2 package CharClass::Matcher;
6 use warnings FATAL => 'all';
8 $Data::Dumper::Useqq= 1;
13 require './regen/regen_lib.pl';
14 require './regen/charset_translations.pl';
15 require "./regen/regcharclass_multi_char_folds.pl";
19 CharClass::Matcher -- Generate C macros that match character classes efficiently
23 perl regen/regcharclass.pl
27 Dynamically generates macros for detecting special charclasses
28 in latin-1, utf8, and codepoint forms. Macros can be set to return
29 the length (in bytes) of the matched codepoint, and/or the codepoint itself.
31 To regenerate F<regcharclass.h>, run this script from perl-root. No arguments
34 Using WHATEVER as an example the following macros can be produced, depending
35 on the input parameters (how to get each is described by internal comments at
36 the C<__DATA__> line):
40 =item C<is_WHATEVER(s,is_utf8)>
42 =item C<is_WHATEVER_safe(s,e,is_utf8)>
44 Do a lookup as appropriate based on the C<is_utf8> flag. When possible
45 comparisons involving octet<128 are done before checking the C<is_utf8>
46 flag, hopefully saving time.
48 The version without the C<_safe> suffix should be used only when the input is
49 known to be well-formed.
51 =item C<is_WHATEVER_utf8(s)>
53 =item C<is_WHATEVER_utf8_safe(s,e)>
55 Do a lookup assuming the string is encoded in (normalized) UTF8.
57 The version without the C<_safe> suffix should be used only when the input is
58 known to be well-formed.
60 =item C<is_WHATEVER_latin1(s)>
62 =item C<is_WHATEVER_latin1_safe(s,e)>
64 Do a lookup assuming the string is encoded in latin-1 (aka plan octets).
66 The version without the C<_safe> suffix should be used only when it is known
67 that C<s> contains at least one character.
69 =item C<is_WHATEVER_cp(cp)>
71 Check to see if the string matches a given codepoint (hypothetically a
72 U32). The condition is constructed as to "break out" as early as
73 possible if the codepoint is out of range of the condition.
77 (cp==X || (cp>X && (cp==Y || (cp>Y && ...))))
79 Thus if the character is X+1 only two comparisons will be done. Making
80 matching lookups slower, but non-matching faster.
82 =item C<what_len_WHATEVER_FOO(arg1, ..., len)>
84 A variant form of each of the macro types described above can be generated, in
85 which the code point is returned by the macro, and an extra parameter (in the
86 final position) is added, which is a pointer for the macro to set the byte
87 length of the returned code point.
89 These forms all have a C<what_len> prefix instead of the C<is_>, for example
90 C<what_len_WHATEVER_safe(s,e,is_utf8,len)> and
91 C<what_len_WHATEVER_utf8(s,len)>.
93 These forms should not be used I<except> on small sets of mostly widely
94 separated code points; otherwise the code generated is inefficient. For these
95 cases, it is best to use the C<is_> forms, and then find the code point with
96 C<utf8_to_uvchr_buf>(). This program can fail with a "deep recursion"
97 message on the worst of the inappropriate sets. Examine the generated macro
98 to see if it is acceptable.
100 =item C<what_WHATEVER_FOO(arg1, ...)>
102 A variant form of each of the C<is_> macro types described above can be generated, in
103 which the code point and not the length is returned by the macro. These have
104 the same caveat as L</what_len_WHATEVER_FOO(arg1, ..., len)>, plus they should
105 not be used where the set contains a NULL, as 0 is returned for two different
106 cases: a) the set doesn't include the input code point; b) the set does
107 include it, and it is a NULL.
111 The above isn't quite complete, as for specialized purposes one can get a
112 macro like C<is_WHATEVER_utf8_no_length_checks(s)>, which assumes that it is
113 already known that there is enough space to hold the character starting at
114 C<s>, but otherwise checks that it is well-formed. In other words, this is
115 intermediary in checking between C<is_WHATEVER_utf8(s)> and
116 C<is_WHATEVER_utf8_safe(s,e)>.
120 perltidy -st -bt=1 -bbt=0 -pt=0 -sbt=1 -ce -nwls== "%f"
125 Author: Yves Orton (demerphq) 2007. Maintained by Perl5 Porters.
129 No tests directly here (although the regex engine will fail tests
130 if this code is broken). Insufficient documentation and no Getopts
131 handler for using the module as a script.
135 You may distribute under the terms of either the GNU General Public
136 License or the Artistic License, as specified in the README file.
140 # Sub naming convention:
141 # __func : private subroutine, can not be called as a method
142 # _func : private method, not meant for external use
143 # func : public method.
146 #-------------------------------------------------------------------------------
148 # ($cp,$n,$l,$u)=__uni_latin($str);
150 # Return a list of arrays, each of which when interpreted correctly
151 # represent the string in some given encoding with specific conditions.
153 # $cp - list of codepoints that make up the string.
154 # $n - list of octets that make up the string if all codepoints are invariant
155 # regardless of if the string is in UTF-8 or not.
156 # $l - list of octets that make up the string in latin1 encoding if all
157 # codepoints < 256, and at least one codepoint is UTF-8 variant.
158 # $u - list of octets that make up the string in utf8 if any codepoint is
162 #-----------+----------
163 # 0 - 127 : $n (127/128 are the values for ASCII platforms)
175 my $only_has_invariants = 1;
176 for my $ch ( split //, $str ) {
178 $max= $cp if $max < $cp;
184 push @cp, $a2n->[$cp];
188 $only_has_invariants = ($charset =~ /ascii/i) ? $max < 128 : $max < 160;
189 if ($only_has_invariants) {
192 $l= [@cp] if $max && $max < 256;
195 for my $ch ( split //, $str ) {
196 push @u, map { ord } split //, cp_2_utfbytes(ord $ch, $charset);
200 return ( \@cp, \@cp_high, $n, $l, $u );
204 # $clean= __clean($expr);
206 # Cleanup a ternary expression, removing unnecessary parens and apply some
207 # simplifications using regexes.
216 $parens= qr/ (?> \( (?> (?: (?> [^()]+ ) | (??{ $parens }) )* ) \) ) /x;
218 ## remove redundant parens
219 1 while $expr =~ s/ \( \s* ( $parens ) \s* \) /$1/gx;
222 # repeatedly simplify conditions like
223 # ( (cond1) ? ( (cond2) ? X : Y ) : Y )
225 # ( ( (cond1) && (cond2) ) ? X : Y )
226 # Also similarly handles expressions like:
227 # : (cond1) ? ( (cond2) ? X : Y ) : Y )
228 # Note the inclusion of the close paren in ([:()]) and the open paren in ([()]) is
229 # purely to ensure we have a balanced set of parens in the expression which makes
230 # it easier to understand the pattern in an editor that understands paren's, we do
231 # not expect either of these cases to actually fire. - Yves
237 \? \s* ($parens|[^()?:\s]+?) \s*
238 : \s* ($parens|[^()?:\s]+?) \s*
242 /$1 ( $2 && $3 ) ? $4 : $5 $6/gx;
243 #$expr=~s/\(\(U8\*\)s\)\[(\d+)\]/S$1/g if length $expr > 8000;
244 #$expr=~s/\s+//g if length $expr > 8000;
246 die "Expression too long" if length $expr > 8000;
252 # $text= __macro(@args);
253 # Join args together by newlines, and then neatly add backslashes to the end
254 # of every line as expected by the C pre-processor for #define's.
258 my $str= join "\n", @_;
260 my @lines= map { s/\s+$//; s/\t/ /g; $_ } split /\n/, $str;
261 my $last= pop @lines;
262 $str= join "\n", ( map { sprintf "%-76s\\", $_ } @lines ), $last;
263 1 while $str =~ s/^(\t*) {8}/$1\t/gm;
268 # my $op=__incrdepth($op);
270 # take an 'op' hashref and add one to it and all its childrens depths.
275 return unless ref $op;
277 __incrdepth( $op->{yes} );
278 __incrdepth( $op->{no} );
282 # join two branches of an opcode together with a condition, incrementing
283 # the depth on the yes branch when we do so.
284 # returns the new root opcode of the tree.
286 my ( $cond, $yes, $no )= @_;
290 yes => __incrdepth( $yes ),
299 no => __incrdepth($no),
305 my $hex_fmt= "0x%02X";
312 # Format 'arg' using the printable character if it has one, or a %x if
313 # not, returning a string containing the result
315 # Return what always returned for an unexpected argument
316 return $hex_fmt unless defined $arg && $arg !~ /\D/;
318 # We convert only things inside Latin1
321 # Find the ASCII equivalent of this argument (as the current character
322 # set might not be ASCII)
323 my $char = chr $self->{n2a}->[$arg];
325 # If printable, return it, escaping \ and '
326 return "'$char'" if $char =~ /[^\\'[:^print:]]/a;
327 return "'\\\\'" if $char eq "\\";
328 return "'\''" if $char eq "'";
330 # Handle the mnemonic controls
331 my $pos = index("\a\b\e\f\n\r\t\cK", $char);
332 return "'\\" . substr("abefnrtv", $pos, 1) . "'" if $pos >= 0;
335 # Otherwise, just the input, formatted
336 return sprintf $hex_fmt, $arg;
343 # my $obj=CLASS->new(op=>'SOMENAME',title=>'blah',txt=>[..]);
345 # Create a new CharClass::Matcher object by parsing the text in
346 # the txt array. Currently applies the following rules:
348 # Element starts with C<0x>, line is evaled the result treated as
349 # a number which is passed to chr().
351 # Element starts with C<">, line is evaled and the result treated
354 # Each string is then stored in the 'strs' subhash as a hash record
355 # made up of the results of __uni_latin1, using the keynames
356 # 'low','latin1','utf8', as well as the synthesized 'LATIN1', 'high', and
357 # 'UTF8' which hold a merge of 'low' and their lowercase equivalents.
359 # Size data is tracked per type in the 'size' subhash.
363 my %n2a; # Inversion of a2n, for each character set
369 die "in " . __PACKAGE__ . " constructor '$_;' is a mandatory field"
375 title => $opt{title} || '',
378 my $charset = $opt{charset};
379 my $a2n = get_a2n($charset);
381 # We need to construct the map going the other way if not already done
382 unless (defined $n2a{$charset}) {
383 for (my $i = 0; $i < 256; $i++) {
384 $n2a{$charset}->[$a2n->[$i]] = $i;
388 foreach my $txt ( @{ $opt{txt} } ) {
390 if ( $str =~ /^[""]/ ) {
392 } elsif ($str =~ / - /x ) { # A range: Replace this element on the
393 # list with its expansion
394 my ($lower, $upper) = $str =~ / 0x (.+?) \s* - \s* 0x (.+) /x;
395 die "Format must be like '0xDEAD - 0xBEAF'; instead was '$str'" if ! defined $lower || ! defined $upper;
396 foreach my $cp (hex $lower .. hex $upper) {
397 push @{$opt{txt}}, sprintf "0x%X", $cp;
400 } elsif ($str =~ s/ ^ N (?= 0x ) //x ) {
401 # Otherwise undocumented, a leading N means is already in the
402 # native character set; don't convert.
404 } elsif ( $str =~ /^0x/ ) {
407 } elsif ( $str =~ / \s* \\p \{ ( .*? ) \} /x) {
409 use Unicode::UCD qw(prop_invlist);
411 my @invlist = prop_invlist($property, '_perl_core_internal_ok');
414 # An empty return could mean an unknown property, or merely
415 # that it is empty. Call in scalar context to differentiate
416 my $count = prop_invlist($property, '_perl_core_internal_ok');
417 die "$property not found" unless defined $count;
420 # Replace this element on the list with the property's expansion
421 for (my $i = 0; $i < @invlist; $i += 2) {
422 foreach my $cp ($invlist[$i] .. $invlist[$i+1] - 1) {
424 # prop_invlist() returns native values; add leading 'N'
426 push @{$opt{txt}}, sprintf "N0x%X", $cp;
430 } elsif ($str =~ / ^ do \s+ ( .* ) /x) {
431 die "do '$1' failed: $!$@" if ! do $1 or $@;
433 } elsif ($str =~ / ^ & \s* ( .* ) /x) { # user-furnished sub() call
434 my @results = eval "$1";
435 die "eval '$1' failed: $@" if $@;
436 push @{$opt{txt}}, @results;
439 die "Unparsable line: $txt\n";
441 my ( $cp, $cp_high, $low, $latin1, $utf8 )= __uni_latin1($charset, $a2n, $str );
442 my $UTF8= $low || $utf8;
443 my $LATIN1= $low || $latin1;
444 my $high = (scalar grep { $_ < 256 } @$cp) ? 0 : $utf8;
445 #die Dumper($txt,$cp,$low,$latin1,$utf8)
446 # if $txt=~/NEL/ or $utf8 and @$utf8>3;
448 @{ $self->{strs}{$str} }{qw( str txt low utf8 latin1 high cp cp_high UTF8 LATIN1 )}=
449 ( $str, $txt, $low, $utf8, $latin1, $high, $cp, $cp_high, $UTF8, $LATIN1 );
450 my $rec= $self->{strs}{$str};
451 foreach my $key ( qw(low utf8 latin1 high cp cp_high UTF8 LATIN1) ) {
452 $self->{size}{$key}{ 0 + @{ $self->{strs}{$str}{$key} } }++
453 if $self->{strs}{$str}{$key};
455 $self->{has_multi} ||= @$cp > 1;
456 $self->{has_ascii} ||= $latin1 && @$latin1;
457 $self->{has_low} ||= $low && @$low;
458 $self->{has_high} ||= !$low && !$latin1;
460 $self->{n2a} = $n2a{$charset};
461 $self->{count}= 0 + keys %{ $self->{strs} };
465 # my $trie = make_trie($type,$maxlen);
467 # using the data stored in the object build a trie of a specific type,
468 # and with specific maximum depth. The trie is made up the elements of
469 # the given types array for each string in the object (assuming it is
472 # returns the trie, or undef if there was no relevant data in the object.
476 my ( $self, $type, $maxlen )= @_;
478 my $strs= $self->{strs};
480 foreach my $rec ( values %$strs ) {
481 die "panic: unknown type '$type'"
482 if !exists $rec->{$type};
483 my $dat= $rec->{$type};
485 next if $maxlen && @$dat > $maxlen;
487 foreach my $elem ( @$dat ) {
488 $node->{$elem} ||= {};
489 $node= $node->{$elem};
491 $node->{''}= $rec->{str};
493 return 0 + keys( %trie ) ? \%trie : undef;
499 # This returns a list of the positions of the bits in the input word that
505 push @positions, $position if $word & 1;
512 # my $optree= _optree()
514 # recursively convert a trie to an optree where every node represents
520 my ( $self, $trie, $test_type, $ret_type, $else, $depth )= @_;
521 return unless defined $trie;
522 if ( $self->{has_multi} and $ret_type =~ /cp|both/ ) {
523 die "Can't do 'cp' optree from multi-codepoint strings";
526 $else= 0 unless defined $else;
527 $depth= 0 unless defined $depth;
529 # if we have an empty string as a key it means we are in an
530 # accepting state and unless we can match further on should
531 # return the value of the '' key.
532 if (exists $trie->{''} ) {
533 # we can now update the "else" value, anything failing to match
534 # after this point should return the value from this.
535 if ( $ret_type eq 'cp' ) {
536 $else= $self->{strs}{ $trie->{''} }{cp}[0];
537 $else= $self->val_fmt($else) if $else > 9;
538 } elsif ( $ret_type eq 'len' ) {
540 } elsif ( $ret_type eq 'both') {
541 $else= $self->{strs}{ $trie->{''} }{cp}[0];
542 $else= $self->val_fmt($else) if $else > 9;
543 $else= "len=$depth, $else";
546 # extract the meaningful keys from the trie, filter out '' as
547 # it means we are an accepting state (end of sequence).
548 my @conds= sort { $a <=> $b } grep { length $_ } keys %$trie;
550 # if we haven't any keys there is no further we can match and we
551 # can return the "else" value.
552 return $else if !@conds;
554 my $test = $test_type =~ /^cp/ ? "cp" : "((const U8*)s)[$depth]";
556 # First we loop over the possible keys/conditions and find out what they
557 # look like; we group conditions with the same optree together.
560 local $Data::Dumper::Sortkeys=1;
561 foreach my $cond ( @conds ) {
563 # get the optree for this child/condition
564 my $res= $self->_optree( $trie->{$cond}, $test_type, $ret_type, $else, $depth + 1 );
565 # convert it to a string with Dumper
566 my $res_code= Dumper( $res );
568 push @{$dmp_res{$res_code}{vals}}, $cond;
569 if (!$dmp_res{$res_code}{optree}) {
570 $dmp_res{$res_code}{optree}= $res;
571 push @res_order, $res_code;
575 # now that we have deduped the optrees we construct a new optree containing the merged
579 foreach my $res_code_idx (0 .. $#res_order) {
580 my $res_code= $res_order[$res_code_idx];
581 $node->{vals}= $dmp_res{$res_code}{vals};
582 $node->{test}= $test;
583 $node->{yes}= $dmp_res{$res_code}{optree};
584 $node->{depth}= $depth;
585 if ($res_code_idx < $#res_order) {
586 $node= $node->{no}= {};
596 # my $optree= optree(%opts);
598 # Convert a trie to an optree, wrapper for _optree
603 my $trie= $self->make_trie( $opt{type}, $opt{max_depth} );
604 $opt{ret_type} ||= 'len';
605 my $test_type= $opt{type} =~ /^cp/ ? 'cp' : 'depth';
606 return $self->_optree( $trie, $test_type, $opt{ret_type}, $opt{else}, 0 );
609 # my $optree= generic_optree(%opts);
611 # build a "generic" optree out of the three 'low', 'latin1', 'utf8'
612 # sets of strings, including a branch for handling the string type check.
619 $opt{ret_type} ||= 'len';
620 my $test_type= 'depth';
621 my $else= $opt{else} || 0;
623 my $latin1= $self->make_trie( 'latin1', $opt{max_depth} );
624 my $utf8= $self->make_trie( 'utf8', $opt{max_depth} );
626 $_= $self->_optree( $_, $test_type, $opt{ret_type}, $else, 0 )
630 $else= __cond_join( "( is_utf8 )", $utf8, $latin1 || $else );
631 } elsif ( $latin1 ) {
632 $else= __cond_join( "!( is_utf8 )", $latin1, $else );
634 if ($opt{type} eq 'generic') {
635 my $low= $self->make_trie( 'low', $opt{max_depth} );
637 $else= $self->_optree( $low, $test_type, $opt{ret_type}, $else, 0 );
646 # create a string length guarded optree.
652 my $type= $opt{type};
654 die "Can't do a length_optree on type 'cp', makes no sense."
657 my $else= ( $opt{else} ||= 0 );
659 return $else if $self->{count} == 0;
661 my $method = $type =~ /generic/ ? 'generic_optree' : 'optree';
662 if ($method eq 'optree' && scalar keys %{$self->{size}{$type}} == 1) {
664 # Here is non-generic output (meaning that we are only generating one
665 # type), and all things that match have the same number ('size') of
666 # bytes. The length guard is simply that we have that number of
668 my @size = keys %{$self->{size}{$type}};
669 my $cond= "((e) - (s)) >= $size[0]";
670 my $optree = $self->$method(%opt);
671 $else= __cond_join( $cond, $optree, $else );
673 elsif ($self->{has_multi}) {
676 # Here, there can be a match of a multiple character string. We use
677 # the traditional method which is to have a branch for each possible
678 # size (longest first) and test for the legal values for that size.
680 %{ $self->{size}{low} || {} },
681 %{ $self->{size}{latin1} || {} },
682 %{ $self->{size}{utf8} || {} }
684 if ($method eq 'generic_optree') {
685 @size= sort { $a <=> $b } keys %sizes;
687 @size= sort { $a <=> $b } keys %{ $self->{size}{$type} };
689 for my $size ( @size ) {
690 my $optree= $self->$method( %opt, type => $type, max_depth => $size );
691 my $cond= "((e)-(s) > " . ( $size - 1 ).")";
692 $else= __cond_join( $cond, $optree, $else );
698 # Here, has more than one possible size, and only matches a single
699 # character. For non-utf8, the needed length is 1; for utf8, it is
700 # found by array lookup 'UTF8SKIP'.
702 # If want just the code points above 255, set up to look for those;
703 # otherwise assume will be looking for all non-UTF-8-invariant code
705 my $trie_type = ($type eq 'high') ? 'high' : 'utf8';
707 # If we do want more than the 0-255 range, find those, and if they
709 if ($opt{type} !~ /latin1/i && ($utf8 = $self->make_trie($trie_type, 0))) {
711 # ... get them into an optree, and set them up as the 'else' clause
712 $utf8 = $self->_optree( $utf8, 'depth', $opt{ret_type}, 0, 0 );
715 # UTF8_IS_START(*s) && ((e) - (s)) >= UTF8SKIP(s))";
716 # to avoid doing the UTF8SKIP and subsequent branches for invariants
717 # that don't match. But the current macros that get generated
718 # have only a few things that can match past this, so I (khw)
719 # don't think it is worth it. (Even better would be to use
720 # calculate_mask(keys %$utf8) instead of UTF8_IS_START, and use it
721 # if it saves a bunch. We assume that input text likely to be
723 my $cond = "LIKELY(((e) - (s)) >= UTF8SKIP(s))";
724 $else = __cond_join($cond, $utf8, $else);
726 # For 'generic', we also will want the latin1 UTF-8 variants for
727 # the case where the input isn't UTF-8.
729 if ($method eq 'generic_optree') {
730 $latin1 = $self->make_trie( 'latin1', 1);
731 $latin1= $self->_optree( $latin1, 'depth', $opt{ret_type}, 0, 0 );
734 # If we want the UTF-8 invariants, get those.
736 if ($opt{type} !~ /non_low|high/
737 && ($low= $self->make_trie( 'low', 1)))
739 $low= $self->_optree( $low, 'depth', $opt{ret_type}, 0, 0 );
741 # Expand out the UTF-8 invariants as a string so that we
742 # can use them as the conditional
743 $low = $self->_cond_as_str( $low, 0, \%opt);
745 # If there are Latin1 variants, add a test for them.
747 $else = __cond_join("(! is_utf8 )", $latin1, $else);
749 elsif ($method eq 'generic_optree') {
751 # Otherwise for 'generic' only we know that what
752 # follows must be valid for just UTF-8 strings,
753 $else->{test} = "( is_utf8 && $else->{test} )";
756 # If the invariants match, we are done; otherwise we have
757 # to go to the 'else' clause.
758 $else = __cond_join($low, 1, $else);
760 elsif ($latin1) { # Here, didn't want or didn't have invariants,
761 # but we do have latin variants
762 $else = __cond_join("(! is_utf8)", $latin1, $else);
765 # We need at least one byte available to start off the tests
766 $else = __cond_join("LIKELY((e) > (s))", $else, 0);
768 else { # Here, we don't want or there aren't any variants. A single
769 # byte available is enough.
770 my $cond= "((e) > (s))";
771 my $optree = $self->$method(%opt);
772 $else= __cond_join( $cond, $optree, $else );
779 sub calculate_mask(@) {
780 # Look at the input list of byte values. This routine returns an array of
781 # mask/base pairs to generate that list.
784 my $list_count = @list;
786 # Consider a set of byte values, A, B, C .... If we want to determine if
787 # <c> is one of them, we can write c==A || c==B || c==C .... If the
788 # values are consecutive, we can shorten that to inRANGE(c, 'A', 'Z'),
789 # which uses far fewer branches. If only some of them are consecutive we
790 # can still save some branches by creating range tests for just those that
791 # are consecutive. _cond_as_str() does this work for looking for ranges.
793 # Another approach is to look at the bit patterns for A, B, C .... and see
794 # if they have some commonalities. That's what this function does. For
795 # example, consider a set consisting of the bytes
796 # 0x42, 0x43, 0x62, and 0x63. We could write:
797 # inRANGE(c, 0x42, 0x43) || inRANGE(c, 0x62, 0x63)
798 # which through the magic of casting has not 4, but 2 tests. But the
799 # following mask/compare also works, and has just one test:
801 # The reason it works is that the set consists of exactly the 4 bit
802 # patterns which have either 0 or 1 in the two bit positions that are 0 in
803 # the mask. They have the same value in each bit position where the mask
804 # is 1. The comparison makes sure that the result matches all bytes which
805 # match those six 1 bits exactly. This can be applied to bytes that
806 # differ in 1 through all 8 bit positions. In order to be a candidate for
807 # this optimization, the number of bytes in the set must be a power of 2.
809 # It may be that the bytes needing to be matched can't be done with a
810 # single mask. But it may be possible to have two (or more) sets, each
811 # with a separate mask. This function attempts to find some way to save
812 # some branches using the mask technique. If not, it returns an empty
813 # list; if so, it returns a list consisting of
814 # [ [compare1, mask1], [compare2, mask2], ...
815 # [compare_n, undef], [compare_m, undef], ...
817 # The <mask> is undef in the above for those bytes that must be tested
820 # This function does not attempt to find the optimal set. To do so would
821 # probably require testing all possible combinations, and keeping track of
822 # the current best one.
824 # There are probably much better algorithms, but this is the one I (khw)
825 # came up with. We start with doing a bit-wise compare of every byte in
826 # the set with every other byte. The results are sorted into arrays of
827 # all those that differ by the same bit positions. These are stored in a
828 # hash with the each key being the bits they differ in. Here is the hash
829 # for the 0x53, 0x54, 0x73, 0x74 set:
857 # The set consisting of values which differ in the 4 bit positions 0, 1,
858 # 2, and 5 from some other value in the set consists of all 4 values.
859 # Likewise all 4 values differ from some other value in the 3 bit
860 # positions 0, 1, and 2; and all 4 values differ from some other value in
861 # the single bit position 5. The keys at the uppermost level in the above
862 # hash, 1, 3, and 4, give the number of bit positions that each sub-key
863 # below it has. For example, the 4 key could have as its value an array
864 # consisting of "0,1,2,5", "0,1,2,6", and "3,4,6,7", if the inputs were
865 # such. The best optimization will group the most values into a single
866 # mask. The most values will be the ones that differ in the most
867 # positions, the ones with the largest value for the topmost key. These
868 # keys, are thus just for convenience of sorting by that number, and do
869 # not have any bearing on the core of the algorithm.
871 # We start with an element from largest number of differing bits. The
872 # largest in this case is 4 bits, and there is only one situation in this
873 # set which has 4 differing bits, "0,1,2,5". We look for any subset of
874 # this set which has 16 values that differ in these 4 bits. There aren't
875 # any, because there are only 4 values in the entire set. We then look at
876 # the next possible thing, which is 3 bits differing in positions "0,1,2".
877 # We look for a subset that has 8 values that differ in these 3 bits.
878 # Again there are none. So we go to look for the next possible thing,
879 # which is a subset of 2**1 values that differ only in bit position 5. 83
880 # and 115 do, so we calculate a mask and base for those and remove them
881 # from every set. Since there is only the one set remaining, we remove
882 # them from just this one. We then look to see if there is another set of
883 # 2 values that differ in bit position 5. 84 and 116 do, so we calculate
884 # a mask and base for those and remove them from every set (again only
885 # this set remains in this example). The set is now empty, and there are
886 # no more sets to look at, so we are done.
888 if ($list_count == 256) { # All 256 is trivially masked
894 # Generate bits-differing lists for each element compared against each
896 for my $i (0 .. $list_count - 2) {
897 for my $j ($i + 1 .. $list_count - 1) {
898 my @bits_that_differ = pop_count($list[$i] ^ $list[$j]);
899 my $differ_count = @bits_that_differ;
900 my $key = join ",", @bits_that_differ;
901 push @{$hash{$differ_count}{$key}}, $list[$i] unless grep { $_ == $list[$i] } @{$hash{$differ_count}{$key}};
902 push @{$hash{$differ_count}{$key}}, $list[$j];
906 print STDERR __LINE__, ": calculate_mask() called: List of values grouped by differing bits: ", Dumper \%hash if DEBUG;
909 foreach my $count (reverse sort { $a <=> $b } keys %hash) {
910 my $need = 2 ** $count; # Need 8 values for 3 differing bits, etc
911 foreach my $bits (sort keys $hash{$count}->%*) {
913 print STDERR __LINE__, ": For $count bit(s) difference ($bits), need $need; have ", scalar @{$hash{$count}{$bits}}, "\n" if DEBUG;
915 # Look only as long as there are at least as many elements in the
916 # subset as are needed
917 while ((my $cur_count = @{$hash{$count}{$bits}}) >= $need) {
919 print STDERR __LINE__, ": Looking at bit positions ($bits): ", Dumper $hash{$count}{$bits} if DEBUG;
921 # Start with the first element in it
922 my $try_base = $hash{$count}{$bits}[0];
923 my @subset = $try_base;
925 # If it succeeds, we return a mask and a base to compare
926 # against the masked value. That base will be the AND of
927 # every element in the subset. Initialize to the one element
929 my $compare = $try_base;
931 # We are trying to find a subset of this that has <need>
932 # elements that differ in the bit positions given by the
933 # string $bits, which is comma separated.
934 my @bits = split ",", $bits;
936 TRY: # Look through the remainder of the list for other
937 # elements that differ only by these bit positions.
939 for (my $i = 1; $i < $cur_count; $i++) {
940 my $try_this = $hash{$count}{$bits}[$i];
941 my @positions = pop_count($try_base ^ $try_this);
943 print STDERR __LINE__, ": $try_base vs $try_this: is (", join(',', @positions), ") a subset of ($bits)?" if DEBUG;;
945 foreach my $pos (@positions) {
946 unless (grep { $pos == $_ } @bits) {
947 print STDERR " No\n" if DEBUG;
948 my $remaining = $cur_count - $i - 1;
949 if ($remaining && @subset + $remaining < $need) {
950 print STDERR __LINE__, ": Can stop trying $try_base, because even if all the remaining $remaining values work, they wouldn't add up to the needed $need when combined with the existing ", scalar @subset, " ones\n" if DEBUG;
957 print STDERR " Yes\n" if DEBUG;
958 push @subset, $try_this;
960 # Add this to the mask base, in case it ultimately
962 $compare &= $try_this;
965 print STDERR __LINE__, ": subset (", join(", ", @subset), ") has ", scalar @subset, " elements; needs $need\n" if DEBUG;
967 if (@subset < $need) {
968 shift @{$hash{$count}{$bits}};
969 next; # Try with next value
974 foreach my $position (@bits) {
975 $mask |= 1 << $position;
977 $mask = ~$mask & 0xFF;
978 push @final_results, [$compare, $mask];
980 printf STDERR "%d: Got it: compare=%d=0x%X; mask=%X\n", __LINE__, $compare, $compare, $mask if DEBUG;
982 # These values are now spoken for. Remove them from future
984 foreach my $remove_count (sort keys %hash) {
985 foreach my $bits (sort keys %{$hash{$remove_count}}) {
986 foreach my $to_remove (@subset) {
987 @{$hash{$remove_count}{$bits}} = grep { $_ != $to_remove } @{$hash{$remove_count}{$bits}};
995 # Any values that remain in the list are ones that have to be tested for
998 foreach my $count (reverse sort { $a <=> $b } keys %hash) {
999 foreach my $bits (sort keys $hash{$count}->%*) {
1000 foreach my $remaining (@{$hash{$count}{$bits}}) {
1002 # If we already know about this value, just ignore it.
1003 next if grep { $remaining == $_ } @individuals;
1005 # Otherwise it needs to be returned as something to match
1007 push @final_results, [$remaining, undef];
1008 push @individuals, $remaining;
1013 # Sort by increasing numeric value
1014 @final_results = sort { $a->[0] <=> $b->[0] } @final_results;
1016 print STDERR __LINE__, ": Final return: ", Dumper \@final_results if DEBUG;
1018 return @final_results;
1022 # turn a list of conditions into a text expression
1023 # - merges ranges of conditions, and joins the result with ||
1025 my ( $self, $op, $combine, $opts_ref )= @_;
1026 my $cond= $op->{vals};
1027 my $test= $op->{test};
1028 my $is_cp_ret = $opts_ref->{ret_type} eq "cp";
1029 return "( $test )" if !defined $cond;
1034 # We skip this if there are optimizations that
1035 # we can apply (below) to the individual ranges
1036 if ( ($is_cp_ret || $combine) && @ranges && ref $ranges[-1]) {
1037 if ( $ranges[-1][0] == $ranges[-1][1] ) {
1038 $ranges[-1]= $ranges[-1][0];
1039 } elsif ( $ranges[-1][0] + 1 == $ranges[-1][1] ) {
1040 $ranges[-1]= $ranges[-1][0];
1041 push @ranges, $ranges[-1] + 1;
1045 for my $condition ( @$cond ) {
1046 if ( !@ranges || $condition != $ranges[-1][1] + 1 ) {
1048 push @ranges, [ $condition, $condition ];
1055 return $self->_combine( $test, @ranges )
1061 ? "isRANGE( $test, "
1062 . $self->val_fmt($_[0]) . ", "
1063 . $self->val_fmt($_[1]) . " )"
1064 : $self->val_fmt($_) . " == $test";
1067 return "( " . join( " || ", @ranges ) . " )";
1070 # If the input set has certain characteristics, we can optimize tests
1071 # for it. This doesn't apply if returning the code point, as we want
1072 # each element of the set individually. The code above is for this
1075 return 1 if @$cond == 256; # If all bytes match, is trivially true
1080 # See if the entire set shares optimizable characteristics, and if so,
1081 # return the optimization. There is no need to do this on sets with
1082 # just a single range, as that can be expressed with a single
1084 @masks = calculate_mask(@$cond);
1086 # Stringify the output of calculate_mask()
1089 foreach my $mask_ref (@masks) {
1090 if (defined $mask_ref->[1]) {
1091 push @return, "( ( $test & "
1092 . $self->val_fmt($mask_ref->[1]) . " ) == "
1093 . $self->val_fmt($mask_ref->[0]) . " )";
1095 else { # An undefined mask means to use the value as-is
1096 push @return, "$test == " . $self->val_fmt($mask_ref->[0]);
1100 # The best possible case below for specifying this set of values via
1101 # ranges is 1 branch per range. If our mask method yielded better
1102 # results, there is no sense trying something that is bound to be
1104 if (@return < @ranges) {
1105 return "( " . join( " || ", @return ) . " )";
1112 # Here, there was no entire-class optimization that was clearly better
1113 # than doing things by ranges. Look at each range.
1114 my $range_count_extra = 0;
1115 for (my $i = 0; $i < @ranges; $i++) {
1116 if (! ref $ranges[$i]) { # Trivial case: no range
1117 $ranges[$i] = $self->val_fmt($ranges[$i]) . " == $test";
1119 elsif ($ranges[$i]->[0] == $ranges[$i]->[1]) {
1120 $ranges[$i] = # Trivial case: single element range
1121 $self->val_fmt($ranges[$i]->[0]) . " == $test";
1123 elsif ($ranges[$i]->[0] == 0) {
1124 # If the range matches all 256 possible bytes, it is trivially
1126 return 1 if $ranges[0]->[1] == 0xFF; # @ranges must be 1 in
1128 $ranges[$i] = "( $test <= "
1129 . $self->val_fmt($ranges[$i]->[1]) . " )";
1131 elsif ($ranges[$i]->[1] == 255) {
1133 # Similarly the max possible is 255, so can omit an upper bound
1134 # test if the calculated max is the max possible one.
1135 $ranges[$i] = "( $test >= " . $self->val_fmt($ranges[0]->[0]) . " )";
1140 # Well-formed UTF-8 continuation bytes on ascii platforms must be
1141 # in the range 0x80 .. 0xBF. If we know that the input is
1142 # well-formed (indicated by not trying to be 'safe'), we can omit
1143 # tests that verify that the input is within either of these
1144 # bounds. (No legal UTF-8 character can begin with anything in
1145 # this range, so we don't have to worry about this being a
1146 # continuation byte or not.)
1147 if ($opts_ref->{charset} =~ /ascii/i
1148 && (! $opts_ref->{safe} && ! $opts_ref->{no_length_checks})
1149 && $opts_ref->{type} =~ / ^ (?: utf8 | high ) $ /xi)
1151 # If the range is the entire legal range, it matches any legal
1152 # byte, so we can omit both tests. (This should happen only
1153 # if the number of ranges is 1.)
1154 if ($ranges[$i]->[0] == 0x80 && $ranges[$i]->[1] == 0xBF) {
1159 # Here, it isn't the full range of legal continuation bytes. We
1160 # could just assume that there's nothing outside of the legal
1161 # bounds. But inRANGE() allows us to have a single conditional,
1162 # so the only cost of making sure it's a legal UTF-8 continuation
1163 # byte is an extra subtraction instruction, a trivial expense.
1164 $ranges[$i] = "inRANGE($test, "
1165 . $self->val_fmt($ranges[$i]->[0]) .", "
1166 . $self->val_fmt($ranges[$i]->[1]) . ")";
1170 # We have generated the list of bytes in two ways; one trying to use masks
1171 # to cut the number of branches down, and the other to look at individual
1172 # ranges (some of which could be cut down by using a mask for just it).
1173 # We return whichever method uses the fewest branches.
1175 . join( " || ", (@masks && @masks < @ranges + $range_count_extra)
1182 # recursively turn a list of conditions into a fast break-out condition
1183 # used by _cond_as_str() for 'cp' type macros.
1185 my ( $self, $test, @cond )= @_;
1187 my $item= shift @cond;
1189 if ( ref $item ) { # @item should be a 2-element array giving range start
1191 if ($item->[0] == 0) { # UV's are never negative, so skip "0 <= "
1192 # test which could generate a compiler warning
1193 # that test is always true
1194 $cstr= "$test <= " . $self->val_fmt($item->[1]);
1197 $cstr = "inRANGE($test, "
1198 . $self->val_fmt($item->[0]) . ", "
1199 . $self->val_fmt($item->[1]) . ")";
1201 $gtv= $self->val_fmt($item->[1]);
1203 $cstr= $self->val_fmt($item) . " == $test";
1204 $gtv= $self->val_fmt($item)
1207 my $combine= $self->_combine( $test, @cond );
1209 return "( $cstr || ( $gtv < $test &&\n"
1210 . $combine . " ) )";
1212 return "( $cstr || $combine )";
1220 # recursively convert an optree to text with reasonably neat formatting
1222 my ( $self, $op, $combine, $brace, $opts_ref, $def, $submacros )= @_;
1223 return 0 if ! defined $op; # The set is empty
1227 my $cond= $self->_cond_as_str( $op, $combine, $opts_ref );
1228 #no warnings 'recursion'; # This would allow really really inefficient
1229 # code to be generated. See pod
1230 my $yes= $self->_render( $op->{yes}, $combine, 1, $opts_ref, $def, $submacros );
1231 return $yes if $cond eq '1';
1233 my $no= $self->_render( $op->{no}, $combine, 0, $opts_ref, $def, $submacros );
1234 return "( $cond )" if $yes eq '1' and $no eq '0';
1235 my ( $lb, $rb )= $brace ? ( "( ", " )" ) : ( "", "" );
1236 return "$lb$cond ? $yes : $no$rb"
1237 if !ref( $op->{yes} ) && !ref( $op->{no} );
1239 my $ind= "\n" . ( $ind1 x $op->{depth} );
1241 if ( ref $op->{yes} ) {
1242 $yes= $ind . $ind1 . $yes;
1247 my $str= "$lb$cond ?$yes$ind: $no$rb";
1248 if (length $str > 6000) {
1249 push @$submacros, sprintf "#define $def\n( %s )", "_part" . (my $yes_idx= 0+@$submacros), $yes;
1250 push @$submacros, sprintf "#define $def\n( %s )", "_part" . (my $no_idx= 0+@$submacros), $no;
1251 return sprintf "%s%s ? $def : $def%s", $lb, $cond, "_part$yes_idx", "_part$no_idx", $rb;
1256 # $expr=render($op,$combine)
1258 # convert an optree to text with reasonably neat formatting. If $combine
1259 # is true then the condition is created using "fast breakouts" which
1260 # produce uglier expressions that are more efficient for common case,
1261 # longer lists such as that resulting from type 'cp' output.
1262 # Currently only used for type 'cp' macros.
1264 my ( $self, $op, $combine, $opts_ref, $def_fmt )= @_;
1267 my $macro= sprintf "#define $def_fmt\n( %s )", "", $self->_render( $op, $combine, 0, $opts_ref, $def_fmt, \@submacros );
1269 return join "\n\n", map { "/*** GENERATED CODE ***/\n" . __macro( __clean( $_ ) ) } @submacros, $macro;
1273 # make a macro of a given type.
1274 # calls into make_trie and (generic_|length_)optree as needed
1276 # type : 'cp','cp_high', 'generic','high','low','latin1','utf8','LATIN1','UTF8'
1277 # ret_type : 'cp' or 'len'
1278 # safe : don't assume is well-formed UTF-8, so don't skip any range
1279 # checks, and add length guards to macro
1280 # no_length_checks : like safe, but don't add length guards.
1282 # type defaults to 'generic', and ret_type to 'len' unless type is 'cp'
1283 # in which case it defaults to 'cp' as well.
1285 # It is illegal to do a type 'cp' macro on a pattern with multi-codepoint
1286 # sequences in it, as the generated macro will accept only a single codepoint
1289 # It is also illegal to do a non-safe macro on a pattern with multi-codepoint
1290 # sequences in it, as even if it is known to be well-formed, we need to not
1291 # run off the end of the buffer when, say, the buffer ends with the first two
1292 # characters, but three are looked at by the macro.
1294 # returns the macro.
1300 my $type= $opts{type} || 'generic';
1301 if ($self->{has_multi}) {
1302 if ($type =~ /^cp/) {
1303 die "Can't do a 'cp' on multi-codepoint character class '$self->{op}'"
1305 elsif (! $opts{safe}) {
1306 die "'safe' is required on multi-codepoint character class '$self->{op}'"
1309 my $ret_type= $opts{ret_type} || ( $opts{type} =~ /^cp/ ? 'cp' : 'len' );
1311 if ( $opts{safe} ) {
1312 $method= 'length_optree';
1313 } elsif ( $type =~ /generic/ ) {
1314 $method= 'generic_optree';
1318 my @args= $type =~ /^cp/ ? 'cp' : 's';
1319 push @args, "e" if $opts{safe};
1320 push @args, "is_utf8" if $type =~ /generic/;
1321 push @args, "len" if $ret_type eq 'both';
1322 my $pfx= $ret_type eq 'both' ? 'what_len_' :
1323 $ret_type eq 'cp' ? 'what_' : 'is_';
1324 my $ext= $type =~ /generic/ ? '' : '_' . lc( $type );
1325 $ext .= '_non_low' if $type eq 'generic_non_low';
1326 $ext .= "_safe" if $opts{safe};
1327 $ext .= "_no_length_checks" if $opts{no_length_checks};
1328 my $argstr= join ",", @args;
1329 my $def_fmt="$pfx$self->{op}$ext%s($argstr)";
1330 my $optree= $self->$method( %opts, type => $type, ret_type => $ret_type );
1331 return $self->render( $optree, ($type =~ /^cp/) ? 1 : 0, \%opts, $def_fmt );
1334 # if we aren't being used as a module (highly likely) then process
1335 # the __DATA__ below and produce macros in regcharclass.h
1336 # if an argument is provided to the script then it is assumed to
1337 # be the path of the file to output to, if the arg is '-' outputs
1341 my $path= shift @ARGV || "regcharclass.h";
1343 if ( $path eq '-' ) {
1346 $out_fh = open_new( $path );
1348 print $out_fh read_only_top( lang => 'C', by => $0,
1349 file => 'regcharclass.h', style => '*',
1350 copyright => [2007, 2011],
1352 WARNING: These macros are for internal Perl core use only, and may be
1353 changed or removed without notice.
1356 print $out_fh "\n#ifndef PERL_REGCHARCLASS_H_ /* Guard against nested #includes */\n#define PERL_REGCHARCLASS_H_\n";
1358 my ( $op, $title, @txt, @types, %mods );
1362 my $charset = shift;
1364 # Skip if to compile on a different platform.
1365 return if delete $mods{only_ascii_platform} && $charset !~ /ascii/i;
1366 return if delete $mods{only_ebcdic_platform} && $charset !~ /ebcdic/i;
1368 print $out_fh "/*\n\t$op: $title\n\n";
1369 print $out_fh join "\n", ( map { "\t$_" } @txt ), "*/", "";
1370 my $obj= __PACKAGE__->new( op => $op, title => $title, txt => \@txt, charset => $charset);
1372 #die Dumper(\@types,\%mods);
1375 push @mods, 'safe' if delete $mods{safe};
1376 push @mods, 'no_length_checks' if delete $mods{no_length_checks};
1377 unshift @mods, 'fast' if delete $mods{fast} || ! @mods; # Default to 'fast'
1382 die "Unknown modifiers: ", join ", ", map { "'$_'" } sort keys %mods;
1385 foreach my $type_spec ( @types ) {
1386 my ( $type, $ret )= split /-/, $type_spec;
1388 foreach my $mod ( @mods ) {
1390 # 'safe' is irrelevant with code point macros, so skip if
1391 # there is also a 'fast', but don't skip if this is the only
1392 # way a cp macro will get generated. Below we convert 'safe'
1393 # to 'fast' in this instance
1394 next if $type =~ /^cp/
1395 && ($mod eq 'safe' || $mod eq 'no_length_checks')
1396 && grep { 'fast' =~ $_ } @mods;
1398 my $macro= $obj->make_macro(
1401 safe => $mod eq 'safe' && $type !~ /^cp/,
1402 charset => $charset,
1403 no_length_checks => $mod eq 'no_length_checks' && $type !~ /^cp/,
1405 print $out_fh $macro, "\n";
1411 foreach my $charset (get_supported_code_pages()) {
1418 print $out_fh "\n", get_conditional_compile_line_start($charset);
1419 my @data_copy = @data;
1421 s/^ \s* (?: \# .* ) ? $ //x; # squeeze out comment and blanks
1425 $doit->($charset) unless $first_time; # This starts a new
1426 # definition; do the
1429 ( $op, $title )= split /\s*:\s*/, $_, 2;
1431 } elsif ( s/^=>// ) {
1432 my ( $type, $modifier )= split /:/, $_;
1433 @types= split ' ', $type;
1435 map { $mods{$_} = 1 } split ' ', $modifier;
1441 print $out_fh get_conditional_compile_line_end();
1444 print $out_fh "\n#endif /* PERL_REGCHARCLASS_H_ */\n";
1447 print $out_fh "/* ex: set ro: */\n";
1449 # Some of the sources for these macros come from Unicode tables
1450 my $sources_list = "lib/unicore/mktables.lst";
1451 my @sources = ($0, qw(lib/unicore/mktables
1453 regen/regcharclass_multi_char_folds.pl
1454 regen/charset_translations.pl
1457 # Depend on mktables’ own sources. It’s a shorter list of files than
1458 # those that Unicode::UCD uses.
1459 if (! open my $mktables_list, '<', $sources_list) {
1461 # This should force a rebuild once $sources_list exists
1462 push @sources, $sources_list;
1465 while(<$mktables_list>) {
1468 push @sources, "lib/unicore/$_" if /^[^#]/;
1472 read_only_bottom_close_and_rename($out_fh, \@sources)
1476 # The form of the input is a series of definitions to make macros for.
1477 # The first line gives the base name of the macro, followed by a colon, and
1478 # then text to be used in comments associated with the macro that are its
1479 # title or description. In all cases the first (perhaps only) parameter to
1480 # the macro is a pointer to the first byte of the code point it is to test to
1481 # see if it is in the class determined by the macro. In the case of non-UTF8,
1482 # the code point consists only of a single byte.
1484 # The second line must begin with a '=>' and be followed by the types of
1485 # macro(s) to be generated; these are specified below. A colon follows the
1486 # types, followed by the modifiers, also specified below. At least one
1487 # modifier is required.
1489 # The subsequent lines give what code points go into the class defined by the
1490 # macro. Multiple characters may be specified via a string like "\x0D\x0A",
1491 # enclosed in quotes. Otherwise the lines consist of one of:
1492 # 1) a single Unicode code point, prefaced by 0x
1493 # 2) a single range of Unicode code points separated by a minus (and
1495 # 3) a single Unicode property specified in the standard Perl form
1497 # 4) a line like 'do path'. This will do a 'do' on the file given by
1498 # 'path'. It is assumed that this does nothing but load subroutines
1499 # (See item 5 below). The reason 'require path' is not used instead is
1500 # because 'do' doesn't assume that path is in @INC.
1501 # 5) a subroutine call
1502 # &pkg::foo(arg1, ...)
1503 # where pkg::foo was loaded by a 'do' line (item 4). The subroutine
1504 # returns an array of entries of forms like items 1-3 above. This
1505 # allows more complex inputs than achievable from the other input types.
1507 # A blank line or one whose first non-blank character is '#' is a comment.
1508 # The definition of the macro is terminated by a line unlike those described.
1511 # low generate a macro whose name is 'is_BASE_low' and defines a
1512 # class that includes only ASCII-range chars. (BASE is the
1513 # input macro base name.)
1514 # latin1 generate a macro whose name is 'is_BASE_latin1' and defines a
1515 # class that includes only upper-Latin1-range chars. It is not
1516 # designed to take a UTF-8 input parameter.
1517 # high generate a macro whose name is 'is_BASE_high' and defines a
1518 # class that includes all relevant code points that are above
1519 # the Latin1 range. This is for very specialized uses only.
1520 # It is designed to take only an input UTF-8 parameter.
1521 # utf8 generate a macro whose name is 'is_BASE_utf8' and defines a
1522 # class that includes all relevant characters that aren't ASCII.
1523 # It is designed to take only an input UTF-8 parameter.
1524 # LATIN1 generate a macro whose name is 'is_BASE_latin1' and defines a
1525 # class that includes both ASCII and upper-Latin1-range chars.
1526 # It is not designed to take a UTF-8 input parameter.
1527 # UTF8 generate a macro whose name is 'is_BASE_utf8' and defines a
1528 # class that can include any code point, adding the 'low' ones
1529 # to what 'utf8' works on. It is designed to take only an input
1531 # generic generate a macro whose name is 'is_BASE". It has a 2nd,
1532 # boolean, parameter which indicates if the first one points to
1533 # a UTF-8 string or not. Thus it works in all circumstances.
1534 # generic_non_low generate a macro whose name is 'is_BASE_non_low". It has
1535 # a 2nd, boolean, parameter which indicates if the first one
1536 # points to a UTF-8 string or not. It excludes any ASCII-range
1537 # matches, but otherwise it works in all circumstances.
1538 # cp generate a macro whose name is 'is_BASE_cp' and defines a
1539 # class that returns true if the UV parameter is a member of the
1540 # class; false if not.
1541 # cp_high like cp, but it is assumed that it is known that the UV
1542 # parameter is above Latin1. The name of the generated macro is
1543 # 'is_BASE_cp_high'. This is different from high-cp, derived
1545 # A macro of the given type is generated for each type listed in the input.
1546 # The default return value is the number of octets read to generate the match.
1547 # Append "-cp" to the type to have it instead return the matched codepoint.
1548 # The macro name is changed to 'what_BASE...'. See pod for
1550 # Appending '-both" instead adds an extra parameter to the end of the argument
1551 # list, which is a pointer as to where to store the number of
1552 # bytes matched, while also returning the code point. The macro
1553 # name is changed to 'what_len_BASE...'. See pod for caveats
1556 # safe The input string is not necessarily valid UTF-8. In
1557 # particular an extra parameter (always the 2nd) to the macro is
1558 # required, which points to one beyond the end of the string.
1559 # The macro will make sure not to read off the end of the
1560 # string. In the case of non-UTF8, it makes sure that the
1561 # string has at least one byte in it. The macro name has
1562 # '_safe' appended to it.
1563 # no_length_checks The input string is not necessarily valid UTF-8, but it
1564 # is to be assumed that the length has already been checked and
1566 # fast The input string is valid UTF-8. No bounds checking is done,
1567 # and the macro can make assumptions that lead to faster
1569 # only_ascii_platform Skip this definition if the character set is for
1570 # a non-ASCII platform.
1571 # only_ebcdic_platform Skip this definition if the character set is for
1572 # a non-EBCDIC platform.
1573 # No modifier need be specified; fast is assumed for this case. If both
1574 # 'fast', and 'safe' are specified, two macros will be created for each
1577 # If run on a non-ASCII platform will automatically convert the Unicode input
1578 # to native. The documentation above is slightly wrong in this case. 'low'
1579 # actually refers to code points whose UTF-8 representation is the same as the
1580 # non-UTF-8 version (invariants); and 'latin1' refers to all the rest of the
1581 # code points less than 256.
1583 1; # in the unlikely case we are being used as a module
1586 # This is no longer used, but retained in case it is needed some day.
1587 # TRICKYFOLD: Problematic fold case letters. When adding to this list, also should add them to regcomp.c and fold_grind.t
1588 # => generic cp generic-cp generic-both :fast safe
1589 # 0x00DF # LATIN SMALL LETTER SHARP S
1590 # 0x0390 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
1591 # 0x03B0 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
1592 # 0x1E9E # LATIN CAPITAL LETTER SHARP S, because maps to same as 00DF
1593 # 0x1FD3 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND OXIA; maps same as 0390
1594 # 0x1FE3 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND OXIA; maps same as 03B0
1596 LNBREAK: Line Break: \R
1597 => generic UTF8 LATIN1 : safe
1598 "\x0D\x0A" # CRLF - Network (Windows) line ending
1601 HORIZWS: Horizontal Whitespace: \h \H
1602 => high cp_high : fast
1605 VERTWS: Vertical Whitespace: \v \V
1606 => high cp_high : fast
1609 XDIGIT: Hexadecimal digits
1610 => high cp_high : fast
1613 XPERLSPACE: \p{XPerlSpace}
1614 => high cp_high : fast
1617 NONCHAR: Non character code points
1621 SURROGATE: Surrogate code points
1625 QUOTEMETA: Meta-characters that \Q should quote
1629 MULTI_CHAR_FOLD: multi-char strings that are folded to by a single character
1631 ®charclass_multi_char_folds::multi_char_folds('u', 'a')
1633 MULTI_CHAR_FOLD: multi-char strings that are folded to by a single character
1635 ®charclass_multi_char_folds::multi_char_folds('l', 'a')
1637 THREE_CHAR_FOLD: A three-character multi-char fold
1639 ®charclass_multi_char_folds::multi_char_folds('u', '3')
1641 THREE_CHAR_FOLD: A three-character multi-char fold
1643 ®charclass_multi_char_folds::multi_char_folds('l', '3')
1645 THREE_CHAR_FOLD_HEAD: The first two of three-character multi-char folds
1647 ®charclass_multi_char_folds::multi_char_folds('u', 'h')
1649 THREE_CHAR_FOLD_HEAD: The first two of three-character multi-char folds
1651 ®charclass_multi_char_folds::multi_char_folds('l', 'h')
1653 #THREE_CHAR_FOLD_NON_FINAL: The first or middle character of multi-char folds
1655 #®charclass_multi_char_folds::multi_char_folds('u', 'fm')
1657 #THREE_CHAR_FOLD_NON_FINAL: The first or middle character of multi-char folds
1659 #®charclass_multi_char_folds::multi_char_folds('l', 'fm')
1661 FOLDS_TO_MULTI: characters that fold to multi-char strings
1663 \p{_Perl_Folds_To_Multi_Char}
1665 PROBLEMATIC_LOCALE_FOLD : characters whose fold is problematic under locale
1667 \p{_Perl_Problematic_Locale_Folds}
1669 PROBLEMATIC_LOCALE_FOLDEDS_START : The first folded character of folds which are problematic under locale
1671 \p{_Perl_Problematic_Locale_Foldeds_Start}
1673 PATWS: pattern white space
1677 HANGUL_ED: Hangul syllables whose first character is \xED
1678 => UTF8 :only_ascii_platform safe