1 # EXTRACT VARIOUSLY DELIMITED TEXT SEQUENCES FROM STRINGS.
2 # FOR FULL DOCUMENTATION SEE Balanced.pod
7 package Text::Balanced;
10 use vars qw { $VERSION @ISA %EXPORT_TAGS };
13 @ISA = qw ( Exporter );
15 %EXPORT_TAGS = ( ALL => [ qw(
30 Exporter::export_ok_tags('ALL');
34 sub _match_bracketed($$$$$$);
35 sub _match_variable($$);
36 sub _match_codeblock($$$$$$$);
37 sub _match_quotelike($$$$);
40 require Carp; goto &Carp::carp;
44 require Carp; goto &Carp::croak;
47 # HANDLE RETURN VALUES IN VARIOUS CONTEXTS
50 my ($message, $pos) = @_;
51 $@ = bless { error=>$message, pos=>$pos }, "Text::Balanced::ErrorMsg";
56 my ($wantarray, $textref, $message, $pos) = @_;
57 _failmsg $message, $pos if $message;
58 return (undef,$$textref,undef) if $wantarray;
65 my ($wantarray,$textref) = splice @_, 0, 2;
66 my ($extrapos, $extralen) = @_>18 ? splice(@_, -2, 2) : (0,0);
67 my ($startlen) = $_[5];
69 my $remainderpos = $_[2];
73 while (my ($from, $len) = splice @_, 0, 2)
75 push @res, substr($$textref,$from,$len);
77 if ($extralen) { # CORRECT FILLET
78 my $extra = substr($res[0], $extrapos-$oppos, $extralen, "\n");
79 $res[1] = "$extra$res[1]";
80 eval { substr($$textref,$remainderpos,0) = $extra;
81 substr($$textref,$extrapos,$extralen,"\n")} ;
82 #REARRANGE HERE DOC AND FILLET IF POSSIBLE
83 pos($$textref) = $remainderpos-$extralen+1; # RESET \G
86 pos($$textref) = $remainderpos; # RESET \G
92 my $match = substr($$textref,$_[0],$_[1]);
93 substr($match,$extrapos-$_[0]-$startlen,$extralen,"") if $extralen;
95 ? substr($$textref, $extrapos, $extralen)."\n" : "";
96 eval {substr($$textref,$_[4],$_[1]+$_[5])=$extra} ; #CHOP OUT PREFIX & MATCH, IF POSSIBLE
97 pos($$textref) = $_[4]; # RESET \G
102 # BUILD A PATTERN MATCHING A SIMPLE DELIMITED STRING
104 sub gen_delimited_pat($;$) # ($delimiters;$escapes)
106 my ($dels, $escs) = @_;
107 return "" unless $dels =~ /\S/;
108 $escs = '\\' unless $escs;
109 $escs .= substr($escs,-1) x (length($dels)-length($escs));
112 for ($i=0; $i<length $dels; $i++)
114 my $del = quotemeta substr($dels,$i,1);
115 my $esc = quotemeta substr($escs,$i,1);
118 push @pat, "$del(?:[^$del]*(?:(?:$del$del)[^$del]*)*)$del";
122 push @pat, "$del(?:[^$esc$del]*(?:$esc.[^$esc$del]*)*)$del";
125 my $pat = join '|', @pat;
129 *delimited_pat = \&gen_delimited_pat;
132 # THE EXTRACTION FUNCTIONS
134 sub extract_delimited (;$$$$)
136 my $textref = defined $_[0] ? \$_[0] : \$_;
137 my $wantarray = wantarray;
138 my $del = defined $_[1] ? $_[1] : qq{\'\"\`};
139 my $pre = defined $_[2] ? $_[2] : '\s*';
140 my $esc = defined $_[3] ? $_[3] : qq{\\};
141 my $pat = gen_delimited_pat($del, $esc);
142 my $startpos = pos $$textref || 0;
143 return _fail($wantarray, $textref, "Not a delimited pattern", 0)
144 unless $$textref =~ m/\G($pre)($pat)/gc;
145 my $prelen = length($1);
146 my $matchpos = $startpos+$prelen;
147 my $endpos = pos $$textref;
148 return _succeed $wantarray, $textref,
149 $matchpos, $endpos-$matchpos, # MATCH
150 $endpos, length($$textref)-$endpos, # REMAINDER
151 $startpos, $prelen; # PREFIX
154 sub extract_bracketed (;$$$)
156 my $textref = defined $_[0] ? \$_[0] : \$_;
157 my $ldel = defined $_[1] ? $_[1] : '{([<';
158 my $pre = defined $_[2] ? $_[2] : '\s*';
159 my $wantarray = wantarray;
162 $ldel =~ s/'//g and $qdel .= q{'};
163 $ldel =~ s/"//g and $qdel .= q{"};
164 $ldel =~ s/`//g and $qdel .= q{`};
165 $ldel =~ s/q//g and $quotelike = 1;
166 $ldel =~ tr/[](){}<>\0-\377/[[(({{<</ds;
168 unless ($rdel =~ tr/[({</])}>/)
170 return _fail $wantarray, $textref,
171 "Did not find a suitable bracket in delimiter: \"$_[1]\"",
175 $ldel = join('|', map { quotemeta $_ } split('', $ldel));
176 $rdel = join('|', map { quotemeta $_ } split('', $rdel));
179 my $startpos = pos $$textref || 0;
180 my @match = _match_bracketed($textref,$pre, $ldel, $qdel, $quotelike, $rdel);
182 return _fail ($wantarray, $textref) unless @match;
184 return _succeed ( $wantarray, $textref,
185 $match[2], $match[5]+2, # MATCH
186 @match[8,9], # REMAINDER
187 @match[0,1], # PREFIX
191 sub _match_bracketed($$$$$$) # $textref, $pre, $ldel, $qdel, $quotelike, $rdel
193 my ($textref, $pre, $ldel, $qdel, $quotelike, $rdel) = @_;
194 my ($startpos, $ldelpos, $endpos) = (pos $$textref = pos $$textref||0);
195 unless ($$textref =~ m/\G$pre/gc)
197 _failmsg "Did not find prefix: /$pre/", $startpos;
201 $ldelpos = pos $$textref;
203 unless ($$textref =~ m/\G($ldel)/gc)
205 _failmsg "Did not find opening bracket after prefix: \"$pre\"",
207 pos $$textref = $startpos;
211 my @nesting = ( $1 );
212 my $textlen = length $$textref;
213 while (pos $$textref < $textlen)
215 next if $$textref =~ m/\G\\./gcs;
217 if ($$textref =~ m/\G($ldel)/gc)
221 elsif ($$textref =~ m/\G($rdel)/gc)
223 my ($found, $brackettype) = ($1, $1);
226 _failmsg "Unmatched closing bracket: \"$found\"",
228 pos $$textref = $startpos;
231 my $expected = pop(@nesting);
232 $expected =~ tr/({[</)}]>/;
233 if ($expected ne $brackettype)
235 _failmsg qq{Mismatched closing bracket: expected "$expected" but found "$found"},
237 pos $$textref = $startpos;
240 last if $#nesting < 0;
242 elsif ($qdel && $$textref =~ m/\G([$qdel])/gc)
244 $$textref =~ m/\G[^\\$1]*(?:\\.[^\\$1]*)*(\Q$1\E)/gsc and next;
245 _failmsg "Unmatched embedded quote ($1)",
247 pos $$textref = $startpos;
250 elsif ($quotelike && _match_quotelike($textref,"",1,0))
255 else { $$textref =~ m/\G(?:[a-zA-Z0-9]+|.)/gcs }
259 _failmsg "Unmatched opening bracket(s): "
260 . join("..",@nesting)."..",
262 pos $$textref = $startpos;
266 $endpos = pos $$textref;
269 $startpos, $ldelpos-$startpos, # PREFIX
270 $ldelpos, 1, # OPENING BRACKET
271 $ldelpos+1, $endpos-$ldelpos-2, # CONTENTS
272 $endpos-1, 1, # CLOSING BRACKET
273 $endpos, length($$textref)-$endpos, # REMAINDER
279 my $brack = reverse $_[0];
280 $brack =~ tr/[({</])}>/;
284 my $XMLNAME = q{[a-zA-Z_:][a-zA-Z0-9_:.-]*};
286 sub extract_tagged (;$$$$$) # ($text, $opentag, $closetag, $pre, \%options)
288 my $textref = defined $_[0] ? \$_[0] : \$_;
291 my $pre = defined $_[3] ? $_[3] : '\s*';
292 my %options = defined $_[4] ? %{$_[4]} : ();
293 my $omode = defined $options{fail} ? $options{fail} : '';
294 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
295 : defined($options{reject}) ? $options{reject}
298 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
299 : defined($options{ignore}) ? $options{ignore}
303 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
306 my @match = _match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
308 return _fail(wantarray, $textref) unless @match;
309 return _succeed wantarray, $textref,
310 $match[2], $match[3]+$match[5]+$match[7], # MATCH
311 @match[8..9,0..1,2..7]; # REM, PRE, BITS
314 sub _match_tagged # ($$$$$$$)
316 my ($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore) = @_;
319 my ($startpos, $opentagpos, $textpos, $parapos, $closetagpos, $endpos) = ( pos($$textref) = pos($$textref)||0 );
321 unless ($$textref =~ m/\G($pre)/gc)
323 _failmsg "Did not find prefix: /$pre/", pos $$textref;
327 $opentagpos = pos($$textref);
329 unless ($$textref =~ m/\G$ldel/gc)
331 _failmsg "Did not find opening tag: /$ldel/", pos $$textref;
335 $textpos = pos($$textref);
340 unless ($rdelspec =~ s/\A([[(<{]+)($XMLNAME).*/ quotemeta "$1\/$2". _revbracket($1) /oes)
342 _failmsg "Unable to construct closing tag to match: $rdel",
349 $rdelspec = eval "qq{$rdel}" || do {
351 for (qw,~ ! ^ & * ) _ + - = } ] : " ; ' > . ? / | ',)
352 { next if $rdel =~ /\Q$_/; $del = $_; last }
354 croak ("Can't interpolate right delimiter $rdel")
356 eval "qq$del$rdel$del";
360 while (pos($$textref) < length($$textref))
362 next if $$textref =~ m/\G\\./gc;
364 if ($$textref =~ m/\G(\n[ \t]*\n)/gc )
366 $parapos = pos($$textref) - length($1)
367 unless defined $parapos;
369 elsif ($$textref =~ m/\G($rdelspec)/gc )
371 $closetagpos = pos($$textref)-length($1);
374 elsif ($ignore && $$textref =~ m/\G(?:$ignore)/gc)
378 elsif ($bad && $$textref =~ m/\G($bad)/gcs)
380 pos($$textref) -= length($1); # CUT OFF WHATEVER CAUSED THE SHORTNESS
381 goto short if ($omode eq 'PARA' || $omode eq 'MAX');
382 _failmsg "Found invalid nested tag: $1", pos $$textref;
385 elsif ($$textref =~ m/\G($ldel)/gc)
388 pos($$textref) -= length($tag); # REWIND TO NESTED TAG
389 unless (_match_tagged(@_)) # MATCH NESTED TAG
391 goto short if $omode eq 'PARA' || $omode eq 'MAX';
392 _failmsg "Found unbalanced nested tag: $tag",
397 else { $$textref =~ m/./gcs }
401 $closetagpos = pos($$textref);
402 goto matched if $omode eq 'MAX';
403 goto failed unless $omode eq 'PARA';
405 if (defined $parapos) { pos($$textref) = $parapos }
406 else { $parapos = pos($$textref) }
409 $startpos, $opentagpos-$startpos, # PREFIX
410 $opentagpos, $textpos-$opentagpos, # OPENING TAG
411 $textpos, $parapos-$textpos, # TEXT
412 $parapos, 0, # NO CLOSING TAG
413 $parapos, length($$textref)-$parapos, # REMAINDER
417 $endpos = pos($$textref);
419 $startpos, $opentagpos-$startpos, # PREFIX
420 $opentagpos, $textpos-$opentagpos, # OPENING TAG
421 $textpos, $closetagpos-$textpos, # TEXT
422 $closetagpos, $endpos-$closetagpos, # CLOSING TAG
423 $endpos, length($$textref)-$endpos, # REMAINDER
427 _failmsg "Did not find closing tag", pos $$textref unless $@;
428 pos($$textref) = $startpos;
432 sub extract_variable (;$$)
434 my $textref = defined $_[0] ? \$_[0] : \$_;
435 return ("","","") unless defined $$textref;
436 my $pre = defined $_[1] ? $_[1] : '\s*';
438 my @match = _match_variable($textref,$pre);
440 return _fail wantarray, $textref unless @match;
442 return _succeed wantarray, $textref,
443 @match[2..3,4..5,0..1]; # MATCH, REMAINDER, PREFIX
446 sub _match_variable($$)
451 my ($textref, $pre) = @_;
452 my $startpos = pos($$textref) = pos($$textref)||0;
453 unless ($$textref =~ m/\G($pre)/gc)
455 _failmsg "Did not find prefix: /$pre/", pos $$textref;
458 my $varpos = pos($$textref);
459 unless ($$textref =~ m{\G\$\s*(?!::)(\d+|[][&`'+*./|,";%=~:?!\@<>()-]|\^[a-z]?)}gci)
461 unless ($$textref =~ m/\G((\$#?|[*\@\%]|\\&)+)/gc)
463 _failmsg "Did not find leading dereferencer", pos $$textref;
464 pos $$textref = $startpos;
469 unless ($$textref =~ m/\G\s*(?:::|')?(?:[_a-z]\w*(?:::|'))*[_a-z]\w*/gci
470 or _match_codeblock($textref, "", '\{', '\}', '\{', '\}', 0)
471 or $deref eq '$#' or $deref eq '$$' )
473 _failmsg "Bad identifier after dereferencer", pos $$textref;
474 pos $$textref = $startpos;
481 next if $$textref =~ m/\G\s*(?:->)?\s*[{]\w+[}]/gc;
482 next if _match_codeblock($textref,
483 qr/\s*->\s*(?:[_a-zA-Z]\w+\s*)?/,
484 qr/[({[]/, qr/[)}\]]/,
485 qr/[({[]/, qr/[)}\]]/, 0);
486 next if _match_codeblock($textref,
487 qr/\s*/, qr/[{[]/, qr/[}\]]/,
488 qr/[{[]/, qr/[}\]]/, 0);
489 next if _match_variable($textref,'\s*->\s*');
490 next if $$textref =~ m/\G\s*->\s*\w+(?![{([])/gc;
494 my $endpos = pos($$textref);
495 return ($startpos, $varpos-$startpos,
496 $varpos, $endpos-$varpos,
497 $endpos, length($$textref)-$endpos
501 sub extract_codeblock (;$$$$$)
503 my $textref = defined $_[0] ? \$_[0] : \$_;
504 my $wantarray = wantarray;
505 my $ldel_inner = defined $_[1] ? $_[1] : '{';
506 my $pre = defined $_[2] ? $_[2] : '\s*';
507 my $ldel_outer = defined $_[3] ? $_[3] : $ldel_inner;
509 my $rdel_inner = $ldel_inner;
510 my $rdel_outer = $ldel_outer;
512 for ($ldel_inner, $ldel_outer) { tr/[]()<>{}\0-\377/[[((<<{{/ds }
513 for ($rdel_inner, $rdel_outer) { tr/[]()<>{}\0-\377/]]))>>}}/ds }
514 for ($ldel_inner, $ldel_outer, $rdel_inner, $rdel_outer)
516 $_ = '('.join('|',map { quotemeta $_ } split('',$_)).')'
520 my @match = _match_codeblock($textref, $pre,
521 $ldel_outer, $rdel_outer,
522 $ldel_inner, $rdel_inner,
524 return _fail($wantarray, $textref) unless @match;
525 return _succeed($wantarray, $textref,
526 @match[2..3,4..5,0..1] # MATCH, REMAINDER, PREFIX
531 sub _match_codeblock($$$$$$$)
533 my ($textref, $pre, $ldel_outer, $rdel_outer, $ldel_inner, $rdel_inner, $rd) = @_;
534 my $startpos = pos($$textref) = pos($$textref) || 0;
535 unless ($$textref =~ m/\G($pre)/gc)
537 _failmsg qq{Did not match prefix /$pre/ at"} .
538 substr($$textref,pos($$textref),20) .
543 my $codepos = pos($$textref);
544 unless ($$textref =~ m/\G($ldel_outer)/gc) # OUTERMOST DELIMITER
546 _failmsg qq{Did not find expected opening bracket at "} .
547 substr($$textref,pos($$textref),20) .
550 pos $$textref = $startpos;
554 $closing =~ tr/([<{/)]>}/;
557 while (pos($$textref) < length($$textref))
560 if ($rd && $$textref =~ m#\G(\Q(?)\E|\Q(s?)\E|\Q(s)\E)#gc)
566 if ($$textref =~ m/\G\s*#.*/gc)
571 if ($$textref =~ m/\G\s*($rdel_outer)/gc)
573 unless ($matched = ($closing && $1 eq $closing) )
575 next if $1 eq '>'; # MIGHT BE A "LESS THAN"
576 _failmsg q{Mismatched closing bracket at "} .
577 substr($$textref,pos($$textref),20) .
578 qq{...". Expected '$closing'},
584 if (_match_variable($textref,'\s*') ||
585 _match_quotelike($textref,'\s*',$patvalid,$patvalid) )
592 # NEED TO COVER MANY MORE CASES HERE!!!
593 # NB 'case' is included here, because in Switch.pm,
594 # it's followed by a term, not an op
596 if ($$textref =~ m#\G\s*(?!$ldel_inner)
600 | (\*\*|&&|\|\||<<|>>)=?
601 | case|split|grep|map|return
609 if ( _match_codeblock($textref, '\s*', $ldel_inner, $rdel_inner, $ldel_inner, $rdel_inner, $rd) )
615 if ($$textref =~ m/\G\s*$ldel_outer/gc)
617 _failmsg q{Improperly nested codeblock at "} .
618 substr($$textref,pos($$textref),20) .
625 $$textref =~ m/\G\s*(\w+|[-=>]>|.|\Z)/gc;
627 continue { $@ = undef }
631 _failmsg 'No match found for opening bracket', pos $$textref
636 my $endpos = pos($$textref);
637 return ( $startpos, $codepos-$startpos,
638 $codepos, $endpos-$codepos,
639 $endpos, length($$textref)-$endpos,
645 'none' => '[cgimsox]*',
647 's' => '[cegimsox]*',
657 sub extract_quotelike (;$$)
659 my $textref = $_[0] ? \$_[0] : \$_;
660 my $wantarray = wantarray;
661 my $pre = defined $_[1] ? $_[1] : '\s*';
663 my @match = _match_quotelike($textref,$pre,1,0);
664 return _fail($wantarray, $textref) unless @match;
665 return _succeed($wantarray, $textref,
666 $match[2], $match[18]-$match[2], # MATCH
667 @match[18,19], # REMAINDER
668 @match[0,1], # PREFIX
669 @match[2..17], # THE BITS
670 @match[20,21], # ANY FILLET?
674 sub _match_quotelike($$$$) # ($textref, $prepat, $allow_raw_match)
676 my ($textref, $pre, $rawmatch, $qmark) = @_;
678 my ($textlen,$startpos,
680 $preld1pos,$ld1pos,$str1pos,$rd1pos,
681 $preld2pos,$ld2pos,$str2pos,$rd2pos,
682 $modpos) = ( length($$textref), pos($$textref) = pos($$textref) || 0 );
684 unless ($$textref =~ m/\G($pre)/gc)
686 _failmsg qq{Did not find prefix /$pre/ at "} .
687 substr($$textref, pos($$textref), 20) .
692 $oppos = pos($$textref);
694 my $initial = substr($$textref,$oppos,1);
696 if ($initial && $initial =~ m|^[\"\'\`]|
697 || $rawmatch && $initial =~ m|^/|
698 || $qmark && $initial =~ m|^\?|)
700 unless ($$textref =~ m/ \Q$initial\E [^\\$initial]* (\\.[^\\$initial]*)* \Q$initial\E /gcsx)
702 _failmsg qq{Did not find closing delimiter to match '$initial' at "} .
703 substr($$textref, $oppos, 20) .
706 pos $$textref = $startpos;
709 $modpos= pos($$textref);
712 if ($initial eq '/' || $initial eq '?')
714 $$textref =~ m/\G$mods{none}/gc
717 my $endpos = pos($$textref);
719 $startpos, $oppos-$startpos, # PREFIX
720 $oppos, 0, # NO OPERATOR
721 $oppos, 1, # LEFT DEL
722 $oppos+1, $rd1pos-$oppos-1, # STR/PAT
723 $rd1pos, 1, # RIGHT DEL
724 $modpos, 0, # NO 2ND LDEL
725 $modpos, 0, # NO 2ND STR
726 $modpos, 0, # NO 2ND RDEL
727 $modpos, $endpos-$modpos, # MODIFIERS
728 $endpos, $textlen-$endpos, # REMAINDER
732 unless ($$textref =~ m{\G(\b(?:m|s|qq|qx|qw|q|qr|tr|y)\b(?=\s*\S)|<<)}gc)
734 _failmsg q{No quotelike operator found after prefix at "} .
735 substr($$textref, pos($$textref), 20) .
738 pos $$textref = $startpos;
743 $preld1pos = pos($$textref);
745 $ld1pos = pos($$textref);
747 if ($$textref =~ m{\G([A-Za-z_]\w*)}gc) {
750 elsif ($$textref =~ m{ \G ' ([^'\\]* (?:\\.[^'\\]*)*) '
751 | \G " ([^"\\]* (?:\\.[^"\\]*)*) "
752 | \G ` ([^`\\]* (?:\\.[^`\\]*)*) `
759 my $extrapos = pos($$textref);
760 $$textref =~ m{.*\n}gc;
761 $str1pos = pos($$textref)--;
762 unless ($$textref =~ m{.*?\n(?=\Q$label\E\n)}gc) {
763 _failmsg qq{Missing here doc terminator ('$label') after "} .
764 substr($$textref, $startpos, 20) .
767 pos $$textref = $startpos;
770 $rd1pos = pos($$textref);
771 $$textref =~ m{\Q$label\E\n}gc;
772 $ld2pos = pos($$textref);
774 $startpos, $oppos-$startpos, # PREFIX
775 $oppos, length($op), # OPERATOR
776 $ld1pos, $extrapos-$ld1pos, # LEFT DEL
777 $str1pos, $rd1pos-$str1pos, # STR/PAT
778 $rd1pos, $ld2pos-$rd1pos, # RIGHT DEL
779 $ld2pos, 0, # NO 2ND LDEL
780 $ld2pos, 0, # NO 2ND STR
781 $ld2pos, 0, # NO 2ND RDEL
782 $ld2pos, 0, # NO MODIFIERS
783 $ld2pos, $textlen-$ld2pos, # REMAINDER
784 $extrapos, $str1pos-$extrapos, # FILLETED BIT
788 $$textref =~ m/\G\s*/gc;
789 $ld1pos = pos($$textref);
790 $str1pos = $ld1pos+1;
792 unless ($$textref =~ m/\G(\S)/gc) # SHOULD USE LOOKAHEAD
794 _failmsg "No block delimiter found after quotelike $op",
796 pos $$textref = $startpos;
799 pos($$textref) = $ld1pos; # HAVE TO DO THIS BECAUSE LOOKAHEAD BROKEN
800 my ($ldel1, $rdel1) = ("\Q$1","\Q$1");
801 if ($ldel1 =~ /[[(<{]/)
803 $rdel1 =~ tr/[({</])}>/;
804 defined(_match_bracketed($textref,"",$ldel1,"","",$rdel1))
805 || do { pos $$textref = $startpos; return };
809 $$textref =~ /$ldel1[^\\$ldel1]*(\\.[^\\$ldel1]*)*$ldel1/gcs
810 || do { pos $$textref = $startpos; return };
812 $ld2pos = $rd1pos = pos($$textref)-1;
814 my $second_arg = $op =~ /s|tr|y/ ? 1 : 0;
818 if ($ldel1 =~ /[[(<{]/)
820 unless ($$textref =~ /\G\s*(\S)/gc) # SHOULD USE LOOKAHEAD
822 _failmsg "Missing second block for quotelike $op",
824 pos $$textref = $startpos;
827 $ldel2 = $rdel2 = "\Q$1";
828 $rdel2 =~ tr/[({</])}>/;
832 $ldel2 = $rdel2 = $ldel1;
834 $str2pos = $ld2pos+1;
836 if ($ldel2 =~ /[[(<{]/)
838 pos($$textref)--; # OVERCOME BROKEN LOOKAHEAD
839 defined(_match_bracketed($textref,"",$ldel2,"","",$rdel2))
840 || do { pos $$textref = $startpos; return };
844 $$textref =~ /[^\\$ldel2]*(\\.[^\\$ldel2]*)*$ldel2/gcs
845 || do { pos $$textref = $startpos; return };
847 $rd2pos = pos($$textref)-1;
851 $ld2pos = $str2pos = $rd2pos = $rd1pos;
854 $modpos = pos $$textref;
856 $$textref =~ m/\G($mods{$op})/gc;
857 my $endpos = pos $$textref;
860 $startpos, $oppos-$startpos, # PREFIX
861 $oppos, length($op), # OPERATOR
862 $ld1pos, 1, # LEFT DEL
863 $str1pos, $rd1pos-$str1pos, # STR/PAT
864 $rd1pos, 1, # RIGHT DEL
865 $ld2pos, $second_arg, # 2ND LDEL (MAYBE)
866 $str2pos, $rd2pos-$str2pos, # 2ND STR (MAYBE)
867 $rd2pos, $second_arg, # 2ND RDEL (MAYBE)
868 $modpos, $endpos-$modpos, # MODIFIERS
869 $endpos, $textlen-$endpos, # REMAINDER
875 sub { extract_variable($_[0], '') },
876 sub { extract_quotelike($_[0],'') },
877 sub { extract_codeblock($_[0],'{}','') },
880 sub extract_multiple (;$$$$) # ($text, $functions_ref, $max_fields, $ignoreunknown)
882 my $textref = defined($_[0]) ? \$_[0] : \$_;
884 my ($lastpos, $firstpos);
889 my @func = defined $_[1] ? @{$_[1]} : @{$def_func};
890 my $max = defined $_[2] && $_[2]>0 ? $_[2] : 1_000_000_000;
897 carp ("extract_multiple reset maximal count to 1 in scalar context")
898 if $^W && defined($_[2]) && $max > 1;
907 foreach $func ( @func )
909 if (ref($func) eq 'HASH')
911 push @class, (keys %$func)[0];
912 $func = (values %$func)[0];
920 FIELD: while (pos($$textref) < length($$textref))
924 foreach my $i ( 0..$#func )
929 $lastpos = pos $$textref;
930 if (ref($func) eq 'CODE')
931 { ($field,$rem,$pref) = @bits = $func->($$textref) }
932 elsif (ref($func) eq 'Text::Balanced::Extractor')
933 { @bits = $field = $func->extract($$textref) }
934 elsif( $$textref =~ m/\G$func/gc )
935 { @bits = $field = defined($1) ? $1 : $& }
937 if (defined($field) && length($field))
941 if length($pref) && !defined($unkpos);
944 push @fields, substr($$textref, $unkpos, $lastpos-$unkpos).$pref;
945 $firstpos = $unkpos unless defined $firstpos;
947 last FIELD if @fields == $max;
951 ? bless (\$field, $class)
953 $firstpos = $lastpos unless defined $firstpos;
954 $lastpos = pos $$textref;
955 last FIELD if @fields == $max;
959 if ($$textref =~ /\G(.)/gcs)
961 $unkpos = pos($$textref)-1
962 unless $igunk || defined $unkpos;
968 push @fields, substr($$textref, $unkpos);
969 $firstpos = $unkpos unless defined $firstpos;
970 $lastpos = length $$textref;
975 pos $$textref = $lastpos;
976 return @fields if wantarray;
979 eval { substr($$textref,$firstpos,$lastpos-$firstpos)="";
980 pos $$textref = $firstpos };
985 sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options)
989 my $pre = defined $_[2] ? $_[2] : '\s*';
990 my %options = defined $_[3] ? %{$_[3]} : ();
991 my $omode = defined $options{fail} ? $options{fail} : '';
992 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
993 : defined($options{reject}) ? $options{reject}
996 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
997 : defined($options{ignore}) ? $options{ignore}
1001 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
1004 for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ }
1009 my $textref = defined $_[0] ? \$_[0] : \$_;
1010 my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
1012 return _fail(wantarray, $textref) unless @match;
1013 return _succeed wantarray, $textref,
1014 $match[2], $match[3]+$match[5]+$match[7], # MATCH
1015 @match[8..9,0..1,2..7]; # REM, PRE, BITS
1018 bless $closure, 'Text::Balanced::Extractor';
1021 package Text::Balanced::Extractor;
1023 sub extract($$) # ($self, $text)
1028 package Text::Balanced::ErrorMsg;
1030 use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" };
1038 Text::Balanced - Extract delimited text sequences from strings.
1043 use Text::Balanced qw (
1056 # Extract the initial substring of $text that is delimited by
1057 # two (unescaped) instances of the first character in $delim.
1059 ($extracted, $remainder) = extract_delimited($text,$delim);
1062 # Extract the initial substring of $text that is bracketed
1063 # with a delimiter(s) specified by $delim (where the string
1064 # in $delim contains one or more of '(){}[]<>').
1066 ($extracted, $remainder) = extract_bracketed($text,$delim);
1069 # Extract the initial substring of $text that is bounded by
1072 ($extracted, $remainder) = extract_tagged($text);
1075 # Extract the initial substring of $text that is bounded by
1076 # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags
1078 ($extracted, $remainder) =
1079 extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]});
1082 # Extract the initial substring of $text that represents a
1083 # Perl "quote or quote-like operation"
1085 ($extracted, $remainder) = extract_quotelike($text);
1088 # Extract the initial substring of $text that represents a block
1089 # of Perl code, bracketed by any of character(s) specified by $delim
1090 # (where the string $delim contains one or more of '(){}[]<>').
1092 ($extracted, $remainder) = extract_codeblock($text,$delim);
1095 # Extract the initial substrings of $text that would be extracted by
1096 # one or more sequential applications of the specified functions
1097 # or regular expressions
1099 @extracted = extract_multiple($text,
1100 [ \&extract_bracketed,
1101 \&extract_quotelike,
1102 \&some_other_extractor_sub,
1107 # Create a string representing an optimized pattern (a la Friedl)
1108 # that matches a substring delimited by any of the specified characters
1109 # (in this case: any type of quote or a slash)
1111 $patstring = gen_delimited_pat(q{'"`/});
1114 # Generate a reference to an anonymous sub that is just like extract_tagged
1115 # but pre-compiled and optimized for a specific pair of tags, and consequently
1116 # much faster (i.e. 3 times faster). It uses qr// for better performance on
1117 # repeated calls, so it only works under Perl 5.005 or later.
1119 $extract_head = gen_extract_tagged('<HEAD>','</HEAD>');
1121 ($extracted, $remainder) = $extract_head->($text);
1126 The various C<extract_...> subroutines may be used to
1127 extract a delimited substring, possibly after skipping a
1128 specified prefix string. By default, that prefix is
1129 optional whitespace (C</\s*/>), but you can change it to whatever
1130 you wish (see below).
1132 The substring to be extracted must appear at the
1133 current C<pos> location of the string's variable
1134 (or at index zero, if no C<pos> position is defined).
1135 In other words, the C<extract_...> subroutines I<don't>
1136 extract the first occurrence of a substring anywhere
1137 in a string (like an unanchored regex would). Rather,
1138 they extract an occurrence of the substring appearing
1139 immediately at the current matching position in the
1140 string (like a C<\G>-anchored regex would).
1144 =head2 General behaviour in list contexts
1146 In a list context, all the subroutines return a list, the first three
1147 elements of which are always:
1153 The extracted string, including the specified delimiters.
1154 If the extraction fails C<undef> is returned.
1158 The remainder of the input string (i.e. the characters after the
1159 extracted string). On failure, the entire string is returned.
1163 The skipped prefix (i.e. the characters before the extracted string).
1164 On failure, C<undef> is returned.
1168 Note that in a list context, the contents of the original input text (the first
1169 argument) are not modified in any way.
1171 However, if the input text was passed in a variable, that variable's
1172 C<pos> value is updated to point at the first character after the
1173 extracted text. That means that in a list context the various
1174 subroutines can be used much like regular expressions. For example:
1176 while ( $next = (extract_quotelike($text))[0] )
1178 # process next quote-like (in $next)
1182 =head2 General behaviour in scalar and void contexts
1184 In a scalar context, the extracted string is returned, having first been
1185 removed from the input text. Thus, the following code also processes
1186 each quote-like operation, but actually removes them from $text:
1188 while ( $next = extract_quotelike($text) )
1190 # process next quote-like (in $next)
1193 Note that if the input text is a read-only string (i.e. a literal),
1194 no attempt is made to remove the extracted text.
1196 In a void context the behaviour of the extraction subroutines is
1197 exactly the same as in a scalar context, except (of course) that the
1198 extracted substring is not returned.
1200 =head2 A note about prefixes
1202 Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.)
1203 This can bite you if you're expecting a prefix specification like
1204 '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix
1205 pattern will only succeed if the <H1> tag is on the current line, since
1206 . normally doesn't match newlines.
1208 To overcome this limitation, you need to turn on /s matching within
1209 the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)'
1212 =head2 C<extract_delimited>
1214 The C<extract_delimited> function formalizes the common idiom
1215 of extracting a single-character-delimited substring from the start of
1216 a string. For example, to extract a single-quote delimited string, the
1217 following code is typically used:
1219 ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s;
1222 but with C<extract_delimited> it can be simplified to:
1224 ($extracted,$remainder) = extract_delimited($text, "'");
1226 C<extract_delimited> takes up to four scalars (the input text, the
1227 delimiters, a prefix pattern to be skipped, and any escape characters)
1228 and extracts the initial substring of the text that
1229 is appropriately delimited. If the delimiter string has multiple
1230 characters, the first one encountered in the text is taken to delimit
1232 The third argument specifies a prefix pattern that is to be skipped
1233 (but must be present!) before the substring is extracted.
1234 The final argument specifies the escape character to be used for each
1237 All arguments are optional. If the escape characters are not specified,
1238 every delimiter is escaped with a backslash (C<\>).
1239 If the prefix is not specified, the
1240 pattern C<'\s*'> - optional whitespace - is used. If the delimiter set
1241 is also not specified, the set C</["'`]/> is used. If the text to be processed
1242 is not specified either, C<$_> is used.
1244 In list context, C<extract_delimited> returns a array of three
1245 elements, the extracted substring (I<including the surrounding
1246 delimiters>), the remainder of the text, and the skipped prefix (if
1247 any). If a suitable delimited substring is not found, the first
1248 element of the array is the empty string, the second is the complete
1249 original text, and the prefix returned in the third element is an
1252 In a scalar context, just the extracted substring is returned. In
1253 a void context, the extracted substring (and any prefix) are simply
1254 removed from the beginning of the first argument.
1258 # Remove a single-quoted substring from the very beginning of $text:
1260 $substring = extract_delimited($text, "'", '');
1262 # Remove a single-quoted Pascalish substring (i.e. one in which
1263 # doubling the quote character escapes it) from the very
1264 # beginning of $text:
1266 $substring = extract_delimited($text, "'", '', "'");
1268 # Extract a single- or double- quoted substring from the
1269 # beginning of $text, optionally after some whitespace
1270 # (note the list context to protect $text from modification):
1272 ($substring) = extract_delimited $text, q{"'};
1275 # Delete the substring delimited by the first '/' in $text:
1277 $text = join '', (extract_delimited($text,'/','[^/]*')[2,1];
1279 Note that this last example is I<not> the same as deleting the first
1280 quote-like pattern. For instance, if C<$text> contained the string:
1282 "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }"
1284 then after the deletion it would contain:
1286 "if ('.$UNIXCMD/s) { $cmd = $1; }"
1290 "if ('./cmd' =~ ms) { $cmd = $1; }"
1293 See L<"extract_quotelike"> for a (partial) solution to this problem.
1296 =head2 C<extract_bracketed>
1298 Like C<"extract_delimited">, the C<extract_bracketed> function takes
1299 up to three optional scalar arguments: a string to extract from, a delimiter
1300 specifier, and a prefix pattern. As before, a missing prefix defaults to
1301 optional whitespace and a missing text defaults to C<$_>. However, a missing
1302 delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below).
1304 C<extract_bracketed> extracts a balanced-bracket-delimited
1305 substring (using any one (or more) of the user-specified delimiter
1306 brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also
1307 respect quoted unbalanced brackets (see below).
1309 A "delimiter bracket" is a bracket in list of delimiters passed as
1310 C<extract_bracketed>'s second argument. Delimiter brackets are
1311 specified by giving either the left or right (or both!) versions
1312 of the required bracket(s). Note that the order in which
1313 two or more delimiter brackets are specified is not significant.
1315 A "balanced-bracket-delimited substring" is a substring bounded by
1316 matched brackets, such that any other (left or right) delimiter
1317 bracket I<within> the substring is also matched by an opposite
1318 (right or left) delimiter bracket I<at the same level of nesting>. Any
1319 type of bracket not in the delimiter list is treated as an ordinary
1322 In other words, each type of bracket specified as a delimiter must be
1323 balanced and correctly nested within the substring, and any other kind of
1324 ("non-delimiter") bracket in the substring is ignored.
1326 For example, given the string:
1328 $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }";
1330 then a call to C<extract_bracketed> in a list context:
1332 @result = extract_bracketed( $text, '{}' );
1336 ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" )
1338 since both sets of C<'{..}'> brackets are properly nested and evenly balanced.
1339 (In a scalar context just the first element of the array would be returned. In
1340 a void context, C<$text> would be replaced by an empty string.)
1342 Likewise the call in:
1344 @result = extract_bracketed( $text, '{[' );
1346 would return the same result, since all sets of both types of specified
1347 delimiter brackets are correctly nested and balanced.
1349 However, the call in:
1351 @result = extract_bracketed( $text, '{([<' );
1353 would fail, returning:
1355 ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" );
1357 because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and
1358 the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would
1359 return an empty string. In a void context, C<$text> would be unchanged.)
1361 Note that the embedded single-quotes in the string don't help in this
1362 case, since they have not been specified as acceptable delimiters and are
1363 therefore treated as non-delimiter characters (and ignored).
1365 However, if a particular species of quote character is included in the
1366 delimiter specification, then that type of quote will be correctly handled.
1367 for example, if C<$text> is:
1369 $text = '<A HREF=">>>>">link</A>';
1373 @result = extract_bracketed( $text, '<">' );
1377 ( '<A HREF=">>>>">', 'link</A>', "" )
1379 as expected. Without the specification of C<"> as an embedded quoter:
1381 @result = extract_bracketed( $text, '<>' );
1383 the result would be:
1385 ( '<A HREF=">', '>>>">link</A>', "" )
1387 In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like
1388 quoting (i.e. q{string}, qq{string}, etc) can be specified by including the
1389 letter 'q' as a delimiter. Hence:
1391 @result = extract_bracketed( $text, '<q>' );
1393 would correctly match something like this:
1395 $text = '<leftop: conj /and/ conj>';
1397 See also: C<"extract_quotelike"> and C<"extract_codeblock">.
1400 =head2 C<extract_variable>
1402 C<extract_variable> extracts any valid Perl variable or
1403 variable-involved expression, including scalars, arrays, hashes, array
1404 accesses, hash look-ups, method calls through objects, subroutine calls
1405 through subroutine references, etc.
1407 The subroutine takes up to two optional arguments:
1413 A string to be processed (C<$_> if the string is omitted or C<undef>)
1417 A string specifying a pattern to be matched as a prefix (which is to be
1418 skipped). If omitted, optional whitespace is skipped.
1422 On success in a list context, an array of 3 elements is returned. The
1429 the extracted variable, or variablish expression
1433 the remainder of the input text,
1437 the prefix substring (if any),
1441 On failure, all of these values (except the remaining text) are C<undef>.
1443 In a scalar context, C<extract_variable> returns just the complete
1444 substring that matched a variablish expression. C<undef> is returned on
1445 failure. In addition, the original input text has the returned substring
1446 (and any prefix) removed from it.
1448 In a void context, the input text just has the matched substring (and
1449 any specified prefix) removed.
1452 =head2 C<extract_tagged>
1454 C<extract_tagged> extracts and segments text between (balanced)
1457 The subroutine takes up to five optional arguments:
1463 A string to be processed (C<$_> if the string is omitted or C<undef>)
1467 A string specifying a pattern to be matched as the opening tag.
1468 If the pattern string is omitted (or C<undef>) then a pattern
1469 that matches any standard XML tag is used.
1473 A string specifying a pattern to be matched at the closing tag.
1474 If the pattern string is omitted (or C<undef>) then the closing
1475 tag is constructed by inserting a C</> after any leading bracket
1476 characters in the actual opening tag that was matched (I<not> the pattern
1477 that matched the tag). For example, if the opening tag pattern
1478 is specified as C<'{{\w+}}'> and actually matched the opening tag
1479 C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">.
1483 A string specifying a pattern to be matched as a prefix (which is to be
1484 skipped). If omitted, optional whitespace is skipped.
1488 A hash reference containing various parsing options (see below)
1492 The various options that can be specified are:
1496 =item C<reject =E<gt> $listref>
1498 The list reference contains one or more strings specifying patterns
1499 that must I<not> appear within the tagged text.
1501 For example, to extract
1502 an HTML link (which should not contain nested links) use:
1504 extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} );
1506 =item C<ignore =E<gt> $listref>
1508 The list reference contains one or more strings specifying patterns
1509 that are I<not> be be treated as nested tags within the tagged text
1510 (even if they would match the start tag pattern).
1512 For example, to extract an arbitrary XML tag, but ignore "empty" elements:
1514 extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} );
1516 (also see L<"gen_delimited_pat"> below).
1519 =item C<fail =E<gt> $str>
1521 The C<fail> option indicates the action to be taken if a matching end
1522 tag is not encountered (i.e. before the end of the string or some
1523 C<reject> pattern matches). By default, a failure to match a closing
1524 tag causes C<extract_tagged> to immediately fail.
1526 However, if the string value associated with <reject> is "MAX", then
1527 C<extract_tagged> returns the complete text up to the point of failure.
1528 If the string is "PARA", C<extract_tagged> returns only the first paragraph
1529 after the tag (up to the first line that is either empty or contains
1530 only whitespace characters).
1531 If the string is "", the the default behaviour (i.e. failure) is reinstated.
1533 For example, suppose the start tag "/para" introduces a paragraph, which then
1534 continues until the next "/endpara" tag or until another "/para" tag is
1537 $text = "/para line 1\n\nline 3\n/para line 4";
1539 extract_tagged($text, '/para', '/endpara', undef,
1540 {reject => '/para', fail => MAX );
1542 # EXTRACTED: "/para line 1\n\nline 3\n"
1544 Suppose instead, that if no matching "/endpara" tag is found, the "/para"
1545 tag refers only to the immediately following paragraph:
1547 $text = "/para line 1\n\nline 3\n/para line 4";
1549 extract_tagged($text, '/para', '/endpara', undef,
1550 {reject => '/para', fail => MAX );
1552 # EXTRACTED: "/para line 1\n"
1554 Note that the specified C<fail> behaviour applies to nested tags as well.
1558 On success in a list context, an array of 6 elements is returned. The elements are:
1564 the extracted tagged substring (including the outermost tags),
1568 the remainder of the input text,
1572 the prefix substring (if any),
1580 the text between the opening and closing tags
1584 the closing tag (or "" if no closing tag was found)
1588 On failure, all of these values (except the remaining text) are C<undef>.
1590 In a scalar context, C<extract_tagged> returns just the complete
1591 substring that matched a tagged text (including the start and end
1592 tags). C<undef> is returned on failure. In addition, the original input
1593 text has the returned substring (and any prefix) removed from it.
1595 In a void context, the input text just has the matched substring (and
1596 any specified prefix) removed.
1599 =head2 C<gen_extract_tagged>
1601 (Note: This subroutine is only available under Perl5.005)
1603 C<gen_extract_tagged> generates a new anonymous subroutine which
1604 extracts text between (balanced) specified tags. In other words,
1605 it generates a function identical in function to C<extract_tagged>.
1607 The difference between C<extract_tagged> and the anonymous
1608 subroutines generated by
1609 C<gen_extract_tagged>, is that those generated subroutines:
1615 do not have to reparse tag specification or parsing options every time
1616 they are called (whereas C<extract_tagged> has to effectively rebuild
1617 its tag parser on every call);
1621 make use of the new qr// construct to pre-compile the regexes they use
1622 (whereas C<extract_tagged> uses standard string variable interpolation
1623 to create tag-matching patterns).
1627 The subroutine takes up to four optional arguments (the same set as
1628 C<extract_tagged> except for the string to be processed). It returns
1629 a reference to a subroutine which in turn takes a single argument (the text to
1632 In other words, the implementation of C<extract_tagged> is exactly
1638 $extractor = gen_extract_tagged(@_);
1639 return $extractor->($text);
1642 (although C<extract_tagged> is not currently implemented that way, in order
1643 to preserve pre-5.005 compatibility).
1645 Using C<gen_extract_tagged> to create extraction functions for specific tags
1646 is a good idea if those functions are going to be called more than once, since
1647 their performance is typically twice as good as the more general-purpose
1651 =head2 C<extract_quotelike>
1653 C<extract_quotelike> attempts to recognize, extract, and segment any
1654 one of the various Perl quotes and quotelike operators (see
1655 L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket
1656 delimiters (for the quotelike operators), and trailing modifiers are
1657 all caught. For example, in:
1659 extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #'
1661 extract_quotelike ' "You said, \"Use sed\"." '
1663 extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; '
1665 extract_quotelike ' tr/\\\/\\\\/\\\//ds; '
1667 the full Perl quotelike operations are all extracted correctly.
1669 Note too that, when using the /x modifier on a regex, any comment
1670 containing the current pattern delimiter will cause the regex to be
1671 immediately terminated. In other words:
1674 (?i) # CASE INSENSITIVE
1675 [a-z_] # LEADING ALPHABETIC/UNDERSCORE
1676 [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS
1679 will be extracted as if it were:
1682 (?i) # CASE INSENSITIVE
1683 [a-z_] # LEADING ALPHABETIC/'
1685 This behaviour is identical to that of the actual compiler.
1687 C<extract_quotelike> takes two arguments: the text to be processed and
1688 a prefix to be matched at the very beginning of the text. If no prefix
1689 is specified, optional whitespace is the default. If no text is given,
1692 In a list context, an array of 11 elements is returned. The elements are:
1698 the extracted quotelike substring (including trailing modifiers),
1702 the remainder of the input text,
1706 the prefix substring (if any),
1710 the name of the quotelike operator (if any),
1714 the left delimiter of the first block of the operation,
1718 the text of the first block of the operation
1719 (that is, the contents of
1720 a quote, the regex of a match or substitution or the target list of a
1725 the right delimiter of the first block of the operation,
1729 the left delimiter of the second block of the operation
1730 (that is, if it is a C<s>, C<tr>, or C<y>),
1734 the text of the second block of the operation
1735 (that is, the replacement of a substitution or the translation list
1740 the right delimiter of the second block of the operation (if any),
1744 the trailing modifiers on the operation (if any).
1748 For each of the fields marked "(if any)" the default value on success is
1750 On failure, all of these values (except the remaining text) are C<undef>.
1753 In a scalar context, C<extract_quotelike> returns just the complete substring
1754 that matched a quotelike operation (or C<undef> on failure). In a scalar or
1755 void context, the input text has the same substring (and any specified
1760 # Remove the first quotelike literal that appears in text
1762 $quotelike = extract_quotelike($text,'.*?');
1764 # Replace one or more leading whitespace-separated quotelike
1765 # literals in $_ with "<QLL>"
1767 do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@;
1770 # Isolate the search pattern in a quotelike operation from $text
1772 ($op,$pat) = (extract_quotelike $text)[3,5];
1775 print "search pattern: $pat\n";
1779 print "$op is not a pattern matching operation\n";
1783 =head2 C<extract_quotelike> and "here documents"
1785 C<extract_quotelike> can successfully extract "here documents" from an input
1786 string, but with an important caveat in list contexts.
1788 Unlike other types of quote-like literals, a here document is rarely
1789 a contiguous substring. For example, a typical piece of code using
1790 here document might look like this:
1793 This is the message.
1797 Given this as an input string in a scalar context, C<extract_quotelike>
1798 would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG",
1799 leaving the string " || die;\nexit;" in the original variable. In other words,
1800 the two separate pieces of the here document are successfully extracted and
1803 In a list context, C<extract_quotelike> would return the list
1809 "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document,
1810 including fore and aft delimiters),
1814 " || die;\nexit;" (i.e. the remainder of the input text, concatenated),
1818 "" (i.e. the prefix substring -- trivial in this case),
1822 "<<" (i.e. the "name" of the quotelike operator)
1826 "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes),
1830 "This is the message.\n" (i.e. the text of the here document),
1834 "EOMSG" (i.e. the right delimiter of the here document),
1838 "" (a here document has no second left delimiter, second text, second right
1839 delimiter, or trailing modifiers).
1843 However, the matching position of the input variable would be set to
1844 "exit;" (i.e. I<after> the closing delimiter of the here document),
1845 which would cause the earlier " || die;\nexit;" to be skipped in any
1846 sequence of code fragment extractions.
1848 To avoid this problem, when it encounters a here document whilst
1849 extracting from a modifiable string, C<extract_quotelike> silently
1850 rearranges the string to an equivalent piece of Perl:
1853 This is the message.
1858 in which the here document I<is> contiguous. It still leaves the
1859 matching position after the here document, but now the rest of the line
1860 on which the here document starts is not skipped.
1862 To prevent <extract_quotelike> from mucking about with the input in this way
1863 (this is the only case where a list-context C<extract_quotelike> does so),
1864 you can pass the input variable as an interpolated literal:
1866 $quotelike = extract_quotelike("$var");
1869 =head2 C<extract_codeblock>
1871 C<extract_codeblock> attempts to recognize and extract a balanced
1872 bracket delimited substring that may contain unbalanced brackets
1873 inside Perl quotes or quotelike operations. That is, C<extract_codeblock>
1874 is like a combination of C<"extract_bracketed"> and
1875 C<"extract_quotelike">.
1877 C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>:
1878 a text to process, a set of delimiter brackets to look for, and a prefix to
1879 match first. It also takes an optional fourth parameter, which allows the
1880 outermost delimiter brackets to be specified separately (see below).
1882 Omitting the first argument (input text) means process C<$_> instead.
1883 Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used.
1884 Omitting the third argument (prefix argument) implies optional whitespace at the start.
1885 Omitting the fourth argument (outermost delimiter brackets) indicates that the
1886 value of the second argument is to be used for the outermost delimiters.
1888 Once the prefix an dthe outermost opening delimiter bracket have been
1889 recognized, code blocks are extracted by stepping through the input text and
1890 trying the following alternatives in sequence:
1896 Try and match a closing delimiter bracket. If the bracket was the same
1897 species as the last opening bracket, return the substring to that
1898 point. If the bracket was mismatched, return an error.
1902 Try to match a quote or quotelike operator. If found, call
1903 C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return
1904 the error it returned. Otherwise go back to step 1.
1908 Try to match an opening delimiter bracket. If found, call
1909 C<extract_codeblock> recursively to eat the embedded block. If the
1910 recursive call fails, return an error. Otherwise, go back to step 1.
1914 Unconditionally match a bareword or any other single character, and
1915 then go back to step 1.
1922 # Find a while loop in the text
1924 if ($text =~ s/.*?while\s*\{/{/)
1926 $loop = "while " . extract_codeblock($text);
1929 # Remove the first round-bracketed list (which may include
1930 # round- or curly-bracketed code blocks or quotelike operators)
1932 extract_codeblock $text, "(){}", '[^(]*';
1935 The ability to specify a different outermost delimiter bracket is useful
1936 in some circumstances. For example, in the Parse::RecDescent module,
1937 parser actions which are to be performed only on a successful parse
1938 are specified using a C<E<lt>defer:...E<gt>> directive. For example:
1940 sentence: subject verb object
1941 <defer: {$::theVerb = $item{verb}} >
1943 Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code
1944 within the C<E<lt>defer:...E<gt>> directive, but there's a problem.
1946 A deferred action like this:
1948 <defer: {if ($count>10) {$count--}} >
1950 will be incorrectly parsed as:
1952 <defer: {if ($count>
1954 because the "less than" operator is interpreted as a closing delimiter.
1956 But, by extracting the directive using
1957 S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>>
1958 the '>' character is only treated as a delimited at the outermost
1959 level of the code block, so the directive is parsed correctly.
1961 =head2 C<extract_multiple>
1963 The C<extract_multiple> subroutine takes a string to be processed and a
1964 list of extractors (subroutines or regular expressions) to apply to that string.
1966 In an array context C<extract_multiple> returns an array of substrings
1967 of the original string, as extracted by the specified extractors.
1968 In a scalar context, C<extract_multiple> returns the first
1969 substring successfully extracted from the original string. In both
1970 scalar and void contexts the original string has the first successfully
1971 extracted substring removed from it. In all contexts
1972 C<extract_multiple> starts at the current C<pos> of the string, and
1973 sets that C<pos> appropriately after it matches.
1975 Hence, the aim of of a call to C<extract_multiple> in a list context
1976 is to split the processed string into as many non-overlapping fields as
1977 possible, by repeatedly applying each of the specified extractors
1978 to the remainder of the string. Thus C<extract_multiple> is
1979 a generalized form of Perl's C<split> subroutine.
1981 The subroutine takes up to four optional arguments:
1987 A string to be processed (C<$_> if the string is omitted or C<undef>)
1991 A reference to a list of subroutine references and/or qr// objects and/or
1992 literal strings and/or hash references, specifying the extractors
1993 to be used to split the string. If this argument is omitted (or
1997 sub { extract_variable($_[0], '') },
1998 sub { extract_quotelike($_[0],'') },
1999 sub { extract_codeblock($_[0],'{}','') },
2007 An number specifying the maximum number of fields to return. If this
2008 argument is omitted (or C<undef>), split continues as long as possible.
2010 If the third argument is I<N>, then extraction continues until I<N> fields
2011 have been successfully extracted, or until the string has been completely
2014 Note that in scalar and void contexts the value of this argument is
2015 automatically reset to 1 (under C<-w>, a warning is issued if the argument
2020 A value indicating whether unmatched substrings (see below) within the
2021 text should be skipped or returned as fields. If the value is true,
2022 such substrings are skipped. Otherwise, they are returned.
2026 The extraction process works by applying each extractor in
2027 sequence to the text string.
2029 If the extractor is a subroutine it is called in a list context and is
2030 expected to return a list of a single element, namely the extracted
2031 text. It may optionally also return two further arguments: a string
2032 representing the text left after extraction (like $' for a pattern
2033 match), and a string representing any prefix skipped before the
2034 extraction (like $` in a pattern match). Note that this is designed
2035 to facilitate the use of other Text::Balanced subroutines with
2036 C<extract_multiple>. Note too that the value returned by an extractor
2037 subroutine need not bear any relationship to the corresponding substring
2038 of the original text (see examples below).
2040 If the extractor is a precompiled regular expression or a string,
2041 it is matched against the text in a scalar context with a leading
2042 '\G' and the gc modifiers enabled. The extracted value is either
2043 $1 if that variable is defined after the match, or else the
2044 complete match (i.e. $&).
2046 If the extractor is a hash reference, it must contain exactly one element.
2047 The value of that element is one of the
2048 above extractor types (subroutine reference, regular expression, or string).
2049 The key of that element is the name of a class into which the successful
2050 return value of the extractor will be blessed.
2052 If an extractor returns a defined value, that value is immediately
2053 treated as the next extracted field and pushed onto the list of fields.
2054 If the extractor was specified in a hash reference, the field is also
2055 blessed into the appropriate class,
2057 If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is
2058 assumed to have failed to extract.
2059 If none of the extractor subroutines succeeds, then one
2060 character is extracted from the start of the text and the extraction
2061 subroutines reapplied. Characters which are thus removed are accumulated and
2062 eventually become the next field (unless the fourth argument is true, in which
2063 case they are discarded).
2065 For example, the following extracts substrings that are valid Perl variables:
2067 @fields = extract_multiple($text,
2068 [ sub { extract_variable($_[0]) } ],
2071 This example separates a text into fields which are quote delimited,
2072 curly bracketed, and anything else. The delimited and bracketed
2073 parts are also blessed to identify them (the "anything else" is unblessed):
2075 @fields = extract_multiple($text,
2077 { Delim => sub { extract_delimited($_[0],q{'"}) } },
2078 { Brack => sub { extract_bracketed($_[0],'{}') } },
2081 This call extracts the next single substring that is a valid Perl quotelike
2082 operator (and removes it from $text):
2084 $quotelike = extract_multiple($text,
2086 sub { extract_quotelike($_[0]) },
2089 Finally, here is yet another way to do comma-separated value parsing:
2091 @fields = extract_multiple($csv_text,
2093 sub { extract_delimited($_[0],q{'"}) },
2098 The list in the second argument means:
2099 I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">.
2100 The undef third argument means:
2101 I<"...as many times as possible...">,
2102 and the true value in the fourth argument means
2103 I<"...discarding anything else that appears (i.e. the commas)">.
2105 If you wanted the commas preserved as separate fields (i.e. like split
2106 does if your split pattern has capturing parentheses), you would
2107 just make the last parameter undefined (or remove it).
2110 =head2 C<gen_delimited_pat>
2112 The C<gen_delimited_pat> subroutine takes a single (string) argument and
2113 > builds a Friedl-style optimized regex that matches a string delimited
2114 by any one of the characters in the single argument. For example:
2116 gen_delimited_pat(q{'"})
2120 (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\')
2122 Note that the specified delimiters are automatically quotemeta'd.
2124 A typical use of C<gen_delimited_pat> would be to build special purpose tags
2125 for C<extract_tagged>. For example, to properly ignore "empty" XML elements
2126 (which might contain quoted strings):
2128 my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>';
2130 extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} );
2133 C<gen_delimited_pat> may also be called with an optional second argument,
2134 which specifies the "escape" character(s) to be used for each delimiter.
2135 For example to match a Pascal-style string (where ' is the delimiter
2136 and '' is a literal ' within the string):
2138 gen_delimited_pat(q{'},q{'});
2140 Different escape characters can be specified for different delimiters.
2141 For example, to specify that '/' is the escape for single quotes
2142 and '%' is the escape for double quotes:
2144 gen_delimited_pat(q{'"},q{/%});
2146 If more delimiters than escape chars are specified, the last escape char
2147 is used for the remaining delimiters.
2148 If no escape char is specified for a given specified delimiter, '\' is used.
2150 =head2 C<delimited_pat>
2152 Note that C<gen_delimited_pat> was previously called C<delimited_pat>.
2153 That name may still be used, but is now deprecated.
2158 In a list context, all the functions return C<(undef,$original_text)>
2159 on failure. In a scalar context, failure is indicated by returning C<undef>
2160 (in this case the input text is not modified in any way).
2162 In addition, on failure in I<any> context, the C<$@> variable is set.
2163 Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed
2165 Accessing C<$@-E<gt>{pos}> returns the offset into the original string at
2166 which the error was detected (although not necessarily where it occurred!)
2167 Printing C<$@> directly produces the error message, with the offset appended.
2168 On success, the C<$@> variable is guaranteed to be C<undef>.
2170 The available diagnostics are:
2174 =item C<Did not find a suitable bracket: "%s">
2176 The delimiter provided to C<extract_bracketed> was not one of
2177 C<'()[]E<lt>E<gt>{}'>.
2179 =item C<Did not find prefix: /%s/>
2181 A non-optional prefix was specified but wasn't found at the start of the text.
2183 =item C<Did not find opening bracket after prefix: "%s">
2185 C<extract_bracketed> or C<extract_codeblock> was expecting a
2186 particular kind of bracket at the start of the text, and didn't find it.
2188 =item C<No quotelike operator found after prefix: "%s">
2190 C<extract_quotelike> didn't find one of the quotelike operators C<q>,
2191 C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring
2194 =item C<Unmatched closing bracket: "%c">
2196 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered
2197 a closing bracket where none was expected.
2199 =item C<Unmatched opening bracket(s): "%s">
2201 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran
2202 out of characters in the text before closing one or more levels of nested
2205 =item C<Unmatched embedded quote (%s)>
2207 C<extract_bracketed> attempted to match an embedded quoted substring, but
2208 failed to find a closing quote to match it.
2210 =item C<Did not find closing delimiter to match '%s'>
2212 C<extract_quotelike> was unable to find a closing delimiter to match the
2213 one that opened the quote-like operation.
2215 =item C<Mismatched closing bracket: expected "%c" but found "%s">
2217 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found
2218 a valid bracket delimiter, but it was the wrong species. This usually
2219 indicates a nesting error, but may indicate incorrect quoting or escaping.
2221 =item C<No block delimiter found after quotelike "%s">
2223 C<extract_quotelike> or C<extract_codeblock> found one of the
2224 quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y>
2225 without a suitable block after it.
2227 =item C<Did not find leading dereferencer>
2229 C<extract_variable> was expecting one of '$', '@', or '%' at the start of
2230 a variable, but didn't find any of them.
2232 =item C<Bad identifier after dereferencer>
2234 C<extract_variable> found a '$', '@', or '%' indicating a variable, but that
2235 character was not followed by a legal Perl identifier.
2237 =item C<Did not find expected opening bracket at %s>
2239 C<extract_codeblock> failed to find any of the outermost opening brackets
2240 that were specified.
2242 =item C<Improperly nested codeblock at %s>
2244 A nested code block was found that started with a delimiter that was specified
2245 as being only to be used as an outermost bracket.
2247 =item C<Missing second block for quotelike "%s">
2249 C<extract_codeblock> or C<extract_quotelike> found one of the
2250 quotelike operators C<s>, C<tr> or C<y> followed by only one block.
2252 =item C<No match found for opening bracket>
2254 C<extract_codeblock> failed to find a closing bracket to match the outermost
2257 =item C<Did not find opening tag: /%s/>
2259 C<extract_tagged> did not find a suitable opening tag (after any specified
2260 prefix was removed).
2262 =item C<Unable to construct closing tag to match: /%s/>
2264 C<extract_tagged> matched the specified opening tag and tried to
2265 modify the matched text to produce a matching closing tag (because
2266 none was specified). It failed to generate the closing tag, almost
2267 certainly because the opening tag did not start with a
2268 bracket of some kind.
2270 =item C<Found invalid nested tag: %s>
2272 C<extract_tagged> found a nested tag that appeared in the "reject" list
2273 (and the failure mode was not "MAX" or "PARA").
2275 =item C<Found unbalanced nested tag: %s>
2277 C<extract_tagged> found a nested opening tag that was not matched by a
2278 corresponding nested closing tag (and the failure mode was not "MAX" or "PARA").
2280 =item C<Did not find closing tag>
2282 C<extract_tagged> reached the end of the text without finding a closing tag
2283 to match the original opening tag (and the failure mode was not
2294 Damian Conway (damian@conway.org)
2297 =head1 BUGS AND IRRITATIONS
2299 There are undoubtedly serious bugs lurking somewhere in this code, if
2300 only because parts of it give the impression of understanding a great deal
2301 more about Perl than they really do.
2303 Bug reports and other feedback are most welcome.
2308 Copyright (c) 1997-2001, Damian Conway. All Rights Reserved.
2309 This module is free software. It may be used, redistributed
2310 and/or modified under the same terms as Perl itself.