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($$$$$$$);
37 sub _match_quotelike($$$$);
40 require Carp; goto &Carp::carp;
44 require Carp; goto &Carp::croak;
47 # HANDLE RETURN VALUES IN VARIOUS CONTEXTS
50 my ($message, $pos) = @_;
51 $@ = bless { error=>$message, pos=>$pos }, "Text::Balanced::ErrorMsg";
56 my ($wantarray, $textref, $message, $pos) = @_;
57 _failmsg $message, $pos if $message;
58 return ("",$$textref,"") 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->($$textref);
932 # print "[$field|$rem]" if $field;
934 elsif (ref($func) eq 'Text::Balanced::Extractor')
935 { @bits = $field = $func->extract($$textref) }
936 elsif( $$textref =~ m/\G$func/gc )
937 { @bits = $field = defined($1) ? $1 : $& }
939 if (defined($field) && length($field))
943 if length($pref) && !defined($unkpos);
946 push @fields, substr($$textref, $unkpos, $lastpos-$unkpos).$pref;
947 $firstpos = $unkpos unless defined $firstpos;
949 last FIELD if @fields == $max;
953 ? bless (\$field, $class)
955 $firstpos = $lastpos unless defined $firstpos;
956 $lastpos = pos $$textref;
957 last FIELD if @fields == $max;
961 if ($$textref =~ /\G(.)/gcs)
963 $unkpos = pos($$textref)-1
964 unless $igunk || defined $unkpos;
970 push @fields, substr($$textref, $unkpos);
971 $firstpos = $unkpos unless defined $firstpos;
972 $lastpos = length $$textref;
977 pos $$textref = $lastpos;
978 return @fields if wantarray;
981 eval { substr($$textref,$firstpos,$lastpos-$firstpos)="";
982 pos $$textref = $firstpos };
987 sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options)
991 my $pre = defined $_[2] ? $_[2] : '\s*';
992 my %options = defined $_[3] ? %{$_[3]} : ();
993 my $omode = defined $options{fail} ? $options{fail} : '';
994 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
995 : defined($options{reject}) ? $options{reject}
998 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
999 : defined($options{ignore}) ? $options{ignore}
1003 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
1006 for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ }
1011 my $textref = defined $_[0] ? \$_[0] : \$_;
1012 my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
1014 return _fail(wantarray, $textref) unless @match;
1015 return _succeed wantarray, $textref,
1016 $match[2], $match[3]+$match[5]+$match[7], # MATCH
1017 @match[8..9,0..1,2..7]; # REM, PRE, BITS
1020 bless $closure, 'Text::Balanced::Extractor';
1023 package Text::Balanced::Extractor;
1025 sub extract($$) # ($self, $text)
1030 package Text::Balanced::ErrorMsg;
1032 use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" };
1040 Text::Balanced - Extract delimited text sequences from strings.
1045 use Text::Balanced qw (
1058 # Extract the initial substring of $text that is delimited by
1059 # two (unescaped) instances of the first character in $delim.
1061 ($extracted, $remainder) = extract_delimited($text,$delim);
1064 # Extract the initial substring of $text that is bracketed
1065 # with a delimiter(s) specified by $delim (where the string
1066 # in $delim contains one or more of '(){}[]<>').
1068 ($extracted, $remainder) = extract_bracketed($text,$delim);
1071 # Extract the initial substring of $text that is bounded by
1074 ($extracted, $remainder) = extract_tagged($text);
1077 # Extract the initial substring of $text that is bounded by
1078 # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags
1080 ($extracted, $remainder) =
1081 extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]});
1084 # Extract the initial substring of $text that represents a
1085 # Perl "quote or quote-like operation"
1087 ($extracted, $remainder) = extract_quotelike($text);
1090 # Extract the initial substring of $text that represents a block
1091 # of Perl code, bracketed by any of character(s) specified by $delim
1092 # (where the string $delim contains one or more of '(){}[]<>').
1094 ($extracted, $remainder) = extract_codeblock($text,$delim);
1097 # Extract the initial substrings of $text that would be extracted by
1098 # one or more sequential applications of the specified functions
1099 # or regular expressions
1101 @extracted = extract_multiple($text,
1102 [ \&extract_bracketed,
1103 \&extract_quotelike,
1104 \&some_other_extractor_sub,
1109 # Create a string representing an optimized pattern (a la Friedl)
1110 # that matches a substring delimited by any of the specified characters
1111 # (in this case: any type of quote or a slash)
1113 $patstring = gen_delimited_pat(q{'"`/});
1116 # Generate a reference to an anonymous sub that is just like extract_tagged
1117 # but pre-compiled and optimized for a specific pair of tags, and consequently
1118 # much faster (i.e. 3 times faster). It uses qr// for better performance on
1119 # repeated calls, so it only works under Perl 5.005 or later.
1121 $extract_head = gen_extract_tagged('<HEAD>','</HEAD>');
1123 ($extracted, $remainder) = $extract_head->($text);
1128 The various C<extract_...> subroutines may be used to
1129 extract a delimited substring, possibly after skipping a
1130 specified prefix string. By default, that prefix is
1131 optional whitespace (C</\s*/>), but you can change it to whatever
1132 you wish (see below).
1134 The substring to be extracted must appear at the
1135 current C<pos> location of the string's variable
1136 (or at index zero, if no C<pos> position is defined).
1137 In other words, the C<extract_...> subroutines I<don't>
1138 extract the first occurrence of a substring anywhere
1139 in a string (like an unanchored regex would). Rather,
1140 they extract an occurrence of the substring appearing
1141 immediately at the current matching position in the
1142 string (like a C<\G>-anchored regex would).
1146 =head2 General behaviour in list contexts
1148 In a list context, all the subroutines return a list, the first three
1149 elements of which are always:
1155 The extracted string, including the specified delimiters.
1156 If the extraction fails an empty string is returned.
1160 The remainder of the input string (i.e. the characters after the
1161 extracted string). On failure, the entire string is returned.
1165 The skipped prefix (i.e. the characters before the extracted string).
1166 On failure, the empty string is returned.
1170 Note that in a list context, the contents of the original input text (the first
1171 argument) are not modified in any way.
1173 However, if the input text was passed in a variable, that variable's
1174 C<pos> value is updated to point at the first character after the
1175 extracted text. That means that in a list context the various
1176 subroutines can be used much like regular expressions. For example:
1178 while ( $next = (extract_quotelike($text))[0] )
1180 # process next quote-like (in $next)
1184 =head2 General behaviour in scalar and void contexts
1186 In a scalar context, the extracted string is returned, having first been
1187 removed from the input text. Thus, the following code also processes
1188 each quote-like operation, but actually removes them from $text:
1190 while ( $next = extract_quotelike($text) )
1192 # process next quote-like (in $next)
1195 Note that if the input text is a read-only string (i.e. a literal),
1196 no attempt is made to remove the extracted text.
1198 In a void context the behaviour of the extraction subroutines is
1199 exactly the same as in a scalar context, except (of course) that the
1200 extracted substring is not returned.
1202 =head2 A note about prefixes
1204 Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.)
1205 This can bite you if you're expecting a prefix specification like
1206 '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix
1207 pattern will only succeed if the <H1> tag is on the current line, since
1208 . normally doesn't match newlines.
1210 To overcome this limitation, you need to turn on /s matching within
1211 the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)'
1214 =head2 C<extract_delimited>
1216 The C<extract_delimited> function formalizes the common idiom
1217 of extracting a single-character-delimited substring from the start of
1218 a string. For example, to extract a single-quote delimited string, the
1219 following code is typically used:
1221 ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s;
1224 but with C<extract_delimited> it can be simplified to:
1226 ($extracted,$remainder) = extract_delimited($text, "'");
1228 C<extract_delimited> takes up to four scalars (the input text, the
1229 delimiters, a prefix pattern to be skipped, and any escape characters)
1230 and extracts the initial substring of the text that
1231 is appropriately delimited. If the delimiter string has multiple
1232 characters, the first one encountered in the text is taken to delimit
1234 The third argument specifies a prefix pattern that is to be skipped
1235 (but must be present!) before the substring is extracted.
1236 The final argument specifies the escape character to be used for each
1239 All arguments are optional. If the escape characters are not specified,
1240 every delimiter is escaped with a backslash (C<\>).
1241 If the prefix is not specified, the
1242 pattern C<'\s*'> - optional whitespace - is used. If the delimiter set
1243 is also not specified, the set C</["'`]/> is used. If the text to be processed
1244 is not specified either, C<$_> is used.
1246 In list context, C<extract_delimited> returns a array of three
1247 elements, the extracted substring (I<including the surrounding
1248 delimiters>), the remainder of the text, and the skipped prefix (if
1249 any). If a suitable delimited substring is not found, the first
1250 element of the array is the empty string, the second is the complete
1251 original text, and the prefix returned in the third element is an
1254 In a scalar context, just the extracted substring is returned. In
1255 a void context, the extracted substring (and any prefix) are simply
1256 removed from the beginning of the first argument.
1260 # Remove a single-quoted substring from the very beginning of $text:
1262 $substring = extract_delimited($text, "'", '');
1264 # Remove a single-quoted Pascalish substring (i.e. one in which
1265 # doubling the quote character escapes it) from the very
1266 # beginning of $text:
1268 $substring = extract_delimited($text, "'", '', "'");
1270 # Extract a single- or double- quoted substring from the
1271 # beginning of $text, optionally after some whitespace
1272 # (note the list context to protect $text from modification):
1274 ($substring) = extract_delimited $text, q{"'};
1277 # Delete the substring delimited by the first '/' in $text:
1279 $text = join '', (extract_delimited($text,'/','[^/]*')[2,1];
1281 Note that this last example is I<not> the same as deleting the first
1282 quote-like pattern. For instance, if C<$text> contained the string:
1284 "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }"
1286 then after the deletion it would contain:
1288 "if ('.$UNIXCMD/s) { $cmd = $1; }"
1292 "if ('./cmd' =~ ms) { $cmd = $1; }"
1295 See L<"extract_quotelike"> for a (partial) solution to this problem.
1298 =head2 C<extract_bracketed>
1300 Like C<"extract_delimited">, the C<extract_bracketed> function takes
1301 up to three optional scalar arguments: a string to extract from, a delimiter
1302 specifier, and a prefix pattern. As before, a missing prefix defaults to
1303 optional whitespace and a missing text defaults to C<$_>. However, a missing
1304 delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below).
1306 C<extract_bracketed> extracts a balanced-bracket-delimited
1307 substring (using any one (or more) of the user-specified delimiter
1308 brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also
1309 respect quoted unbalanced brackets (see below).
1311 A "delimiter bracket" is a bracket in list of delimiters passed as
1312 C<extract_bracketed>'s second argument. Delimiter brackets are
1313 specified by giving either the left or right (or both!) versions
1314 of the required bracket(s). Note that the order in which
1315 two or more delimiter brackets are specified is not significant.
1317 A "balanced-bracket-delimited substring" is a substring bounded by
1318 matched brackets, such that any other (left or right) delimiter
1319 bracket I<within> the substring is also matched by an opposite
1320 (right or left) delimiter bracket I<at the same level of nesting>. Any
1321 type of bracket not in the delimiter list is treated as an ordinary
1324 In other words, each type of bracket specified as a delimiter must be
1325 balanced and correctly nested within the substring, and any other kind of
1326 ("non-delimiter") bracket in the substring is ignored.
1328 For example, given the string:
1330 $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }";
1332 then a call to C<extract_bracketed> in a list context:
1334 @result = extract_bracketed( $text, '{}' );
1338 ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" )
1340 since both sets of C<'{..}'> brackets are properly nested and evenly balanced.
1341 (In a scalar context just the first element of the array would be returned. In
1342 a void context, C<$text> would be replaced by an empty string.)
1344 Likewise the call in:
1346 @result = extract_bracketed( $text, '{[' );
1348 would return the same result, since all sets of both types of specified
1349 delimiter brackets are correctly nested and balanced.
1351 However, the call in:
1353 @result = extract_bracketed( $text, '{([<' );
1355 would fail, returning:
1357 ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" );
1359 because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and
1360 the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would
1361 return an empty string. In a void context, C<$text> would be unchanged.)
1363 Note that the embedded single-quotes in the string don't help in this
1364 case, since they have not been specified as acceptable delimiters and are
1365 therefore treated as non-delimiter characters (and ignored).
1367 However, if a particular species of quote character is included in the
1368 delimiter specification, then that type of quote will be correctly handled.
1369 for example, if C<$text> is:
1371 $text = '<A HREF=">>>>">link</A>';
1375 @result = extract_bracketed( $text, '<">' );
1379 ( '<A HREF=">>>>">', 'link</A>', "" )
1381 as expected. Without the specification of C<"> as an embedded quoter:
1383 @result = extract_bracketed( $text, '<>' );
1385 the result would be:
1387 ( '<A HREF=">', '>>>">link</A>', "" )
1389 In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like
1390 quoting (i.e. q{string}, qq{string}, etc) can be specified by including the
1391 letter 'q' as a delimiter. Hence:
1393 @result = extract_bracketed( $text, '<q>' );
1395 would correctly match something like this:
1397 $text = '<leftop: conj /and/ conj>';
1399 See also: C<"extract_quotelike"> and C<"extract_codeblock">.
1402 =head2 C<extract_variable>
1404 C<extract_variable> extracts any valid Perl variable or
1405 variable-involved expression, including scalars, arrays, hashes, array
1406 accesses, hash look-ups, method calls through objects, subroutine calls
1407 through subroutine references, etc.
1409 The subroutine takes up to two optional arguments:
1415 A string to be processed (C<$_> if the string is omitted or C<undef>)
1419 A string specifying a pattern to be matched as a prefix (which is to be
1420 skipped). If omitted, optional whitespace is skipped.
1424 On success in a list context, an array of 3 elements is returned. The
1431 the extracted variable, or variablish expression
1435 the remainder of the input text,
1439 the prefix substring (if any),
1443 On failure, all of these values (except the remaining text) are C<undef>.
1445 In a scalar context, C<extract_variable> returns just the complete
1446 substring that matched a variablish expression. C<undef> is returned on
1447 failure. In addition, the original input text has the returned substring
1448 (and any prefix) removed from it.
1450 In a void context, the input text just has the matched substring (and
1451 any specified prefix) removed.
1454 =head2 C<extract_tagged>
1456 C<extract_tagged> extracts and segments text between (balanced)
1459 The subroutine takes up to five optional arguments:
1465 A string to be processed (C<$_> if the string is omitted or C<undef>)
1469 A string specifying a pattern to be matched as the opening tag.
1470 If the pattern string is omitted (or C<undef>) then a pattern
1471 that matches any standard XML tag is used.
1475 A string specifying a pattern to be matched at the closing tag.
1476 If the pattern string is omitted (or C<undef>) then the closing
1477 tag is constructed by inserting a C</> after any leading bracket
1478 characters in the actual opening tag that was matched (I<not> the pattern
1479 that matched the tag). For example, if the opening tag pattern
1480 is specified as C<'{{\w+}}'> and actually matched the opening tag
1481 C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">.
1485 A string specifying a pattern to be matched as a prefix (which is to be
1486 skipped). If omitted, optional whitespace is skipped.
1490 A hash reference containing various parsing options (see below)
1494 The various options that can be specified are:
1498 =item C<reject =E<gt> $listref>
1500 The list reference contains one or more strings specifying patterns
1501 that must I<not> appear within the tagged text.
1503 For example, to extract
1504 an HTML link (which should not contain nested links) use:
1506 extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} );
1508 =item C<ignore =E<gt> $listref>
1510 The list reference contains one or more strings specifying patterns
1511 that are I<not> be be treated as nested tags within the tagged text
1512 (even if they would match the start tag pattern).
1514 For example, to extract an arbitrary XML tag, but ignore "empty" elements:
1516 extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} );
1518 (also see L<"gen_delimited_pat"> below).
1521 =item C<fail =E<gt> $str>
1523 The C<fail> option indicates the action to be taken if a matching end
1524 tag is not encountered (i.e. before the end of the string or some
1525 C<reject> pattern matches). By default, a failure to match a closing
1526 tag causes C<extract_tagged> to immediately fail.
1528 However, if the string value associated with <reject> is "MAX", then
1529 C<extract_tagged> returns the complete text up to the point of failure.
1530 If the string is "PARA", C<extract_tagged> returns only the first paragraph
1531 after the tag (up to the first line that is either empty or contains
1532 only whitespace characters).
1533 If the string is "", the the default behaviour (i.e. failure) is reinstated.
1535 For example, suppose the start tag "/para" introduces a paragraph, which then
1536 continues until the next "/endpara" tag or until another "/para" tag is
1539 $text = "/para line 1\n\nline 3\n/para line 4";
1541 extract_tagged($text, '/para', '/endpara', undef,
1542 {reject => '/para', fail => MAX );
1544 # EXTRACTED: "/para line 1\n\nline 3\n"
1546 Suppose instead, that if no matching "/endpara" tag is found, the "/para"
1547 tag refers only to the immediately following paragraph:
1549 $text = "/para line 1\n\nline 3\n/para line 4";
1551 extract_tagged($text, '/para', '/endpara', undef,
1552 {reject => '/para', fail => MAX );
1554 # EXTRACTED: "/para line 1\n"
1556 Note that the specified C<fail> behaviour applies to nested tags as well.
1560 On success in a list context, an array of 6 elements is returned. The elements are:
1566 the extracted tagged substring (including the outermost tags),
1570 the remainder of the input text,
1574 the prefix substring (if any),
1582 the text between the opening and closing tags
1586 the closing tag (or "" if no closing tag was found)
1590 On failure, all of these values (except the remaining text) are C<undef>.
1592 In a scalar context, C<extract_tagged> returns just the complete
1593 substring that matched a tagged text (including the start and end
1594 tags). C<undef> is returned on failure. In addition, the original input
1595 text has the returned substring (and any prefix) removed from it.
1597 In a void context, the input text just has the matched substring (and
1598 any specified prefix) removed.
1601 =head2 C<gen_extract_tagged>
1603 (Note: This subroutine is only available under Perl5.005)
1605 C<gen_extract_tagged> generates a new anonymous subroutine which
1606 extracts text between (balanced) specified tags. In other words,
1607 it generates a function identical in function to C<extract_tagged>.
1609 The difference between C<extract_tagged> and the anonymous
1610 subroutines generated by
1611 C<gen_extract_tagged>, is that those generated subroutines:
1617 do not have to reparse tag specification or parsing options every time
1618 they are called (whereas C<extract_tagged> has to effectively rebuild
1619 its tag parser on every call);
1623 make use of the new qr// construct to pre-compile the regexes they use
1624 (whereas C<extract_tagged> uses standard string variable interpolation
1625 to create tag-matching patterns).
1629 The subroutine takes up to four optional arguments (the same set as
1630 C<extract_tagged> except for the string to be processed). It returns
1631 a reference to a subroutine which in turn takes a single argument (the text to
1634 In other words, the implementation of C<extract_tagged> is exactly
1640 $extractor = gen_extract_tagged(@_);
1641 return $extractor->($text);
1644 (although C<extract_tagged> is not currently implemented that way, in order
1645 to preserve pre-5.005 compatibility).
1647 Using C<gen_extract_tagged> to create extraction functions for specific tags
1648 is a good idea if those functions are going to be called more than once, since
1649 their performance is typically twice as good as the more general-purpose
1653 =head2 C<extract_quotelike>
1655 C<extract_quotelike> attempts to recognize, extract, and segment any
1656 one of the various Perl quotes and quotelike operators (see
1657 L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket
1658 delimiters (for the quotelike operators), and trailing modifiers are
1659 all caught. For example, in:
1661 extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #'
1663 extract_quotelike ' "You said, \"Use sed\"." '
1665 extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; '
1667 extract_quotelike ' tr/\\\/\\\\/\\\//ds; '
1669 the full Perl quotelike operations are all extracted correctly.
1671 Note too that, when using the /x modifier on a regex, any comment
1672 containing the current pattern delimiter will cause the regex to be
1673 immediately terminated. In other words:
1676 (?i) # CASE INSENSITIVE
1677 [a-z_] # LEADING ALPHABETIC/UNDERSCORE
1678 [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS
1681 will be extracted as if it were:
1684 (?i) # CASE INSENSITIVE
1685 [a-z_] # LEADING ALPHABETIC/'
1687 This behaviour is identical to that of the actual compiler.
1689 C<extract_quotelike> takes two arguments: the text to be processed and
1690 a prefix to be matched at the very beginning of the text. If no prefix
1691 is specified, optional whitespace is the default. If no text is given,
1694 In a list context, an array of 11 elements is returned. The elements are:
1700 the extracted quotelike substring (including trailing modifiers),
1704 the remainder of the input text,
1708 the prefix substring (if any),
1712 the name of the quotelike operator (if any),
1716 the left delimiter of the first block of the operation,
1720 the text of the first block of the operation
1721 (that is, the contents of
1722 a quote, the regex of a match or substitution or the target list of a
1727 the right delimiter of the first block of the operation,
1731 the left delimiter of the second block of the operation
1732 (that is, if it is a C<s>, C<tr>, or C<y>),
1736 the text of the second block of the operation
1737 (that is, the replacement of a substitution or the translation list
1742 the right delimiter of the second block of the operation (if any),
1746 the trailing modifiers on the operation (if any).
1750 For each of the fields marked "(if any)" the default value on success is
1752 On failure, all of these values (except the remaining text) are C<undef>.
1755 In a scalar context, C<extract_quotelike> returns just the complete substring
1756 that matched a quotelike operation (or C<undef> on failure). In a scalar or
1757 void context, the input text has the same substring (and any specified
1762 # Remove the first quotelike literal that appears in text
1764 $quotelike = extract_quotelike($text,'.*?');
1766 # Replace one or more leading whitespace-separated quotelike
1767 # literals in $_ with "<QLL>"
1769 do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@;
1772 # Isolate the search pattern in a quotelike operation from $text
1774 ($op,$pat) = (extract_quotelike $text)[3,5];
1777 print "search pattern: $pat\n";
1781 print "$op is not a pattern matching operation\n";
1785 =head2 C<extract_quotelike> and "here documents"
1787 C<extract_quotelike> can successfully extract "here documents" from an input
1788 string, but with an important caveat in list contexts.
1790 Unlike other types of quote-like literals, a here document is rarely
1791 a contiguous substring. For example, a typical piece of code using
1792 here document might look like this:
1795 This is the message.
1799 Given this as an input string in a scalar context, C<extract_quotelike>
1800 would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG",
1801 leaving the string " || die;\nexit;" in the original variable. In other words,
1802 the two separate pieces of the here document are successfully extracted and
1805 In a list context, C<extract_quotelike> would return the list
1811 "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document,
1812 including fore and aft delimiters),
1816 " || die;\nexit;" (i.e. the remainder of the input text, concatenated),
1820 "" (i.e. the prefix substring -- trivial in this case),
1824 "<<" (i.e. the "name" of the quotelike operator)
1828 "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes),
1832 "This is the message.\n" (i.e. the text of the here document),
1836 "EOMSG" (i.e. the right delimiter of the here document),
1840 "" (a here document has no second left delimiter, second text, second right
1841 delimiter, or trailing modifiers).
1845 However, the matching position of the input variable would be set to
1846 "exit;" (i.e. I<after> the closing delimiter of the here document),
1847 which would cause the earlier " || die;\nexit;" to be skipped in any
1848 sequence of code fragment extractions.
1850 To avoid this problem, when it encounters a here document whilst
1851 extracting from a modifiable string, C<extract_quotelike> silently
1852 rearranges the string to an equivalent piece of Perl:
1855 This is the message.
1860 in which the here document I<is> contiguous. It still leaves the
1861 matching position after the here document, but now the rest of the line
1862 on which the here document starts is not skipped.
1864 To prevent <extract_quotelike> from mucking about with the input in this way
1865 (this is the only case where a list-context C<extract_quotelike> does so),
1866 you can pass the input variable as an interpolated literal:
1868 $quotelike = extract_quotelike("$var");
1871 =head2 C<extract_codeblock>
1873 C<extract_codeblock> attempts to recognize and extract a balanced
1874 bracket delimited substring that may contain unbalanced brackets
1875 inside Perl quotes or quotelike operations. That is, C<extract_codeblock>
1876 is like a combination of C<"extract_bracketed"> and
1877 C<"extract_quotelike">.
1879 C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>:
1880 a text to process, a set of delimiter brackets to look for, and a prefix to
1881 match first. It also takes an optional fourth parameter, which allows the
1882 outermost delimiter brackets to be specified separately (see below).
1884 Omitting the first argument (input text) means process C<$_> instead.
1885 Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used.
1886 Omitting the third argument (prefix argument) implies optional whitespace at the start.
1887 Omitting the fourth argument (outermost delimiter brackets) indicates that the
1888 value of the second argument is to be used for the outermost delimiters.
1890 Once the prefix an dthe outermost opening delimiter bracket have been
1891 recognized, code blocks are extracted by stepping through the input text and
1892 trying the following alternatives in sequence:
1898 Try and match a closing delimiter bracket. If the bracket was the same
1899 species as the last opening bracket, return the substring to that
1900 point. If the bracket was mismatched, return an error.
1904 Try to match a quote or quotelike operator. If found, call
1905 C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return
1906 the error it returned. Otherwise go back to step 1.
1910 Try to match an opening delimiter bracket. If found, call
1911 C<extract_codeblock> recursively to eat the embedded block. If the
1912 recursive call fails, return an error. Otherwise, go back to step 1.
1916 Unconditionally match a bareword or any other single character, and
1917 then go back to step 1.
1924 # Find a while loop in the text
1926 if ($text =~ s/.*?while\s*\{/{/)
1928 $loop = "while " . extract_codeblock($text);
1931 # Remove the first round-bracketed list (which may include
1932 # round- or curly-bracketed code blocks or quotelike operators)
1934 extract_codeblock $text, "(){}", '[^(]*';
1937 The ability to specify a different outermost delimiter bracket is useful
1938 in some circumstances. For example, in the Parse::RecDescent module,
1939 parser actions which are to be performed only on a successful parse
1940 are specified using a C<E<lt>defer:...E<gt>> directive. For example:
1942 sentence: subject verb object
1943 <defer: {$::theVerb = $item{verb}} >
1945 Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code
1946 within the C<E<lt>defer:...E<gt>> directive, but there's a problem.
1948 A deferred action like this:
1950 <defer: {if ($count>10) {$count--}} >
1952 will be incorrectly parsed as:
1954 <defer: {if ($count>
1956 because the "less than" operator is interpreted as a closing delimiter.
1958 But, by extracting the directive using
1959 S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>>
1960 the '>' character is only treated as a delimited at the outermost
1961 level of the code block, so the directive is parsed correctly.
1963 =head2 C<extract_multiple>
1965 The C<extract_multiple> subroutine takes a string to be processed and a
1966 list of extractors (subroutines or regular expressions) to apply to that string.
1968 In an array context C<extract_multiple> returns an array of substrings
1969 of the original string, as extracted by the specified extractors.
1970 In a scalar context, C<extract_multiple> returns the first
1971 substring successfully extracted from the original string. In both
1972 scalar and void contexts the original string has the first successfully
1973 extracted substring removed from it. In all contexts
1974 C<extract_multiple> starts at the current C<pos> of the string, and
1975 sets that C<pos> appropriately after it matches.
1977 Hence, the aim of of a call to C<extract_multiple> in a list context
1978 is to split the processed string into as many non-overlapping fields as
1979 possible, by repeatedly applying each of the specified extractors
1980 to the remainder of the string. Thus C<extract_multiple> is
1981 a generalized form of Perl's C<split> subroutine.
1983 The subroutine takes up to four optional arguments:
1989 A string to be processed (C<$_> if the string is omitted or C<undef>)
1993 A reference to a list of subroutine references and/or qr// objects and/or
1994 literal strings and/or hash references, specifying the extractors
1995 to be used to split the string. If this argument is omitted (or
1999 sub { extract_variable($_[0], '') },
2000 sub { extract_quotelike($_[0],'') },
2001 sub { extract_codeblock($_[0],'{}','') },
2009 An number specifying the maximum number of fields to return. If this
2010 argument is omitted (or C<undef>), split continues as long as possible.
2012 If the third argument is I<N>, then extraction continues until I<N> fields
2013 have been successfully extracted, or until the string has been completely
2016 Note that in scalar and void contexts the value of this argument is
2017 automatically reset to 1 (under C<-w>, a warning is issued if the argument
2022 A value indicating whether unmatched substrings (see below) within the
2023 text should be skipped or returned as fields. If the value is true,
2024 such substrings are skipped. Otherwise, they are returned.
2028 The extraction process works by applying each extractor in
2029 sequence to the text string.
2031 If the extractor is a subroutine it is called in a list context and is
2032 expected to return a list of a single element, namely the extracted
2033 text. It may optionally also return two further arguments: a string
2034 representing the text left after extraction (like $' for a pattern
2035 match), and a string representing any prefix skipped before the
2036 extraction (like $` in a pattern match). Note that this is designed
2037 to facilitate the use of other Text::Balanced subroutines with
2038 C<extract_multiple>. Note too that the value returned by an extractor
2039 subroutine need not bear any relationship to the corresponding substring
2040 of the original text (see examples below).
2042 If the extractor is a precompiled regular expression or a string,
2043 it is matched against the text in a scalar context with a leading
2044 '\G' and the gc modifiers enabled. The extracted value is either
2045 $1 if that variable is defined after the match, or else the
2046 complete match (i.e. $&).
2048 If the extractor is a hash reference, it must contain exactly one element.
2049 The value of that element is one of the
2050 above extractor types (subroutine reference, regular expression, or string).
2051 The key of that element is the name of a class into which the successful
2052 return value of the extractor will be blessed.
2054 If an extractor returns a defined value, that value is immediately
2055 treated as the next extracted field and pushed onto the list of fields.
2056 If the extractor was specified in a hash reference, the field is also
2057 blessed into the appropriate class,
2059 If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is
2060 assumed to have failed to extract.
2061 If none of the extractor subroutines succeeds, then one
2062 character is extracted from the start of the text and the extraction
2063 subroutines reapplied. Characters which are thus removed are accumulated and
2064 eventually become the next field (unless the fourth argument is true, in which
2065 case they are discarded).
2067 For example, the following extracts substrings that are valid Perl variables:
2069 @fields = extract_multiple($text,
2070 [ sub { extract_variable($_[0]) } ],
2073 This example separates a text into fields which are quote delimited,
2074 curly bracketed, and anything else. The delimited and bracketed
2075 parts are also blessed to identify them (the "anything else" is unblessed):
2077 @fields = extract_multiple($text,
2079 { Delim => sub { extract_delimited($_[0],q{'"}) } },
2080 { Brack => sub { extract_bracketed($_[0],'{}') } },
2083 This call extracts the next single substring that is a valid Perl quotelike
2084 operator (and removes it from $text):
2086 $quotelike = extract_multiple($text,
2088 sub { extract_quotelike($_[0]) },
2091 Finally, here is yet another way to do comma-separated value parsing:
2093 @fields = extract_multiple($csv_text,
2095 sub { extract_delimited($_[0],q{'"}) },
2100 The list in the second argument means:
2101 I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">.
2102 The undef third argument means:
2103 I<"...as many times as possible...">,
2104 and the true value in the fourth argument means
2105 I<"...discarding anything else that appears (i.e. the commas)">.
2107 If you wanted the commas preserved as separate fields (i.e. like split
2108 does if your split pattern has capturing parentheses), you would
2109 just make the last parameter undefined (or remove it).
2112 =head2 C<gen_delimited_pat>
2114 The C<gen_delimited_pat> subroutine takes a single (string) argument and
2115 > builds a Friedl-style optimized regex that matches a string delimited
2116 by any one of the characters in the single argument. For example:
2118 gen_delimited_pat(q{'"})
2122 (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\')
2124 Note that the specified delimiters are automatically quotemeta'd.
2126 A typical use of C<gen_delimited_pat> would be to build special purpose tags
2127 for C<extract_tagged>. For example, to properly ignore "empty" XML elements
2128 (which might contain quoted strings):
2130 my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>';
2132 extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} );
2135 C<gen_delimited_pat> may also be called with an optional second argument,
2136 which specifies the "escape" character(s) to be used for each delimiter.
2137 For example to match a Pascal-style string (where ' is the delimiter
2138 and '' is a literal ' within the string):
2140 gen_delimited_pat(q{'},q{'});
2142 Different escape characters can be specified for different delimiters.
2143 For example, to specify that '/' is the escape for single quotes
2144 and '%' is the escape for double quotes:
2146 gen_delimited_pat(q{'"},q{/%});
2148 If more delimiters than escape chars are specified, the last escape char
2149 is used for the remaining delimiters.
2150 If no escape char is specified for a given specified delimiter, '\' is used.
2153 C<gen_delimited_pat> was previously called
2154 C<delimited_pat>. That name may still be used, but is now deprecated.
2159 In a list context, all the functions return C<(undef,$original_text)>
2160 on failure. In a scalar context, failure is indicated by returning C<undef>
2161 (in this case the input text is not modified in any way).
2163 In addition, on failure in I<any> context, the C<$@> variable is set.
2164 Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed
2166 Accessing C<$@-E<gt>{pos}> returns the offset into the original string at
2167 which the error was detected (although not necessarily where it occurred!)
2168 Printing C<$@> directly produces the error message, with the offset appended.
2169 On success, the C<$@> variable is guaranteed to be C<undef>.
2171 The available diagnostics are:
2175 =item C<Did not find a suitable bracket: "%s">
2177 The delimiter provided to C<extract_bracketed> was not one of
2178 C<'()[]E<lt>E<gt>{}'>.
2180 =item C<Did not find prefix: /%s/>
2182 A non-optional prefix was specified but wasn't found at the start of the text.
2184 =item C<Did not find opening bracket after prefix: "%s">
2186 C<extract_bracketed> or C<extract_codeblock> was expecting a
2187 particular kind of bracket at the start of the text, and didn't find it.
2189 =item C<No quotelike operator found after prefix: "%s">
2191 C<extract_quotelike> didn't find one of the quotelike operators C<q>,
2192 C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring
2195 =item C<Unmatched closing bracket: "%c">
2197 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered
2198 a closing bracket where none was expected.
2200 =item C<Unmatched opening bracket(s): "%s">
2202 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran
2203 out of characters in the text before closing one or more levels of nested
2206 =item C<Unmatched embedded quote (%s)>
2208 C<extract_bracketed> attempted to match an embedded quoted substring, but
2209 failed to find a closing quote to match it.
2211 =item C<Did not find closing delimiter to match '%s'>
2213 C<extract_quotelike> was unable to find a closing delimiter to match the
2214 one that opened the quote-like operation.
2216 =item C<Mismatched closing bracket: expected "%c" but found "%s">
2218 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found
2219 a valid bracket delimiter, but it was the wrong species. This usually
2220 indicates a nesting error, but may indicate incorrect quoting or escaping.
2222 =item C<No block delimiter found after quotelike "%s">
2224 C<extract_quotelike> or C<extract_codeblock> found one of the
2225 quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y>
2226 without a suitable block after it.
2228 =item C<Did not find leading dereferencer>
2230 C<extract_variable> was expecting one of '$', '@', or '%' at the start of
2231 a variable, but didn't find any of them.
2233 =item C<Bad identifier after dereferencer>
2235 C<extract_variable> found a '$', '@', or '%' indicating a variable, but that
2236 character was not followed by a legal Perl identifier.
2238 =item C<Did not find expected opening bracket at %s>
2240 C<extract_codeblock> failed to find any of the outermost opening brackets
2241 that were specified.
2243 =item C<Improperly nested codeblock at %s>
2245 A nested code block was found that started with a delimiter that was specified
2246 as being only to be used as an outermost bracket.
2248 =item C<Missing second block for quotelike "%s">
2250 C<extract_codeblock> or C<extract_quotelike> found one of the
2251 quotelike operators C<s>, C<tr> or C<y> followed by only one block.
2253 =item C<No match found for opening bracket>
2255 C<extract_codeblock> failed to find a closing bracket to match the outermost
2258 =item C<Did not find opening tag: /%s/>
2260 C<extract_tagged> did not find a suitable opening tag (after any specified
2261 prefix was removed).
2263 =item C<Unable to construct closing tag to match: /%s/>
2265 C<extract_tagged> matched the specified opening tag and tried to
2266 modify the matched text to produce a matching closing tag (because
2267 none was specified). It failed to generate the closing tag, almost
2268 certainly because the opening tag did not start with a
2269 bracket of some kind.
2271 =item C<Found invalid nested tag: %s>
2273 C<extract_tagged> found a nested tag that appeared in the "reject" list
2274 (and the failure mode was not "MAX" or "PARA").
2276 =item C<Found unbalanced nested tag: %s>
2278 C<extract_tagged> found a nested opening tag that was not matched by a
2279 corresponding nested closing tag (and the failure mode was not "MAX" or "PARA").
2281 =item C<Did not find closing tag>
2283 C<extract_tagged> reached the end of the text without finding a closing tag
2284 to match the original opening tag (and the failure mode was not
2295 Damian Conway (damian@conway.org)
2298 =head1 BUGS AND IRRITATIONS
2300 There are undoubtedly serious bugs lurking somewhere in this code, if
2301 only because parts of it give the impression of understanding a great deal
2302 more about Perl than they really do.
2304 Bug reports and other feedback are most welcome.
2309 Copyright (c) 1997-2001, Damian Conway. All Rights Reserved.
2310 This module is free software. It may be used, redistributed
2311 and/or modified under the same terms as Perl itself.