 | ## For building \p{_CombAbove} and \p{_CanonDCIJ} | |
683 | ## | |
684 | my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE") | |
685 | ||
686 | my %CodeToDeco; ## Maps code to decomp. list for chars with first | |
687 | ## decomp. char an "i" or "j" (for \p{_CanonDCIJ}) | |
688 | ||
689 | ## This is filled in as we go.... | |
99598c8c JH |
690 | my $CombAbove = Table->New(Is => '_CombAbove', |
691 | Desc => '(for internal casefolding use)', | |
692 | Fuzzy => 0); | |
a3a8c5f0 | 693 | |
cf25bb62 JH |
694 | while (<IN>) |
695 | { | |
696 | next unless /^[0-9A-Fa-f]+;/; | |
697 | s/\s+$//; | |
698 | ||
699 | my ($hexcode, ## code point in hex (e.g. "0041") | |
700 | $name, ## character name (e.g. "LATIN CAPITAL LETTER A") | |
701 | $cat, ## category (e.g. "Lu") | |
702 | $comb, ## Canonical combining class (e.t. "230") | |
703 | $bidi, ## directional category (e.g. "L") | |
704 | $deco, ## decomposition mapping | |
705 | $decimal, ## decimal digit value | |
706 | $digit, ## digit value | |
707 | $number, ## numeric value | |
708 | $mirrored, ## mirrored | |
709 | $unicode10, ## name in Unicode 1.0 | |
710 | $comment, ## comment field | |
711 | $upper, ## uppercase mapping | |
712 | $lower, ## lowercase mapping | |
713 | $title, ## titlecase mapping | |
714 | ) = split(/\s*;\s*/); | |
715 | ||
2eb5892f JH |
716 | # Note that in Unicode 3.2 there will be names like |
717 | # LINE FEED (LF), which probably means that \N{} needs | |
718 | # to cope also with LINE FEED and LF. | |
719 | $name = $unicode10 if $name eq '<control>' && $unicode10 ne ''; | |
720 | ||
cf25bb62 JH |
721 | my $code = hex($hexcode); |
722 | ||
a3a8c5f0 JH |
723 | if ($comb and $comb == 230) { |
724 | $CombAbove->Append($code); | |
725 | $_Above_HexCodes{$hexcode} = 1; | |
726 | } | |
727 | ||
728 | ## Used in building \p{_CanonDCIJ} | |
729 | if ($deco and $deco =~ m/^006[9A]\b/) { | |
730 | $CodeToDeco{$code} = $deco; | |
731 | } | |
732 | ||
cf25bb62 JH |
733 | ## |
734 | ## There are a few pairs of lines like: | |
735 | ## AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;; | |
736 | ## D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;; | |
737 | ## that define ranges. | |
738 | ## | |
739 | if ($name =~ /^<(.+), (First|Last)>$/) | |
740 | { | |
741 | $name = $1; | |
742 | gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend'); | |
99598c8c | 743 | #New_Prop(In => $name, $General{$name}, Fuzzy => 1); |
cf25bb62 JH |
744 | } |
745 | else | |
746 | { | |
747 | ## normal (single-character) lines | |
748 | gencat($name, $cat, $code, 'Append'); | |
749 | ||
750 | # No Append() here since since several codes may map into one. | |
751 | $To{Upper}->RawAppendRange($code, $code, $upper) if $upper; | |
752 | $To{Lower}->RawAppendRange($code, $code, $lower) if $lower; | |
753 | $To{Title}->RawAppendRange($code, $code, $title) if $title; | |
754 | $To{Digit}->Append($code, $decimal) if length $decimal; | |
755 | ||
756 | $Bidi->Append($code, $bidi); | |
757 | $Comb->Append($code, $comb) if $comb; | |
758 | $Number->Append($code, $number) if length $number; | |
759 | ||
760 | $Mirrored->Append($code) if $mirrored eq "Y"; | |
761 | ||
99598c8c JH |
762 | $Bidi{$bidi} ||= Table->New(Is => "Bidi$bidi", |
763 | Desc => "Bi-directional category '$bidi'", | |
764 | Fuzzy => 0); | |
cf25bb62 JH |
765 | $Bidi{$bidi}->Append($code); |
766 | ||
767 | if ($deco) | |
768 | { | |
769 | $Deco->Append($code, $deco); | |
770 | if ($deco =~/^<(\w+)>/) | |
771 | { | |
772 | $Deco{Compat}->Append($code); | |
773 | ||
99598c8c JH |
774 | $DC{$1} ||= Table->New(Is => "DC$1", |
775 | Desc => "Compatible with '$1'", | |
776 | Fuzzy => 0); | |
cf25bb62 JH |
777 | $DC{$1}->Append($code); |
778 | } | |
779 | else | |
780 | { | |
781 | $Deco{Canon}->Append($code); | |
782 | } | |
783 | } | |
784 | } | |
785 | } | |
786 | close IN; | |
d2d499f5 | 787 | |
cf25bb62 JH |
788 | ## |
789 | ## Tidy up a few special cases.... | |
790 | ## | |
d73e5302 | 791 | |
cf25bb62 | 792 | $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist |
99598c8c JH |
793 | New_Prop(Is => 'Cn', |
794 | $Cat{Cn}, | |
795 | Desc => "General Category 'Cn' [not functional in Perl]", | |
796 | Fuzzy => 0); | |
d73e5302 | 797 | |
cf25bb62 | 798 | ## Unassigned is the same as 'Cn' |
5beb625e | 799 | New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0); |
d73e5302 | 800 | |
cf25bb62 | 801 | $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn})); ## Now merge in Cn into C |
d73e5302 | 802 | |
d73e5302 | 803 | |
cf25bb62 JH |
804 | # L& is Ll, Lu, and Lt. |
805 | New_Prop(Is => 'L&', | |
806 | Table->Merge(@Cat{qw[Ll Lu Lt]}), | |
99598c8c JH |
807 | Desc => '[\p{Ll}\p{Lu}\p{Lt}]', |
808 | Fuzzy => 0); | |
d73e5302 | 809 | |
cf25bb62 | 810 | ## Any and All are all code points. |
99598c8c JH |
811 | my $Any = Table->New(Is => 'Any', |
812 | Desc => sprintf("[\\x{0000}-\\x{%X}]", | |
813 | $LastUnicodeCodepoint), | |
5beb625e | 814 | Fuzzy => 0); |
cf25bb62 | 815 | $Any->RawAppendRange(0, $LastUnicodeCodepoint); |
d73e5302 | 816 | |
5beb625e | 817 | New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0); |
d73e5302 | 818 | |
a3a8c5f0 JH |
819 | ## |
820 | ## Build special properties for Perl's internal case-folding needs: | |
821 | ## \p{_CaseIgnorable} | |
822 | ## \p{_CanonDCIJ} | |
823 | ## \p{_CombAbove} | |
824 | ## _CombAbove was built above. Others are built here.... | |
825 | ## | |
826 | ||
827 | ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010] | |
828 | New_Prop(Is => '_CaseIgnorable', | |
829 | Table->Merge($Cat{Mn}, | |
830 | 0x00AD, #SOFT HYPHEN | |
831 | 0x2010), #HYPHEN | |
99598c8c JH |
832 | Desc => '(for internal casefolding use)', |
833 | Fuzzy => 0); | |
a3a8c5f0 JH |
834 | |
835 | ||
836 | ## \p{_CanonDCIJ} is fairly complex... | |
99598c8c JH |
837 | my $CanonCDIJ = Table->New(Is => '_CanonDCIJ', |
838 | Desc => '(for internal casefolding use)', | |
839 | Fuzzy => 0); | |
a3a8c5f0 JH |
840 | ## It contains the ASCII 'i' and 'j'.... |
841 | $CanonCDIJ->Append(0x0069); # ASCII ord("i") | |
842 | $CanonCDIJ->Append(0x006A); # ASCII ord("j") | |
843 | ## ...and any character with a decomposition that starts with either of | |
844 | ## those code points, but only if the decomposition does not have any | |
845 | ## combining character with the "ABOVE" canonical combining class. | |
846 | for my $code (sort { $a <=> $b} keys %CodeToDeco) | |
847 | { | |
848 | ## Need to ensure that all decomposition characters do not have | |
849 | ## a %HexCodeToComb in %AboveCombClasses. | |
850 | my $want = 1; | |
851 | for my $deco_hexcode (split / /, $CodeToDeco{$code}) | |
852 | { | |
853 | if (exists $_Above_HexCodes{$deco_hexcode}) { | |
854 | ## one of the decmposition chars has an ABOVE combination | |
855 | ## class, so we're not interested in this one | |
856 | $want = 0; | |
857 | last; | |
858 | } | |
859 | } | |
860 | if ($want) { | |
861 | $CanonCDIJ->Append($code); | |
862 | } | |
863 | } | |
864 | ||
865 | ||
d73e5302 | 866 | |
cf25bb62 JH |
867 | ## |
868 | ## Now dump the files. | |
869 | ## | |
870 | $Name->Write("Name.pl"); | |
871 | $Bidi->Write("Bidirectional.pl"); | |
872 | $Comb->Write("CombiningClass.pl"); | |
873 | $Deco->Write("Decomposition.pl"); | |
874 | $Number->Write("Number.pl"); | |
875 | $General->Write("Category.pl"); | |
876 | ||
877 | for my $to (sort keys %To) { | |
878 | $To{$to}->Write("To/$to.pl"); | |
d73e5302 JH |
879 | } |
880 | } | |
881 | ||
cf25bb62 | 882 | ## |
551b6b6f | 883 | ## Process LineBreak.txt |
cf25bb62 | 884 | ## |
551b6b6f | 885 | sub LineBreak_Txt() |
cf25bb62 | 886 | { |
551b6b6f JH |
887 | if (not open IN, "LineBreak.txt") { |
888 | die "$0: LineBreak.txt: $!\n"; | |
cf25bb62 | 889 | } |
d73e5302 | 890 | |
cf25bb62 JH |
891 | my $Lbrk = Table->New(); |
892 | my %Lbrk; | |
d73e5302 | 893 | |
cf25bb62 JH |
894 | while (<IN>) |
895 | { | |
896 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/; | |
d73e5302 | 897 | |
cf25bb62 | 898 | my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3); |
d73e5302 | 899 | |
cf25bb62 | 900 | $Lbrk->Append($first, $lbrk); |
d73e5302 | 901 | |
99598c8c JH |
902 | $Lbrk{$lbrk} ||= Table->New(Is => "Lbrk$lbrk", |
903 | Desc => "Linebreak category '$lbrk'", | |
904 | Fuzzy => 0); | |
cf25bb62 | 905 | $Lbrk{$lbrk}->Append($first); |
d73e5302 | 906 | |
cf25bb62 JH |
907 | if ($last) { |
908 | $Lbrk->Extend($last); | |
909 | $Lbrk{$lbrk}->Extend($last); | |
d73e5302 JH |
910 | } |
911 | } | |
cf25bb62 | 912 | close IN; |
d73e5302 | 913 | |
cf25bb62 JH |
914 | $Lbrk->Write("Lbrk.pl"); |
915 | } | |
d73e5302 | 916 | |
cf25bb62 | 917 | ## |
551b6b6f | 918 | ## Process ArabicShaping.txt. |
cf25bb62 | 919 | ## |
551b6b6f | 920 | sub ArabicShaping_txt() |
cf25bb62 | 921 | { |
551b6b6f JH |
922 | if (not open IN, "ArabicShaping.txt") { |
923 | die "$0: ArabicShaping.txt: $!\n"; | |
cf25bb62 | 924 | } |
d73e5302 | 925 | |
cf25bb62 JH |
926 | my $ArabLink = Table->New(); |
927 | my $ArabLinkGroup = Table->New(); | |
d73e5302 | 928 | |
cf25bb62 JH |
929 | while (<IN>) |
930 | { | |
931 | next unless /^[0-9A-Fa-f]+;/; | |
932 | s/\s+$//; | |
d73e5302 | 933 | |
cf25bb62 JH |
934 | my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/); |
935 | my $code = hex($hexcode); | |
936 | $ArabLink->Append($code, $link); | |
937 | $ArabLinkGroup->Append($code, $linkgroup); | |
d73e5302 | 938 | } |
cf25bb62 JH |
939 | close IN; |
940 | ||
941 | $ArabLink->Write("ArabLink.pl"); | |
942 | $ArabLinkGroup->Write("ArabLnkGrp.pl"); | |
d73e5302 JH |
943 | } |
944 | ||
cf25bb62 JH |
945 | ## |
946 | ## Process Jamo.txt. | |
947 | ## | |
948 | sub Jamo_txt() | |
949 | { | |
950 | if (not open IN, "Jamo.txt") { | |
951 | die "$0: Jamo.txt: $!\n"; | |
952 | } | |
953 | my $Short = Table->New(); | |
d73e5302 | 954 | |
cf25bb62 JH |
955 | while (<IN>) |
956 | { | |
957 | next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/; | |
958 | my ($code, $short) = (hex($1), $2); | |
d73e5302 | 959 | |
cf25bb62 | 960 | $Short->Append($code, $short); |
d73e5302 | 961 | } |
cf25bb62 JH |
962 | close IN; |
963 | $Short->Write("JamoShort.pl"); | |
d73e5302 JH |
964 | } |
965 | ||
cf25bb62 JH |
966 | ## |
967 | ## Process Scripts.txt. | |
968 | ## | |
969 | sub Scripts_txt() | |
970 | { | |
971 | my @ScriptInfo; | |
d73e5302 | 972 | |
cf25bb62 JH |
973 | if (not open(IN, "Scripts.txt")) { |
974 | die "$0: Scripts.txt: $!\n"; | |
975 | } | |
976 | while (<IN>) { | |
977 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
d73e5302 | 978 | |
cf25bb62 JH |
979 | # Wait until all the scripts have been read since |
980 | # they are not listed in numeric order. | |
981 | push @ScriptInfo, [ hex($1), hex($2||""), $3 ]; | |
982 | } | |
983 | close IN; | |
d73e5302 | 984 | |
cf25bb62 | 985 | # Now append the scripts properties in their code point order. |
d73e5302 | 986 | |
cf25bb62 JH |
987 | my %Script; |
988 | my $Scripts = Table->New(); | |
d73e5302 | 989 | |
cf25bb62 JH |
990 | for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo) |
991 | { | |
992 | my ($first, $last, $name) = @$script; | |
993 | $Scripts->Append($first, $name); | |
d73e5302 | 994 | |
99598c8c JH |
995 | $Script{$name} ||= Table->New(Is => $name, |
996 | Desc => "Script '$name'", | |
997 | Fuzzy => 1); | |
cf25bb62 | 998 | $Script{$name}->Append($first, $name); |
d73e5302 | 999 | |
cf25bb62 JH |
1000 | if ($last) { |
1001 | $Scripts->Extend($last); | |
1002 | $Script{$name}->Extend($last); | |
1003 | } | |
1004 | } | |
d73e5302 | 1005 | |
cf25bb62 | 1006 | $Scripts->Write("Scripts.pl"); |
d73e5302 | 1007 | |
cf25bb62 JH |
1008 | ## Common is everything not explicitly assigned to a Script |
1009 | ## | |
1010 | ## ***shouldn't this be intersected with \p{Assigned}? ****** | |
1011 | ## | |
99598c8c JH |
1012 | New_Prop(Is => 'Common', |
1013 | $Scripts->Invert, | |
1014 | Desc => 'Pseudo-Script of codepoints not in other Unicode scripts', | |
1015 | Fuzzy => 1); | |
cf25bb62 | 1016 | } |
d73e5302 | 1017 | |
cf25bb62 JH |
1018 | ## |
1019 | ## Given a name like "Close Punctuation", return a regex (that when applied | |
1020 | ## with /i) matches any valid form of that name (e.g. "ClosePunctuation", | |
1021 | ## "Close-Punctuation", etc.) | |
1022 | ## | |
1023 | ## Accept any space, dash, or underbar where in the official name there is | |
1024 | ## space or a dash (or underbar, but there never is). | |
1025 | ## | |
1026 | ## | |
1027 | sub NameToRegex($) | |
1028 | { | |
1029 | my $Name = shift; | |
1030 | $Name =~ s/[- _]/(?:[-_]|\\s+)?/g; | |
1031 | return $Name; | |
1032 | } | |
d73e5302 | 1033 | |
cf25bb62 JH |
1034 | ## |
1035 | ## Process Blocks.txt. | |
1036 | ## | |
1037 | sub Blocks_txt() | |
1038 | { | |
1039 | my $Blocks = Table->New(); | |
1040 | my %Blocks; | |
d73e5302 | 1041 | |
cf25bb62 JH |
1042 | if (not open IN, "Blocks.txt") { |
1043 | die "$0: Blocks.txt: $!\n"; | |
1044 | } | |
d73e5302 | 1045 | |
cf25bb62 JH |
1046 | while (<IN>) |
1047 | { | |
1048 | #next if not /Private Use$/; | |
1049 | next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/; | |
d73e5302 | 1050 | |
cf25bb62 | 1051 | my ($first, $last, $name) = (hex($1), hex($2), $3); |
d73e5302 | 1052 | |
cf25bb62 | 1053 | $Blocks->Append($first, $name); |
76ccdbe2 | 1054 | |
99598c8c JH |
1055 | $Blocks{$name} ||= Table->New(In => $name, |
1056 | Desc => "Block '$name'", | |
1057 | Fuzzy => 1); | |
cf25bb62 | 1058 | $Blocks{$name}->Append($first, $name); |
76ccdbe2 | 1059 | |
cf25bb62 JH |
1060 | if ($last and $last != $first) { |
1061 | $Blocks->Extend($last); | |
1062 | $Blocks{$name}->Extend($last); | |
d73e5302 | 1063 | } |
d73e5302 | 1064 | } |
cf25bb62 JH |
1065 | close IN; |
1066 | ||
1067 | $Blocks->Write("Blocks.pl"); | |
d73e5302 JH |
1068 | } |
1069 | ||
cf25bb62 JH |
1070 | ## |
1071 | ## Read in the PropList.txt. It contains extended properties not | |
551b6b6f | 1072 | ## listed in the UnicodeData.txt, such as 'Other_Alphabetic': |
cf25bb62 JH |
1073 | ## alphabetic but not of the general category L; many modifiers |
1074 | ## belong to this extended property category: while they are not | |
1075 | ## alphabets, they are alphabetic in nature. | |
1076 | ## | |
1077 | sub PropList_txt() | |
1078 | { | |
1079 | my @PropInfo; | |
1080 | ||
1081 | if (not open IN, "PropList.txt") { | |
1082 | die "$0: PropList.txt: $!\n"; | |
1083 | } | |
d73e5302 | 1084 | |
cf25bb62 JH |
1085 | while (<IN>) |
1086 | { | |
1087 | next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/; | |
d73e5302 | 1088 | |
cf25bb62 JH |
1089 | # Wait until all the extended properties have been read since |
1090 | # they are not listed in numeric order. | |
1091 | push @PropInfo, [ hex($1), hex($2||""), $3 ]; | |
1092 | } | |
1093 | close IN; | |
71d929cb | 1094 | |
cf25bb62 JH |
1095 | # Now append the extended properties in their code point order. |
1096 | my $Props = Table->New(); | |
1097 | my %Prop; | |
71d929cb | 1098 | |
cf25bb62 JH |
1099 | for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo) |
1100 | { | |
1101 | my ($first, $last, $name) = @$prop; | |
1102 | $Props->Append($first, $name); | |
71d929cb | 1103 | |
99598c8c JH |
1104 | $Prop{$name} ||= Table->New(Is => $name, |
1105 | Desc => "Extended property '$name'", | |
1106 | Fuzzy => 1); | |
cf25bb62 | 1107 | $Prop{$name}->Append($first, $name); |
71d929cb | 1108 | |
cf25bb62 JH |
1109 | if ($last) { |
1110 | $Props->Extend($last); | |
1111 | $Prop{$name}->Extend($last); | |
1112 | } | |
71d929cb | 1113 | } |
d73e5302 | 1114 | |
cf25bb62 | 1115 | # Alphabetic is L and Other_Alphabetic. |
99598c8c | 1116 | New_Prop(Is => 'Alphabetic', |
cf25bb62 | 1117 | Table->Merge($Cat{L}, $Prop{Other_Alphabetic}), |
99598c8c JH |
1118 | Desc => '[\p{L}\p{OtherAlphabetic}]', # use canonical names here |
1119 | Fuzzy => 1); | |
cf25bb62 JH |
1120 | |
1121 | # Lowercase is Ll and Other_Lowercase. | |
99598c8c | 1122 | New_Prop(Is => 'Lowercase', |
cf25bb62 | 1123 | Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}), |
99598c8c JH |
1124 | Desc => '[\p{Ll}\p{OtherLowercase}]', # use canonical names here |
1125 | Fuzzy => 1); | |
cf25bb62 JH |
1126 | |
1127 | # Uppercase is Lu and Other_Uppercase. | |
1128 | New_Prop(Is => 'Uppercase', | |
1129 | Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}), | |
99598c8c JH |
1130 | Desc => '[\p{Lu}\p{Other_Uppercase}]', # use canonical names here |
1131 | Fuzzy => 1); | |
cf25bb62 JH |
1132 | |
1133 | # Math is Sm and Other_Math. | |
1134 | New_Prop(Is => 'Math', | |
1135 | Table->Merge($Cat{Sm}, $Prop{Other_Math}), | |
99598c8c JH |
1136 | Desc => '[\p{Sm}\p{OtherMath}]', # use canonical names here |
1137 | Fuzzy => 1); | |
cf25bb62 JH |
1138 | |
1139 | # ID_Start is Ll, Lu, Lt, Lm, Lo, and Nl. | |
1140 | New_Prop(Is => 'ID_Start', | |
1141 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}), | |
99598c8c JH |
1142 | Desc => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}]', |
1143 | Fuzzy => 1); | |
cf25bb62 JH |
1144 | |
1145 | # ID_Continue is ID_Start, Mn, Mc, Nd, and Pc. | |
1146 | New_Prop(Is => 'ID_Continue', | |
1147 | Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]}), | |
99598c8c JH |
1148 | Desc => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}]', |
1149 | Fuzzy => 1); | |
d73e5302 JH |
1150 | } |
1151 | ||
cf25bb62 JH |
1152 | sub Make_GC_Aliases() |
1153 | { | |
1154 | ## | |
1155 | ## The mapping from General Category long forms to short forms is | |
1156 | ## currently hardwired here since no simple data file in the UCD | |
1157 | ## seems to do that. Unicode 3.2 will assumedly correct this. | |
1158 | ## | |
1159 | my %Is = ( | |
d73e5302 | 1160 | 'Letter' => 'L', |
e150c829 JH |
1161 | 'Uppercase_Letter' => 'Lu', |
1162 | 'Lowercase_Letter' => 'Ll', | |
1163 | 'Titlecase_Letter' => 'Lt', | |
1164 | 'Modifier_Letter' => 'Lm', | |
1165 | 'Other_Letter' => 'Lo', | |
d73e5302 JH |
1166 | |
1167 | 'Mark' => 'M', | |
e150c829 JH |
1168 | 'Non_Spacing_Mark' => 'Mn', |
1169 | 'Spacing_Mark' => 'Mc', | |
1170 | 'Enclosing_Mark' => 'Me', | |
d73e5302 JH |
1171 | |
1172 | 'Separator' => 'Z', | |
e150c829 JH |
1173 | 'Space_Separator' => 'Zs', |
1174 | 'Line_Separator' => 'Zl', | |
1175 | 'Paragraph_Separator' => 'Zp', | |
d73e5302 JH |
1176 | |
1177 | 'Number' => 'N', | |
e150c829 JH |
1178 | 'Decimal_Number' => 'Nd', |
1179 | 'Letter_Number' => 'Nl', | |
1180 | 'Other_Number' => 'No', | |
d73e5302 JH |
1181 | |
1182 | 'Punctuation' => 'P', | |
e150c829 JH |
1183 | 'Connector_Punctuation' => 'Pc', |
1184 | 'Dash_Punctuation' => 'Pd', | |
1185 | 'Open_Punctuation' => 'Ps', | |
1186 | 'Close_Punctuation' => 'Pe', | |
1187 | 'Initial_Punctuation' => 'Pi', | |
1188 | 'Final_Punctuation' => 'Pf', | |
1189 | 'Other_Punctuation' => 'Po', | |
d73e5302 JH |
1190 | |
1191 | 'Symbol' => 'S', | |
e150c829 JH |
1192 | 'Math_Symbol' => 'Sm', |
1193 | 'Currency_Symbol' => 'Sc', | |
1194 | 'Modifier_Symbol' => 'Sk', | |
1195 | 'Other_Symbol' => 'So', | |
d73e5302 JH |
1196 | |
1197 | 'Other' => 'C', | |
1198 | 'Control' => 'Cc', | |
1199 | 'Format' => 'Cf', | |
1200 | 'Surrogate' => 'Cs', | |
1201 | 'Private Use' => 'Co', | |
e150c829 | 1202 | 'Unassigned' => 'Cn', |
cf25bb62 | 1203 | ); |
d2d499f5 | 1204 | |
cf25bb62 JH |
1205 | ## make the aliases.... |
1206 | while (my ($Alias, $Name) = each %Is) { | |
99598c8c | 1207 | New_Alias(Is => $Alias, SameAs => $Name, Fuzzy => 1); |
cf25bb62 JH |
1208 | } |
1209 | } | |
d2d499f5 | 1210 | |
5beb625e JH |
1211 | |
1212 | ## | |
1213 | ## These are used in: | |
1214 | ## MakePropTestScript() | |
1215 | ## WriteAllMappings() | |
1216 | ## for making the test script. | |
1217 | ## | |
1218 | my %FuzzyNameToTest; | |
1219 | my %ExactNameToTest; | |
1220 | ||
1221 | ||
1222 | ## This used only for making the test script | |
1223 | sub GenTests($$$$) | |
1224 | { | |
1225 | my $FH = shift; | |
1226 | my $Prop = shift; | |
1227 | my $MatchCode = shift; | |
1228 | my $FailCode = shift; | |
1229 | ||
1230 | if (defined $MatchCode) { | |
1231 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode; | |
1232 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode; | |
1233 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode; | |
1234 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode; | |
1235 | } | |
1236 | if (defined $FailCode) { | |
1237 | printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode; | |
1238 | printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode; | |
1239 | printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode; | |
1240 | printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode; | |
1241 | } | |
1242 | } | |
1243 | ||
1244 | ## This used only for making the test script | |
1245 | sub ExpectError($$) | |
1246 | { | |
1247 | my $FH = shift; | |
1248 | my $prop = shift; | |
1249 | ||
1250 | print $FH qq/Error('\\p{$prop}');\n/; | |
1251 | print $FH qq/Error('\\P{$prop}');\n/; | |
1252 | } | |
1253 | ||
1254 | ## This used only for making the test script | |
1255 | my @GoodSeps = ( | |
1256 | " ", | |
1257 | "-", | |
1258 | " \t ", | |
1259 | "", | |
1260 | "", | |
1261 | "_", | |
1262 | ); | |
1263 | my @BadSeps = ( | |
1264 | "--", | |
1265 | "__", | |
1266 | " _", | |
1267 | "/" | |
1268 | ); | |
1269 | ||
1270 | ## This used only for making the test script | |
1271 | sub RandomlyFuzzifyName($;$) | |
1272 | { | |
1273 | my $Name = shift; | |
1274 | my $WantError = shift; ## if true, make an error | |
1275 | ||
1276 | my @parts; | |
1277 | for my $part (split /[-\s_]+/, $Name) | |
1278 | { | |
1279 | if (@parts) { | |
1280 | if ($WantError and rand() < 0.3) { | |
1281 | push @parts, $BadSeps[rand(@BadSeps)]; | |
1282 | $WantError = 0; | |
1283 | } else { | |
1284 | push @parts, $GoodSeps[rand(@GoodSeps)]; | |
1285 | } | |
1286 | } | |
1287 | my $switch = int rand(4); | |
1288 | if ($switch == 0) { | |
1289 | push @parts, uc $part; | |
1290 | } elsif ($switch == 1) { | |
1291 | push @parts, lc $part; | |
1292 | } elsif ($switch == 2) { | |
1293 | push @parts, ucfirst $part; | |
1294 | } else { | |
1295 | push @parts, $part; | |
1296 | } | |
1297 | } | |
1298 | my $new = join('', @parts); | |
1299 | ||
1300 | if ($WantError) { | |
1301 | if (rand() >= 0.5) { | |
1302 | $new .= $BadSeps[rand(@BadSeps)]; | |
1303 | } else { | |
1304 | $new = $BadSeps[rand(@BadSeps)] . $new; | |
1305 | } | |
1306 | } | |
1307 | return $new; | |
1308 | } | |
1309 | ||
1310 | ## This used only for making the test script | |
1311 | sub MakePropTestScript() | |
1312 | { | |
1313 | ## this written directly -- it's huge. | |
1314 | if (not open OUT, ">TestProp.pl") { | |
1315 | die "$0: TestProp.pl: $!\n"; | |
1316 | } | |
1317 | print OUT <DATA>; | |
1318 | ||
1319 | while (my ($Name, $Table) = each %ExactNameToTest) | |
1320 | { | |
1321 | GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode); | |
1322 | ExpectError(*OUT, uc $Name) if uc $Name ne $Name; | |
1323 | ExpectError(*OUT, lc $Name) if lc $Name ne $Name; | |
1324 | } | |
1325 | ||
1326 | ||
1327 | while (my ($Name, $Table) = each %FuzzyNameToTest) | |
1328 | { | |
1329 | my $Orig = $CanonicalToOrig{$Name}; | |
1330 | my %Names = ( | |
1331 | $Name => 1, | |
1332 | $Orig => 1, | |
1333 | RandomlyFuzzifyName($Orig) => 1 | |
1334 | ); | |
1335 | ||
1336 | for my $N (keys %Names) { | |
1337 | GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode); | |
1338 | } | |
1339 | ||
1340 | ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR')); | |
1341 | } | |
1342 | ||
1343 | print OUT "Finished();\n"; | |
1344 | close OUT; | |
1345 | } | |
1346 | ||
1347 | ||
1348 | ## | |
1349 | ## These are used only in: | |
1350 | ## RegisterFileForName() | |
1351 | ## WriteAllMappings() | |
1352 | ## | |
1353 | my %Exact; ## will become %utf8::Exact; | |
1354 | my %Canonical; ## will become %utf8::Canonical; | |
1355 | my %CaComment; ## Comment for %Canonical entry of same key | |
1356 | ||
1357 | ## | |
1358 | ## Given info about a name and a datafile that it should be associated with, | |
1359 | ## register that assocation in %Exact and %Canonical. | |
1360 | sub RegisterFileForName($$$$) | |
1361 | { | |
1362 | my $Type = shift; | |
1363 | my $Name = shift; | |
1364 | my $IsFuzzy = shift; | |
1365 | my $filename = shift; | |
1366 | ||
1367 | ## | |
1368 | ## Now in details for the mapping. $Type eq 'Is' has the | |
1369 | ## Is removed, as it will be removed in utf8_heavy when this | |
1370 | ## data is being checked. In keeps its "In", but a second | |
1371 | ## sans-In record is written if it doesn't conflict with | |
1372 | ## anything already there. | |
1373 | ## | |
1374 | if (not $IsFuzzy) | |
1375 | { | |
1376 | if ($Type eq 'Is') { | |
1377 | die "oops[$Name]" if $Exact{$Name}; | |
1378 | $Exact{$Name} = $filename; | |
1379 | } else { | |
1380 | die "oops[$Type$Name]" if $Exact{"$Type$Name"}; | |
1381 | $Exact{"$Type$Name"} = $filename; | |
1382 | $Exact{$Name} = $filename if not $Exact{$Name}; | |
1383 | } | |
1384 | } | |
1385 | else | |
1386 | { | |
1387 | my $CName = lc $Name; | |
1388 | if ($Type eq 'Is') { | |
1389 | die "oops[$CName]" if $Canonical{$CName}; | |
1390 | $Canonical{$CName} = $filename; | |
1391 | $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2; | |
1392 | } else { | |
1393 | die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"}; | |
1394 | $Canonical{lc "$Type$CName"} = $filename; | |
1395 | $CaComment{lc "$Type$CName"} = "$Type$Name"; | |
1396 | if (not $Canonical{$CName}) { | |
1397 | $Canonical{$CName} = $filename; | |
1398 | $CaComment{$CName} = "$Type$Name"; | |
1399 | } | |
1400 | } | |
1401 | } | |
1402 | } | |
1403 | ||
cf25bb62 JH |
1404 | ## |
1405 | ## Writes the info accumulated in | |
1406 | ## | |
1407 | ## %TableInfo; | |
1408 | ## %FuzzyNames; | |
1409 | ## %AliasInfo; | |
1410 | ## | |
1411 | ## | |
1412 | sub WriteAllMappings() | |
1413 | { | |
99598c8c JH |
1414 | my @MAP; |
1415 | ||
5beb625e | 1416 | my %BaseNames; ## Base names already used (for avoiding 8.3 conflicts) |
cf25bb62 | 1417 | |
5beb625e JH |
1418 | ## 'Is' *MUST* come first, so its names have precidence over 'In's |
1419 | for my $Type ('Is', 'In') | |
1420 | { | |
1421 | my %RawNameToFile; ## a per-$Type cache | |
cf25bb62 | 1422 | |
5beb625e | 1423 | for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}}) |
cf25bb62 | 1424 | { |
5beb625e | 1425 | ## Note: $Name is already canonical |
99598c8c | 1426 | my $Table = $TableInfo{$Type}->{$Name}; |
5beb625e | 1427 | my $IsFuzzy = $FuzzyNames{$Type}->{$Name}; |
99598c8c JH |
1428 | |
1429 | ## Need an 8.3 safe filename (which means "an 8 safe" $filename) | |
5beb625e | 1430 | my $filename; |
cf25bb62 | 1431 | { |
5beb625e JH |
1432 | ## 'Is' items lose 'Is' from the basename. |
1433 | $filename = $Type eq 'Is' ? $Name : "$Type$Name"; | |
1434 | ||
1435 | $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_" | |
1436 | substr($filename, 8) = '' if length($filename) > 8; | |
1437 | ||
1438 | ## | |
1439 | ## Make sure the basename doesn't conflict with something we | |
1440 | ## might have already written. If we have, say, | |
1441 | ## InGreekExtended1 | |
1442 | ## InGreekExtended2 | |
1443 | ## they become | |
1444 | ## InGreekE | |
1445 | ## InGreek2 | |
1446 | ## | |
1447 | while (my $num = $BaseNames{lc $filename}++) | |
1448 | { | |
1449 | $num++; ## so basenames with numbers start with '2', which | |
1450 | ## just looks more natural. | |
1451 | ## Want to append $num, but if it'll make the basename longer | |
1452 | ## than 8 characters, pre-truncate $filename so that the result | |
1453 | ## is acceptable. | |
1454 | my $delta = length($filename) + length($num) - 8; | |
1455 | if ($delta > 0) { | |
1456 | substr($filename, -$delta) = $num; | |
1457 | } else { | |
1458 | $filename .= $num; | |
1459 | } | |
99598c8c | 1460 | } |
5beb625e | 1461 | }; |
99598c8c JH |
1462 | |
1463 | ## | |
1464 | ## Construct a nice comment to add to the file, and build data | |
1465 | ## for the "./Properties" file along the way. | |
1466 | ## | |
1467 | my $Comment; | |
1468 | { | |
1469 | my $Desc = $TableDesc{$Type}->{$Name} || ""; | |
1470 | ## get list of names this table is reference by | |
1471 | my @Supported = $Name; | |
1472 | while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} }) | |
1473 | { | |
1474 | if ($Orig eq $Name) { | |
1475 | push @Supported, $Alias; | |
1476 | } | |
1477 | } | |
1478 | ||
1479 | my $TypeToShow = $Type eq 'Is' ? "" : $Type; | |
1480 | my $OrigProp; | |
1481 | ||
1482 | $Comment = "This file supports:\n"; | |
1483 | for my $N (@Supported) | |
1484 | { | |
1485 | my $IsFuzzy = $FuzzyNames{$Type}->{$N}; | |
5beb625e | 1486 | my $Prop = "\\p{$TypeToShow$Name}"; |
99598c8c JH |
1487 | $OrigProp = $Prop if not $OrigProp; #cache for aliases |
1488 | if ($IsFuzzy) { | |
1489 | $Comment .= "\t$Prop (and fuzzy permutations)\n"; | |
1490 | } else { | |
1491 | $Comment .= "\t$Prop\n"; | |
1492 | } | |
1493 | my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)"; | |
1494 | ||
1495 | push @MAP, sprintf("%s %-42s %s\n", | |
1496 | $IsFuzzy ? '*' : ' ', $Prop, $MyDesc); | |
1497 | } | |
1498 | if ($Desc) { | |
1499 | $Comment .= "\nMeaning: $Desc\n"; | |
1500 | } | |
1501 | ||
1502 | } | |
cf25bb62 JH |
1503 | ## |
1504 | ## Okay, write the file... | |
1505 | ## | |
5beb625e | 1506 | $Table->Write("lib/$filename.pl", $Comment); |
99598c8c | 1507 | |
5beb625e JH |
1508 | ## and register it |
1509 | $RawNameToFile{$Name} = $filename; | |
1510 | RegisterFileForName($Type => $Name, $IsFuzzy, $filename); | |
cf25bb62 | 1511 | |
5beb625e JH |
1512 | if ($IsFuzzy) |
1513 | { | |
1514 | my $CName = CanonicalName($Type . '_'. $Name); | |
1515 | $FuzzyNameToTest{$Name} = $Table if !$FuzzyNameToTest{$Name}; | |
1516 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; | |
1517 | } else { | |
1518 | $ExactNameToTest{$Name} = $Table; | |
cf25bb62 JH |
1519 | } |
1520 | ||
cf25bb62 JH |
1521 | } |
1522 | ||
5beb625e JH |
1523 | ## Register aliase info |
1524 | for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}}) | |
cf25bb62 | 1525 | { |
5beb625e JH |
1526 | my $Alias = $AliasInfo{$Type}->{$Name}; |
1527 | my $IsFuzzy = $FuzzyNames{$Type}->{$Alias}; | |
1528 | my $filename = $RawNameToFile{$Name}; | |
1529 | die "oops [$Alias]->[$Name]" if not $filename; | |
1530 | RegisterFileForName($Type => $Alias, $IsFuzzy, $filename); | |
1531 | ||
1532 | my $Table = $TableInfo{$Type}->{$Name}; | |
1533 | die "oops" if not $Table; | |
1534 | if ($IsFuzzy) | |
1535 | { | |
1536 | my $CName = CanonicalName($Type .'_'. $Alias); | |
1537 | $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias}; | |
1538 | $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName}; | |
1539 | } else { | |
1540 | $ExactNameToTest{$Alias} = $Table; | |
1541 | } | |
cf25bb62 | 1542 | } |
5beb625e | 1543 | } |
cf25bb62 | 1544 | |
5beb625e JH |
1545 | ## |
1546 | ## Write out the property list | |
1547 | ## | |
1548 | { | |
1549 | my @OUT = ( | |
1550 | "##\n", | |
1551 | "## This file created by $0\n", | |
1552 | "## List of built-in \\p{...}/\\P{...} properties.\n", | |
1553 | "##\n", | |
1554 | "## '*' means name may be 'fuzzy'\n", | |
1555 | "##\n\n", | |
1556 | sort { substr($a,2) cmp substr($b, 2) } @MAP, | |
1557 | ); | |
1558 | WriteIfChanged('Properties', @OUT); | |
1559 | } | |
cf25bb62 | 1560 | |
5beb625e JH |
1561 | use Text::Tabs (); ## using this makes the files about half the size |
1562 | ||
1563 | ## Write Exact.pl | |
1564 | { | |
1565 | my @OUT = ( | |
1566 | $HEADER, | |
1567 | "##\n", | |
1568 | "## Data in this file used by ../utf8_heavy.pl\n", | |
1569 | "##\n\n", | |
1570 | "## Mapping from name to filename in ./lib\n", | |
1571 | "%utf8::Exact = (\n", | |
1572 | ); | |
cf25bb62 | 1573 | |
cf25bb62 JH |
1574 | for my $Name (sort keys %Exact) |
1575 | { | |
1576 | my $File = $Exact{$Name}; | |
5beb625e JH |
1577 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1578 | my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/); | |
1579 | push @OUT, Text::Tabs::unexpand($Text); | |
cf25bb62 | 1580 | } |
5beb625e JH |
1581 | push @OUT, ");\n1;\n"; |
1582 | ||
1583 | WriteIfChanged('Exact.pl', @OUT); | |
1584 | } | |
cf25bb62 | 1585 | |
5beb625e JH |
1586 | ## Write Canonical.pl |
1587 | { | |
1588 | my @OUT = ( | |
1589 | $HEADER, | |
1590 | "##\n", | |
1591 | "## Data in this file used by ../utf8_heavy.pl\n", | |
1592 | "##\n\n", | |
1593 | "## Mapping from lc(canonical name) to filename in ./lib\n", | |
1594 | "%utf8::Canonical = (\n", | |
1595 | ); | |
1596 | my $Trail = ""; ## used just to keep the spacing pretty | |
1597 | for my $Name (sort keys %Canonical) | |
cf25bb62 | 1598 | { |
5beb625e JH |
1599 | my $File = $Canonical{$Name}; |
1600 | if ($CaComment{$Name}) { | |
1601 | push @OUT, "\n" if not $Trail; | |
1602 | push @OUT, " # $CaComment{$Name}\n"; | |
1603 | $Trail = "\n"; | |
1604 | } else { | |
1605 | $Trail = ""; | |
cf25bb62 | 1606 | } |
5beb625e JH |
1607 | $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name "; |
1608 | my $Text = sprintf(" %-41s => %s,\n$Trail", $Name, qq/'$File'/); | |
1609 | push @OUT, Text::Tabs::unexpand($Text); | |
cf25bb62 | 1610 | } |
5beb625e JH |
1611 | push @OUT, ");\n1\n"; |
1612 | WriteIfChanged('Canonical.pl', @OUT); | |
d2d499f5 | 1613 | } |
5beb625e JH |
1614 | |
1615 | MakePropTestScript() if $MakeTestScript; | |
d2d499f5 JH |
1616 | } |
1617 | ||
5beb625e | 1618 | |
551b6b6f | 1619 | sub SpecialCasing_txt() |
cf25bb62 JH |
1620 | { |
1621 | # | |
1622 | # Read in the special cases. | |
1623 | # | |
983ffd37 | 1624 | |
cf25bb62 JH |
1625 | my %CaseInfo; |
1626 | ||
551b6b6f JH |
1627 | if (not open IN, "SpecialCasing.txt") { |
1628 | die "$0: SpecialCasing.txt: $!\n"; | |
cf25bb62 JH |
1629 | } |
1630 | while (<IN>) { | |
1631 | next unless /^[0-9A-Fa-f]+;/; | |
1632 | s/\#.*//; | |
1633 | s/\s+$//; | |
1634 | ||
1635 | my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/); | |
1636 | ||
1637 | if ($condition) { # not implemented yet | |
1638 | print "# SKIPPING $_\n" if $Verbose; | |
1639 | next; | |
1640 | } | |
1641 | ||
1642 | # Wait until all the special cases have been read since | |
1643 | # they are not listed in numeric order. | |
1644 | my $ix = hex($code); | |
a6da3000 JH |
1645 | push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ] |
1646 | unless $code eq $lower; | |
1647 | push @{$CaseInfo{Title}}, [ $ix, $code, $title ] | |
1648 | unless $code eq $title; | |
1649 | push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ] | |
1650 | unless $code eq $upper; | |
cf25bb62 JH |
1651 | } |
1652 | close IN; | |
1653 | ||
1654 | # Now write out the special cases properties in their code point order. | |
1655 | # Prepend them to the To/{Upper,Lower,Title}.pl. | |
1656 | ||
1657 | for my $case (qw(Lower Title Upper)) | |
1658 | { | |
1659 | my $NormalCase = do "To/$case.pl" || die "$0: $@\n"; | |
cf25bb62 | 1660 | |
5beb625e JH |
1661 | my @OUT = ( |
1662 | $HEADER, "\n", | |
1663 | "%utf8::ToSpec$case =\n(\n", | |
1664 | ); | |
cf25bb62 JH |
1665 | |
1666 | for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) { | |
1667 | my ($ix, $code, $to) = @$prop; | |
1668 | my $tostr = | |
1669 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $to; | |
5beb625e | 1670 | push @OUT, sprintf qq['%04X' => "$tostr",\n], $ix; |
5cb851a6 JH |
1671 | # Remove any single-character mappings for |
1672 | # the same character since we are going for | |
1673 | # the special casing rules. | |
1674 | $NormalCase =~ s/^$code\t\t\w+\n//m; | |
cf25bb62 | 1675 | } |
5beb625e JH |
1676 | push @OUT, ( |
1677 | ");\n\n", | |
1678 | "return <<'END';\n", | |
1679 | $NormalCase, | |
1680 | "END\n" | |
1681 | ); | |
1682 | WriteIfChanged("To/$case.pl", @OUT); | |
d2d499f5 | 1683 | } |
d2d499f5 JH |
1684 | } |
1685 | ||
c4051cc5 JH |
1686 | # |
1687 | # Read in the case foldings. | |
1688 | # | |
551b6b6f | 1689 | # We will do full case folding, C + F + I (see CaseFolding.txt). |
c4051cc5 | 1690 | # |
551b6b6f | 1691 | sub CaseFolding_txt() |
cf25bb62 | 1692 | { |
551b6b6f JH |
1693 | if (not open IN, "CaseFolding.txt") { |
1694 | die "$0: CaseFolding.txt: $!\n"; | |
cf25bb62 | 1695 | } |
c4051cc5 | 1696 | |
cf25bb62 | 1697 | my $Fold = Table->New(); |
c4051cc5 JH |
1698 | my %Fold; |
1699 | ||
cf25bb62 | 1700 | while (<IN>) { |
254ba52a | 1701 | # Skip status 'S', simple case folding |
c4051cc5 JH |
1702 | next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/; |
1703 | ||
cf25bb62 | 1704 | my ($code, $status, $fold) = (hex($1), $2, $3); |
c4051cc5 JH |
1705 | |
1706 | if ($status eq 'C') { # Common: one-to-one folding | |
254ba52a | 1707 | # No append() since several codes may fold into one. |
cf25bb62 | 1708 | $Fold->RawAppendRange($code, $code, $fold); |
c4051cc5 | 1709 | } else { # F: full, or I: dotted uppercase I -> dotless lowercase I |
cf25bb62 | 1710 | $Fold{$code} = $fold; |
c4051cc5 JH |
1711 | } |
1712 | } | |
cf25bb62 | 1713 | close IN; |
c4051cc5 | 1714 | |
cf25bb62 | 1715 | $Fold->Write("To/Fold.pl"); |
c4051cc5 JH |
1716 | |
1717 | # | |
1718 | # Prepend the special foldings to the common foldings. | |
1719 | # | |
c4051cc5 | 1720 | my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n"; |
5beb625e JH |
1721 | |
1722 | my @OUT = ( | |
1723 | $HEADER, "\n", | |
1724 | "%utf8::ToSpecFold =\n(\n", | |
1725 | ); | |
cf25bb62 JH |
1726 | for my $code (sort { $a <=> $b } keys %Fold) { |
1727 | my $foldstr = | |
1728 | join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code}; | |
5beb625e | 1729 | push @OUT, sprintf qq['%04X' => "$foldstr",\n], $code; |
c4051cc5 | 1730 | } |
5beb625e JH |
1731 | push @OUT, ( |
1732 | ");\n\n", | |
1733 | "return <<'END';\n", | |
1734 | $CommonFold, | |
1735 | "END\n", | |
1736 | ); | |
1737 | ||
1738 | WriteIfChanged("To/Fold.pl", @OUT); | |
c4051cc5 JH |
1739 | } |
1740 | ||
cf25bb62 JH |
1741 | ## Do it.... |
1742 | ||
44da8cae | 1743 | UnicodeData_Txt(); |
cf25bb62 JH |
1744 | Make_GC_Aliases(); |
1745 | PropList_txt(); | |
1746 | ||
1747 | Scripts_txt(); | |
1748 | Blocks_txt(); | |
1749 | ||
5beb625e JH |
1750 | WriteAllMappings(); |
1751 | ||
551b6b6f JH |
1752 | LineBreak_Txt(); |
1753 | ArabicShaping_txt(); | |
cf25bb62 | 1754 | Jamo_txt(); |
551b6b6f JH |
1755 | SpecialCasing_txt(); |
1756 | CaseFolding_txt(); | |
cf25bb62 | 1757 | |
5beb625e | 1758 | exit(0); |
cf25bb62 | 1759 | |
5beb625e JH |
1760 | ## TRAILING CODE IS USED BY MakePropTestScript() |
1761 | __DATA__ | |
1762 | use strict; | |
1763 | use warnings; | |
1764 | ||
1765 | my $Tests = 0; | |
1766 | my $Fails = 0; | |
cf25bb62 | 1767 | |
5beb625e JH |
1768 | sub Expect($$$) |
1769 | { | |
1770 | my $Expect = shift; | |
1771 | my $String = shift; | |
1772 | my $Regex = shift; | |
1773 | my $Line = (caller)[2]; | |
1774 | ||
1775 | $Tests++; | |
1776 | my $RegObj; | |
1777 | my $result = eval { | |
1778 | $RegObj = qr/$Regex/; | |
1779 | $String =~ $RegObj ? 1 : 0 | |
1780 | }; | |
1781 | ||
1782 | if (not defined $result) { | |
1783 | print "couldn't compile /$Regex/ on $0 line $Line: $@\n"; | |
1784 | $Fails++; | |
1785 | } elsif ($result ^ $Expect) { | |
1786 | print "bad result (expected $Expect) on $0 line $Line: $@\n"; | |
1787 | $Fails++; | |
1788 | } | |
1789 | } | |
d73e5302 | 1790 | |
5beb625e JH |
1791 | sub Error($) |
1792 | { | |
1793 | my $Regex = shift; | |
1794 | $Tests++; | |
1795 | if (eval { 'x' =~ qr/$Regex/; 1 }) { | |
1796 | $Fails++; | |
1797 | my $Line = (caller)[2]; | |
1798 | print "expected error for /$Regex/ on $0 line $Line: $@\n"; | |
1799 | } | |
1800 | } | |
1801 | ||
1802 | sub Finished() | |
1803 | { | |
1804 | if ($Fails == 0) { | |
1805 | print "All $Tests tests passed.\n"; | |
1806 | exit(0); | |
1807 | } else { | |
1808 | print "$Tests tests, $Fails failed!\n"; | |
1809 | exit(-1); | |
1810 | } | |
1811 | } |