1 # EXTRACT VARIOUSLY DELIMITED TEXT SEQUENCES FROM STRINGS.
2 # FOR FULL DOCUMENTATION SEE Balanced.pod
7 package Text::Balanced;
11 use vars qw { $VERSION @ISA %EXPORT_TAGS };
13 use version; $VERSION = qv('1.99.1_1');
14 @ISA = qw ( Exporter );
16 %EXPORT_TAGS = ( ALL => [ qw(
31 Exporter::export_ok_tags('ALL');
35 sub _match_bracketed($$$$$$);
36 sub _match_variable($$);
37 sub _match_codeblock($$$$$$$);
38 sub _match_quotelike($$$$);
40 # HANDLE RETURN VALUES IN VARIOUS CONTEXTS
43 my ($message, $pos) = @_;
44 $@ = bless { error=>$message, pos=>$pos }, "Text::Balanced::ErrorMsg";
49 my ($wantarray, $textref, $message, $pos) = @_;
50 _failmsg $message, $pos if $message;
51 return (undef,$$textref,undef) if $wantarray;
58 my ($wantarray,$textref) = splice @_, 0, 2;
59 my ($extrapos, $extralen) = @_>18 ? splice(@_, -2, 2) : (0,0);
60 my ($startlen, $oppos) = @_[5,6];
61 my $remainderpos = $_[2];
65 while (my ($from, $len) = splice @_, 0, 2)
67 push @res, substr($$textref,$from,$len);
69 if ($extralen) { # CORRECT FILLET
70 my $extra = substr($res[0], $extrapos-$oppos, $extralen, "\n");
71 $res[1] = "$extra$res[1]";
72 eval { substr($$textref,$remainderpos,0) = $extra;
73 substr($$textref,$extrapos,$extralen,"\n")} ;
74 #REARRANGE HERE DOC AND FILLET IF POSSIBLE
75 pos($$textref) = $remainderpos-$extralen+1; # RESET \G
78 pos($$textref) = $remainderpos; # RESET \G
84 my $match = substr($$textref,$_[0],$_[1]);
85 substr($match,$extrapos-$_[0]-$startlen,$extralen,"") if $extralen;
87 ? substr($$textref, $extrapos, $extralen)."\n" : "";
88 eval {substr($$textref,$_[4],$_[1]+$_[5])=$extra} ; #CHOP OUT PREFIX & MATCH, IF POSSIBLE
89 pos($$textref) = $_[4]; # RESET \G
94 # BUILD A PATTERN MATCHING A SIMPLE DELIMITED STRING
96 sub gen_delimited_pat($;$) # ($delimiters;$escapes)
98 my ($dels, $escs) = @_;
99 return "" unless $dels =~ /\S/;
100 $escs = '\\' unless $escs;
101 $escs .= substr($escs,-1) x (length($dels)-length($escs));
104 for ($i=0; $i<length $dels; $i++)
106 my $del = quotemeta substr($dels,$i,1);
107 my $esc = quotemeta substr($escs,$i,1);
110 push @pat, "$del(?:[^$del]*(?:(?:$del$del)[^$del]*)*)$del";
114 push @pat, "$del(?:[^$esc$del]*(?:$esc.[^$esc$del]*)*)$del";
117 my $pat = join '|', @pat;
121 *delimited_pat = \&gen_delimited_pat;
124 # THE EXTRACTION FUNCTIONS
126 sub extract_delimited (;$$$$)
128 my $textref = defined $_[0] ? \$_[0] : \$_;
129 my $wantarray = wantarray;
130 my $del = defined $_[1] ? $_[1] : qq{\'\"\`};
131 my $pre = defined $_[2] ? $_[2] : '\s*';
132 my $esc = defined $_[3] ? $_[3] : qq{\\};
133 my $pat = gen_delimited_pat($del, $esc);
134 my $startpos = pos $$textref || 0;
135 return _fail($wantarray, $textref, "Not a delimited pattern", 0)
136 unless $$textref =~ m/\G($pre)($pat)/gc;
137 my $prelen = length($1);
138 my $matchpos = $startpos+$prelen;
139 my $endpos = pos $$textref;
140 return _succeed $wantarray, $textref,
141 $matchpos, $endpos-$matchpos, # MATCH
142 $endpos, length($$textref)-$endpos, # REMAINDER
143 $startpos, $prelen; # PREFIX
146 sub extract_bracketed (;$$$)
148 my $textref = defined $_[0] ? \$_[0] : \$_;
149 my $ldel = defined $_[1] ? $_[1] : '{([<';
150 my $pre = defined $_[2] ? $_[2] : '\s*';
151 my $wantarray = wantarray;
154 $ldel =~ s/'//g and $qdel .= q{'};
155 $ldel =~ s/"//g and $qdel .= q{"};
156 $ldel =~ s/`//g and $qdel .= q{`};
157 $ldel =~ s/q//g and $quotelike = 1;
158 $ldel =~ tr/[](){}<>\0-\377/[[(({{<</ds;
160 unless ($rdel =~ tr/[({</])}>/)
162 return _fail $wantarray, $textref,
163 "Did not find a suitable bracket in delimiter: \"$_[1]\"",
167 $ldel = join('|', map { quotemeta $_ } split('', $ldel));
168 $rdel = join('|', map { quotemeta $_ } split('', $rdel));
171 my $startpos = pos $$textref || 0;
172 my @match = _match_bracketed($textref,$pre, $ldel, $qdel, $quotelike, $rdel);
174 return _fail ($wantarray, $textref) unless @match;
176 return _succeed ( $wantarray, $textref,
177 $match[2], $match[5]+2, # MATCH
178 @match[8,9], # REMAINDER
179 @match[0,1], # PREFIX
183 sub _match_bracketed($$$$$$) # $textref, $pre, $ldel, $qdel, $quotelike, $rdel
185 my ($textref, $pre, $ldel, $qdel, $quotelike, $rdel) = @_;
186 my ($startpos, $ldelpos, $endpos) = (pos $$textref = pos $$textref||0);
187 unless ($$textref =~ m/\G$pre/gc)
189 _failmsg "Did not find prefix: /$pre/", $startpos;
193 $ldelpos = pos $$textref;
195 unless ($$textref =~ m/\G($ldel)/gc)
197 _failmsg "Did not find opening bracket after prefix: \"$pre\"",
199 pos $$textref = $startpos;
203 my @nesting = ( $1 );
204 my $textlen = length $$textref;
205 while (pos $$textref < $textlen)
207 next if $$textref =~ m/\G\\./gcs;
209 if ($$textref =~ m/\G($ldel)/gc)
213 elsif ($$textref =~ m/\G($rdel)/gc)
215 my ($found, $brackettype) = ($1, $1);
218 _failmsg "Unmatched closing bracket: \"$found\"",
220 pos $$textref = $startpos;
223 my $expected = pop(@nesting);
224 $expected =~ tr/({[</)}]>/;
225 if ($expected ne $brackettype)
227 _failmsg qq{Mismatched closing bracket: expected "$expected" but found "$found"},
229 pos $$textref = $startpos;
232 last if $#nesting < 0;
234 elsif ($qdel && $$textref =~ m/\G([$qdel])/gc)
236 $$textref =~ m/\G[^\\$1]*(?:\\.[^\\$1]*)*(\Q$1\E)/gsc and next;
237 _failmsg "Unmatched embedded quote ($1)",
239 pos $$textref = $startpos;
242 elsif ($quotelike && _match_quotelike($textref,"",1,0))
247 else { $$textref =~ m/\G(?:[a-zA-Z0-9]+|.)/gcs }
251 _failmsg "Unmatched opening bracket(s): "
252 . join("..",@nesting)."..",
254 pos $$textref = $startpos;
258 $endpos = pos $$textref;
261 $startpos, $ldelpos-$startpos, # PREFIX
262 $ldelpos, 1, # OPENING BRACKET
263 $ldelpos+1, $endpos-$ldelpos-2, # CONTENTS
264 $endpos-1, 1, # CLOSING BRACKET
265 $endpos, length($$textref)-$endpos, # REMAINDER
271 my $brack = reverse $_[0];
272 $brack =~ tr/[({</])}>/;
276 my $XMLNAME = q{[a-zA-Z_:][a-zA-Z0-9_:.-]*};
278 sub extract_tagged (;$$$$$) # ($text, $opentag, $closetag, $pre, \%options)
280 my $textref = defined $_[0] ? \$_[0] : \$_;
283 my $pre = defined $_[3] ? $_[3] : '\s*';
284 my %options = defined $_[4] ? %{$_[4]} : ();
285 my $omode = defined $options{fail} ? $options{fail} : '';
286 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
287 : defined($options{reject}) ? $options{reject}
290 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
291 : defined($options{ignore}) ? $options{ignore}
295 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
298 my @match = _match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
300 return _fail(wantarray, $textref) unless @match;
301 return _succeed wantarray, $textref,
302 $match[2], $match[3]+$match[5]+$match[7], # MATCH
303 @match[8..9,0..1,2..7]; # REM, PRE, BITS
306 sub _match_tagged # ($$$$$$$)
308 my ($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore) = @_;
311 my ($startpos, $opentagpos, $textpos, $parapos, $closetagpos, $endpos) = ( pos($$textref) = pos($$textref)||0 );
313 unless ($$textref =~ m/\G($pre)/gc)
315 _failmsg "Did not find prefix: /$pre/", pos $$textref;
319 $opentagpos = pos($$textref);
321 unless ($$textref =~ m/\G$ldel/gc)
323 _failmsg "Did not find opening tag: /$ldel/", pos $$textref;
327 $textpos = pos($$textref);
331 $rdelspec = substr($$textref, $-[0], $+[0] - $-[0]);
332 unless ($rdelspec =~ s/\A([[(<{]+)($XMLNAME).*/ quotemeta "$1\/$2". _revbracket($1) /oes)
334 _failmsg "Unable to construct closing tag to match: $rdel",
341 $rdelspec = eval "qq{$rdel}" || do {
343 for (qw,~ ! ^ & * ) _ + - = } ] : " ; ' > . ? / | ',)
344 { next if $rdel =~ /\Q$_/; $del = $_; last }
347 croak "Can't interpolate right delimiter $rdel"
349 eval "qq$del$rdel$del";
353 while (pos($$textref) < length($$textref))
355 next if $$textref =~ m/\G\\./gc;
357 if ($$textref =~ m/\G(\n[ \t]*\n)/gc )
359 $parapos = pos($$textref) - length($1)
360 unless defined $parapos;
362 elsif ($$textref =~ m/\G($rdelspec)/gc )
364 $closetagpos = pos($$textref)-length($1);
367 elsif ($ignore && $$textref =~ m/\G(?:$ignore)/gc)
371 elsif ($bad && $$textref =~ m/\G($bad)/gcs)
373 pos($$textref) -= length($1); # CUT OFF WHATEVER CAUSED THE SHORTNESS
374 goto short if ($omode eq 'PARA' || $omode eq 'MAX');
375 _failmsg "Found invalid nested tag: $1", pos $$textref;
378 elsif ($$textref =~ m/\G($ldel)/gc)
381 pos($$textref) -= length($tag); # REWIND TO NESTED TAG
382 unless (_match_tagged(@_)) # MATCH NESTED TAG
384 goto short if $omode eq 'PARA' || $omode eq 'MAX';
385 _failmsg "Found unbalanced nested tag: $tag",
390 else { $$textref =~ m/./gcs }
394 $closetagpos = pos($$textref);
395 goto matched if $omode eq 'MAX';
396 goto failed unless $omode eq 'PARA';
398 if (defined $parapos) { pos($$textref) = $parapos }
399 else { $parapos = pos($$textref) }
402 $startpos, $opentagpos-$startpos, # PREFIX
403 $opentagpos, $textpos-$opentagpos, # OPENING TAG
404 $textpos, $parapos-$textpos, # TEXT
405 $parapos, 0, # NO CLOSING TAG
406 $parapos, length($$textref)-$parapos, # REMAINDER
410 $endpos = pos($$textref);
412 $startpos, $opentagpos-$startpos, # PREFIX
413 $opentagpos, $textpos-$opentagpos, # OPENING TAG
414 $textpos, $closetagpos-$textpos, # TEXT
415 $closetagpos, $endpos-$closetagpos, # CLOSING TAG
416 $endpos, length($$textref)-$endpos, # REMAINDER
420 _failmsg "Did not find closing tag", pos $$textref unless $@;
421 pos($$textref) = $startpos;
425 sub extract_variable (;$$)
427 my $textref = defined $_[0] ? \$_[0] : \$_;
428 return ("","","") unless defined $$textref;
429 my $pre = defined $_[1] ? $_[1] : '\s*';
431 my @match = _match_variable($textref,$pre);
433 return _fail wantarray, $textref unless @match;
435 return _succeed wantarray, $textref,
436 @match[2..3,4..5,0..1]; # MATCH, REMAINDER, PREFIX
439 sub _match_variable($$)
444 my ($textref, $pre) = @_;
445 my $startpos = pos($$textref) = pos($$textref)||0;
446 unless ($$textref =~ m/\G($pre)/gc)
448 _failmsg "Did not find prefix: /$pre/", pos $$textref;
451 my $varpos = pos($$textref);
452 unless ($$textref =~ m{\G\$\s*(?!::)(\d+|[][&`'+*./|,";%=~:?!\@<>()-]|\^[a-z]?)}gci)
454 unless ($$textref =~ m/\G((\$#?|[*\@\%]|\\&)+)/gc)
456 _failmsg "Did not find leading dereferencer", pos $$textref;
457 pos $$textref = $startpos;
462 unless ($$textref =~ m/\G\s*(?:::|')?(?:[_a-z]\w*(?:::|'))*[_a-z]\w*/gci
463 or _match_codeblock($textref, "", '\{', '\}', '\{', '\}', 0)
464 or $deref eq '$#' or $deref eq '$$' )
466 _failmsg "Bad identifier after dereferencer", pos $$textref;
467 pos $$textref = $startpos;
474 next if $$textref =~ m/\G\s*(?:->)?\s*[{]\w+[}]/gc;
475 next if _match_codeblock($textref,
476 qr/\s*->\s*(?:[_a-zA-Z]\w+\s*)?/,
477 qr/[({[]/, qr/[)}\]]/,
478 qr/[({[]/, qr/[)}\]]/, 0);
479 next if _match_codeblock($textref,
480 qr/\s*/, qr/[{[]/, qr/[}\]]/,
481 qr/[{[]/, qr/[}\]]/, 0);
482 next if _match_variable($textref,'\s*->\s*');
483 next if $$textref =~ m/\G\s*->\s*\w+(?![{([])/gc;
487 my $endpos = pos($$textref);
488 return ($startpos, $varpos-$startpos,
489 $varpos, $endpos-$varpos,
490 $endpos, length($$textref)-$endpos
494 sub extract_codeblock (;$$$$$)
496 my $textref = defined $_[0] ? \$_[0] : \$_;
497 my $wantarray = wantarray;
498 my $ldel_inner = defined $_[1] ? $_[1] : '{';
499 my $pre = defined $_[2] ? $_[2] : '\s*';
500 my $ldel_outer = defined $_[3] ? $_[3] : $ldel_inner;
502 my $rdel_inner = $ldel_inner;
503 my $rdel_outer = $ldel_outer;
505 for ($ldel_inner, $ldel_outer) { tr/[]()<>{}\0-\377/[[((<<{{/ds }
506 for ($rdel_inner, $rdel_outer) { tr/[]()<>{}\0-\377/]]))>>}}/ds }
507 for ($ldel_inner, $ldel_outer, $rdel_inner, $rdel_outer)
509 $_ = '('.join('|',map { quotemeta $_ } split('',$_)).')'
513 my @match = _match_codeblock($textref, $pre,
514 $ldel_outer, $rdel_outer,
515 $ldel_inner, $rdel_inner,
517 return _fail($wantarray, $textref) unless @match;
518 return _succeed($wantarray, $textref,
519 @match[2..3,4..5,0..1] # MATCH, REMAINDER, PREFIX
524 sub _match_codeblock($$$$$$$)
526 my ($textref, $pre, $ldel_outer, $rdel_outer, $ldel_inner, $rdel_inner, $rd) = @_;
527 my $startpos = pos($$textref) = pos($$textref) || 0;
528 unless ($$textref =~ m/\G($pre)/gc)
530 _failmsg qq{Did not match prefix /$pre/ at"} .
531 substr($$textref,pos($$textref),20) .
536 my $codepos = pos($$textref);
537 unless ($$textref =~ m/\G($ldel_outer)/gc) # OUTERMOST DELIMITER
539 _failmsg qq{Did not find expected opening bracket at "} .
540 substr($$textref,pos($$textref),20) .
543 pos $$textref = $startpos;
547 $closing =~ tr/([<{/)]>}/;
550 while (pos($$textref) < length($$textref))
553 if ($rd && $$textref =~ m#\G(\Q(?)\E|\Q(s?)\E|\Q(s)\E)#gc)
559 if ($$textref =~ m/\G\s*#.*/gc)
564 if ($$textref =~ m/\G\s*($rdel_outer)/gc)
566 unless ($matched = ($closing && $1 eq $closing) )
568 next if $1 eq '>'; # MIGHT BE A "LESS THAN"
569 _failmsg q{Mismatched closing bracket at "} .
570 substr($$textref,pos($$textref),20) .
571 qq{...". Expected '$closing'},
577 if (_match_variable($textref,'\s*') ||
578 _match_quotelike($textref,'\s*',$patvalid,$patvalid) )
585 # NEED TO COVER MANY MORE CASES HERE!!!
586 if ($$textref =~ m#\G\s*(?!$ldel_inner)
590 | (\*\*|&&|\|\||<<|>>)=?
591 | split|grep|map|return
599 if ( _match_codeblock($textref, '\s*', $ldel_inner, $rdel_inner, $ldel_inner, $rdel_inner, $rd) )
605 if ($$textref =~ m/\G\s*$ldel_outer/gc)
607 _failmsg q{Improperly nested codeblock at "} .
608 substr($$textref,pos($$textref),20) .
615 $$textref =~ m/\G\s*(\w+|[-=>]>|.|\Z)/gc;
617 continue { $@ = undef }
621 _failmsg 'No match found for opening bracket', pos $$textref
626 my $endpos = pos($$textref);
627 return ( $startpos, $codepos-$startpos,
628 $codepos, $endpos-$codepos,
629 $endpos, length($$textref)-$endpos,
635 'none' => '[cgimsox]*',
637 's' => '[cegimsox]*',
647 sub extract_quotelike (;$$)
649 my $textref = $_[0] ? \$_[0] : \$_;
650 my $wantarray = wantarray;
651 my $pre = defined $_[1] ? $_[1] : '\s*';
653 my @match = _match_quotelike($textref,$pre,1,0);
654 return _fail($wantarray, $textref) unless @match;
655 return _succeed($wantarray, $textref,
656 $match[2], $match[18]-$match[2], # MATCH
657 @match[18,19], # REMAINDER
658 @match[0,1], # PREFIX
659 @match[2..17], # THE BITS
660 @match[20,21], # ANY FILLET?
664 sub _match_quotelike($$$$) # ($textref, $prepat, $allow_raw_match)
666 my ($textref, $pre, $rawmatch, $qmark) = @_;
668 my ($textlen,$startpos,
670 $preld1pos,$ld1pos,$str1pos,$rd1pos,
671 $preld2pos,$ld2pos,$str2pos,$rd2pos,
672 $modpos) = ( length($$textref), pos($$textref) = pos($$textref) || 0 );
674 unless ($$textref =~ m/\G($pre)/gc)
676 _failmsg qq{Did not find prefix /$pre/ at "} .
677 substr($$textref, pos($$textref), 20) .
682 $oppos = pos($$textref);
684 my $initial = substr($$textref,$oppos,1);
686 if ($initial && $initial =~ m|^[\"\'\`]|
687 || $rawmatch && $initial =~ m|^/|
688 || $qmark && $initial =~ m|^\?|)
690 unless ($$textref =~ m/ \Q$initial\E [^\\$initial]* (\\.[^\\$initial]*)* \Q$initial\E /gcsx)
692 _failmsg qq{Did not find closing delimiter to match '$initial' at "} .
693 substr($$textref, $oppos, 20) .
696 pos $$textref = $startpos;
699 $modpos= pos($$textref);
702 if ($initial eq '/' || $initial eq '?')
704 $$textref =~ m/\G$mods{none}/gc
707 my $endpos = pos($$textref);
709 $startpos, $oppos-$startpos, # PREFIX
710 $oppos, 0, # NO OPERATOR
711 $oppos, 1, # LEFT DEL
712 $oppos+1, $rd1pos-$oppos-1, # STR/PAT
713 $rd1pos, 1, # RIGHT DEL
714 $modpos, 0, # NO 2ND LDEL
715 $modpos, 0, # NO 2ND STR
716 $modpos, 0, # NO 2ND RDEL
717 $modpos, $endpos-$modpos, # MODIFIERS
718 $endpos, $textlen-$endpos, # REMAINDER
722 unless ($$textref =~ m{\G(\b(?:m|s|qq|qx|qw|q|qr|tr|y)\b(?=\s*\S)|<<)}gc)
724 _failmsg q{No quotelike operator found after prefix at "} .
725 substr($$textref, pos($$textref), 20) .
728 pos $$textref = $startpos;
733 $preld1pos = pos($$textref);
735 $ld1pos = pos($$textref);
737 if ($$textref =~ m{\G([A-Za-z_]\w*)}gc) {
740 elsif ($$textref =~ m{ \G ' ([^'\\]* (?:\\.[^'\\]*)*) '
741 | \G " ([^"\\]* (?:\\.[^"\\]*)*) "
742 | \G ` ([^`\\]* (?:\\.[^`\\]*)*) `
749 my $extrapos = pos($$textref);
750 $$textref =~ m{.*\n}gc;
751 $str1pos = pos($$textref)--;
752 unless ($$textref =~ m{.*?\n(?=\Q$label\E\n)}gc) {
753 _failmsg qq{Missing here doc terminator ('$label') after "} .
754 substr($$textref, $startpos, 20) .
757 pos $$textref = $startpos;
760 $rd1pos = pos($$textref);
761 $$textref =~ m{\Q$label\E\n}gc;
762 $ld2pos = pos($$textref);
764 $startpos, $oppos-$startpos, # PREFIX
765 $oppos, length($op), # OPERATOR
766 $ld1pos, $extrapos-$ld1pos, # LEFT DEL
767 $str1pos, $rd1pos-$str1pos, # STR/PAT
768 $rd1pos, $ld2pos-$rd1pos, # RIGHT DEL
769 $ld2pos, 0, # NO 2ND LDEL
770 $ld2pos, 0, # NO 2ND STR
771 $ld2pos, 0, # NO 2ND RDEL
772 $ld2pos, 0, # NO MODIFIERS
773 $ld2pos, $textlen-$ld2pos, # REMAINDER
774 $extrapos, $str1pos-$extrapos, # FILLETED BIT
778 $$textref =~ m/\G\s*/gc;
779 $ld1pos = pos($$textref);
780 $str1pos = $ld1pos+1;
782 unless ($$textref =~ m/\G(\S)/gc) # SHOULD USE LOOKAHEAD
784 _failmsg "No block delimiter found after quotelike $op",
786 pos $$textref = $startpos;
789 pos($$textref) = $ld1pos; # HAVE TO DO THIS BECAUSE LOOKAHEAD BROKEN
790 my ($ldel1, $rdel1) = ("\Q$1","\Q$1");
791 if ($ldel1 =~ /[[(<{]/)
793 $rdel1 =~ tr/[({</])}>/;
794 defined(_match_bracketed($textref,"",$ldel1,"","",$rdel1))
795 || do { pos $$textref = $startpos; return };
799 $$textref =~ /$ldel1[^\\$ldel1]*(\\.[^\\$ldel1]*)*$ldel1/gcs
800 || do { pos $$textref = $startpos; return };
802 $ld2pos = $rd1pos = pos($$textref)-1;
804 my $second_arg = $op =~ /s|tr|y/ ? 1 : 0;
808 if ($ldel1 =~ /[[(<{]/)
810 unless ($$textref =~ /\G\s*(\S)/gc) # SHOULD USE LOOKAHEAD
812 _failmsg "Missing second block for quotelike $op",
814 pos $$textref = $startpos;
817 $ldel2 = $rdel2 = "\Q$1";
818 $rdel2 =~ tr/[({</])}>/;
822 $ldel2 = $rdel2 = $ldel1;
824 $str2pos = $ld2pos+1;
826 if ($ldel2 =~ /[[(<{]/)
828 pos($$textref)--; # OVERCOME BROKEN LOOKAHEAD
829 defined(_match_bracketed($textref,"",$ldel2,"","",$rdel2))
830 || do { pos $$textref = $startpos; return };
834 $$textref =~ /[^\\$ldel2]*(\\.[^\\$ldel2]*)*$ldel2/gcs
835 || do { pos $$textref = $startpos; return };
837 $rd2pos = pos($$textref)-1;
841 $ld2pos = $str2pos = $rd2pos = $rd1pos;
844 $modpos = pos $$textref;
846 $$textref =~ m/\G($mods{$op})/gc;
847 my $endpos = pos $$textref;
850 $startpos, $oppos-$startpos, # PREFIX
851 $oppos, length($op), # OPERATOR
852 $ld1pos, 1, # LEFT DEL
853 $str1pos, $rd1pos-$str1pos, # STR/PAT
854 $rd1pos, 1, # RIGHT DEL
855 $ld2pos, $second_arg, # 2ND LDEL (MAYBE)
856 $str2pos, $rd2pos-$str2pos, # 2ND STR (MAYBE)
857 $rd2pos, $second_arg, # 2ND RDEL (MAYBE)
858 $modpos, $endpos-$modpos, # MODIFIERS
859 $endpos, $textlen-$endpos, # REMAINDER
865 sub { extract_variable($_[0], '') },
866 sub { extract_quotelike($_[0],'') },
867 sub { extract_codeblock($_[0],'{}','') },
870 sub extract_multiple (;$$$$) # ($text, $functions_ref, $max_fields, $ignoreunknown)
872 my $textref = defined($_[0]) ? \$_[0] : \$_;
874 my ($lastpos, $firstpos);
879 my @func = defined $_[1] ? @{$_[1]} : @{$def_func};
880 my $max = defined $_[2] && $_[2]>0 ? $_[2] : 1_000_000_000;
888 carp "extract_multiple reset maximal count to 1 in scalar context"
889 if $^W && defined($_[2]) && $max > 1;
898 foreach $func ( @func )
900 if (ref($func) eq 'HASH')
902 push @class, (keys %$func)[0];
903 $func = (values %$func)[0];
911 FIELD: while (pos($$textref) < length($$textref))
915 foreach my $i ( 0..$#func )
920 $lastpos = pos $$textref;
921 if (ref($func) eq 'CODE')
922 { ($field,$rem,$pref) = @bits = $func->($$textref) }
923 elsif (ref($func) eq 'Text::Balanced::Extractor')
924 { @bits = $field = $func->extract($$textref) }
925 elsif( $$textref =~ m/\G$func/gc )
926 { @bits = $field = defined($1)
928 : substr($$textref, $-[0], $+[0] - $-[0])
931 if (defined($field) && length($field))
935 if length($pref) && !defined($unkpos);
938 push @fields, substr($$textref, $unkpos, $lastpos-$unkpos).$pref;
939 $firstpos = $unkpos unless defined $firstpos;
941 last FIELD if @fields == $max;
945 ? bless (\$field, $class)
947 $firstpos = $lastpos unless defined $firstpos;
948 $lastpos = pos $$textref;
949 last FIELD if @fields == $max;
953 if ($$textref =~ /\G(.)/gcs)
955 $unkpos = pos($$textref)-1
956 unless $igunk || defined $unkpos;
962 push @fields, substr($$textref, $unkpos);
963 $firstpos = $unkpos unless defined $firstpos;
964 $lastpos = length $$textref;
969 pos $$textref = $lastpos;
970 return @fields if wantarray;
973 eval { substr($$textref,$firstpos,$lastpos-$firstpos)="";
974 pos $$textref = $firstpos };
979 sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options)
983 my $pre = defined $_[2] ? $_[2] : '\s*';
984 my %options = defined $_[3] ? %{$_[3]} : ();
985 my $omode = defined $options{fail} ? $options{fail} : '';
986 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
987 : defined($options{reject}) ? $options{reject}
990 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
991 : defined($options{ignore}) ? $options{ignore}
995 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
998 for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ }
1003 my $textref = defined $_[0] ? \$_[0] : \$_;
1004 my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
1006 return _fail(wantarray, $textref) unless @match;
1007 return _succeed wantarray, $textref,
1008 $match[2], $match[3]+$match[5]+$match[7], # MATCH
1009 @match[8..9,0..1,2..7]; # REM, PRE, BITS
1012 bless $closure, 'Text::Balanced::Extractor';
1015 package Text::Balanced::Extractor;
1017 sub extract($$) # ($self, $text)
1022 package Text::Balanced::ErrorMsg;
1024 use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" };
1032 Text::Balanced - Extract delimited text sequences from strings.
1037 use Text::Balanced qw (
1050 # Extract the initial substring of $text that is delimited by
1051 # two (unescaped) instances of the first character in $delim.
1053 ($extracted, $remainder) = extract_delimited($text,$delim);
1056 # Extract the initial substring of $text that is bracketed
1057 # with a delimiter(s) specified by $delim (where the string
1058 # in $delim contains one or more of '(){}[]<>').
1060 ($extracted, $remainder) = extract_bracketed($text,$delim);
1063 # Extract the initial substring of $text that is bounded by
1066 ($extracted, $remainder) = extract_tagged($text);
1069 # Extract the initial substring of $text that is bounded by
1070 # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags
1072 ($extracted, $remainder) =
1073 extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]});
1076 # Extract the initial substring of $text that represents a
1077 # Perl "quote or quote-like operation"
1079 ($extracted, $remainder) = extract_quotelike($text);
1082 # Extract the initial substring of $text that represents a block
1083 # of Perl code, bracketed by any of character(s) specified by $delim
1084 # (where the string $delim contains one or more of '(){}[]<>').
1086 ($extracted, $remainder) = extract_codeblock($text,$delim);
1089 # Extract the initial substrings of $text that would be extracted by
1090 # one or more sequential applications of the specified functions
1091 # or regular expressions
1093 @extracted = extract_multiple($text,
1094 [ \&extract_bracketed,
1095 \&extract_quotelike,
1096 \&some_other_extractor_sub,
1101 # Create a string representing an optimized pattern (a la Friedl)
1102 # that matches a substring delimited by any of the specified characters
1103 # (in this case: any type of quote or a slash)
1105 $patstring = gen_delimited_pat(q{'"`/});
1108 # Generate a reference to an anonymous sub that is just like extract_tagged
1109 # but pre-compiled and optimized for a specific pair of tags, and consequently
1110 # much faster (i.e. 3 times faster). It uses qr// for better performance on
1111 # repeated calls, so it only works under Perl 5.005 or later.
1113 $extract_head = gen_extract_tagged('<HEAD>','</HEAD>');
1115 ($extracted, $remainder) = $extract_head->($text);
1120 The various C<extract_...> subroutines may be used to
1121 extract a delimited substring, possibly after skipping a
1122 specified prefix string. By default, that prefix is
1123 optional whitespace (C</\s*/>), but you can change it to whatever
1124 you wish (see below).
1126 The substring to be extracted must appear at the
1127 current C<pos> location of the string's variable
1128 (or at index zero, if no C<pos> position is defined).
1129 In other words, the C<extract_...> subroutines I<don't>
1130 extract the first occurrence of a substring anywhere
1131 in a string (like an unanchored regex would). Rather,
1132 they extract an occurrence of the substring appearing
1133 immediately at the current matching position in the
1134 string (like a C<\G>-anchored regex would).
1138 =head2 General behaviour in list contexts
1140 In a list context, all the subroutines return a list, the first three
1141 elements of which are always:
1147 The extracted string, including the specified delimiters.
1148 If the extraction fails C<undef> is returned.
1152 The remainder of the input string (i.e. the characters after the
1153 extracted string). On failure, the entire string is returned.
1157 The skipped prefix (i.e. the characters before the extracted string).
1158 On failure, C<undef> is returned.
1162 Note that in a list context, the contents of the original input text (the first
1163 argument) are not modified in any way.
1165 However, if the input text was passed in a variable, that variable's
1166 C<pos> value is updated to point at the first character after the
1167 extracted text. That means that in a list context the various
1168 subroutines can be used much like regular expressions. For example:
1170 while ( $next = (extract_quotelike($text))[0] )
1172 # process next quote-like (in $next)
1176 =head2 General behaviour in scalar and void contexts
1178 In a scalar context, the extracted string is returned, having first been
1179 removed from the input text. Thus, the following code also processes
1180 each quote-like operation, but actually removes them from $text:
1182 while ( $next = extract_quotelike($text) )
1184 # process next quote-like (in $next)
1187 Note that if the input text is a read-only string (i.e. a literal),
1188 no attempt is made to remove the extracted text.
1190 In a void context the behaviour of the extraction subroutines is
1191 exactly the same as in a scalar context, except (of course) that the
1192 extracted substring is not returned.
1194 =head2 A note about prefixes
1196 Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.)
1197 This can bite you if you're expecting a prefix specification like
1198 '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix
1199 pattern will only succeed if the <H1> tag is on the current line, since
1200 . normally doesn't match newlines.
1202 To overcome this limitation, you need to turn on /s matching within
1203 the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)'
1206 =head2 C<extract_delimited>
1208 The C<extract_delimited> function formalizes the common idiom
1209 of extracting a single-character-delimited substring from the start of
1210 a string. For example, to extract a single-quote delimited string, the
1211 following code is typically used:
1213 ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s;
1216 but with C<extract_delimited> it can be simplified to:
1218 ($extracted,$remainder) = extract_delimited($text, "'");
1220 C<extract_delimited> takes up to four scalars (the input text, the
1221 delimiters, a prefix pattern to be skipped, and any escape characters)
1222 and extracts the initial substring of the text that
1223 is appropriately delimited. If the delimiter string has multiple
1224 characters, the first one encountered in the text is taken to delimit
1226 The third argument specifies a prefix pattern that is to be skipped
1227 (but must be present!) before the substring is extracted.
1228 The final argument specifies the escape character to be used for each
1231 All arguments are optional. If the escape characters are not specified,
1232 every delimiter is escaped with a backslash (C<\>).
1233 If the prefix is not specified, the
1234 pattern C<'\s*'> - optional whitespace - is used. If the delimiter set
1235 is also not specified, the set C</["'`]/> is used. If the text to be processed
1236 is not specified either, C<$_> is used.
1238 In list context, C<extract_delimited> returns a array of three
1239 elements, the extracted substring (I<including the surrounding
1240 delimiters>), the remainder of the text, and the skipped prefix (if
1241 any). If a suitable delimited substring is not found, the first
1242 element of the array is the empty string, the second is the complete
1243 original text, and the prefix returned in the third element is an
1246 In a scalar context, just the extracted substring is returned. In
1247 a void context, the extracted substring (and any prefix) are simply
1248 removed from the beginning of the first argument.
1252 # Remove a single-quoted substring from the very beginning of $text:
1254 $substring = extract_delimited($text, "'", '');
1256 # Remove a single-quoted Pascalish substring (i.e. one in which
1257 # doubling the quote character escapes it) from the very
1258 # beginning of $text:
1260 $substring = extract_delimited($text, "'", '', "'");
1262 # Extract a single- or double- quoted substring from the
1263 # beginning of $text, optionally after some whitespace
1264 # (note the list context to protect $text from modification):
1266 ($substring) = extract_delimited $text, q{"'};
1269 # Delete the substring delimited by the first '/' in $text:
1271 $text = join '', (extract_delimited($text,'/','[^/]*')[2,1];
1273 Note that this last example is I<not> the same as deleting the first
1274 quote-like pattern. For instance, if C<$text> contained the string:
1276 "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }"
1278 then after the deletion it would contain:
1280 "if ('.$UNIXCMD/s) { $cmd = $1; }"
1284 "if ('./cmd' =~ ms) { $cmd = $1; }"
1287 See L<"extract_quotelike"> for a (partial) solution to this problem.
1290 =head2 C<extract_bracketed>
1292 Like C<"extract_delimited">, the C<extract_bracketed> function takes
1293 up to three optional scalar arguments: a string to extract from, a delimiter
1294 specifier, and a prefix pattern. As before, a missing prefix defaults to
1295 optional whitespace and a missing text defaults to C<$_>. However, a missing
1296 delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below).
1298 C<extract_bracketed> extracts a balanced-bracket-delimited
1299 substring (using any one (or more) of the user-specified delimiter
1300 brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also
1301 respect quoted unbalanced brackets (see below).
1303 A "delimiter bracket" is a bracket in list of delimiters passed as
1304 C<extract_bracketed>'s second argument. Delimiter brackets are
1305 specified by giving either the left or right (or both!) versions
1306 of the required bracket(s). Note that the order in which
1307 two or more delimiter brackets are specified is not significant.
1309 A "balanced-bracket-delimited substring" is a substring bounded by
1310 matched brackets, such that any other (left or right) delimiter
1311 bracket I<within> the substring is also matched by an opposite
1312 (right or left) delimiter bracket I<at the same level of nesting>. Any
1313 type of bracket not in the delimiter list is treated as an ordinary
1316 In other words, each type of bracket specified as a delimiter must be
1317 balanced and correctly nested within the substring, and any other kind of
1318 ("non-delimiter") bracket in the substring is ignored.
1320 For example, given the string:
1322 $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }";
1324 then a call to C<extract_bracketed> in a list context:
1326 @result = extract_bracketed( $text, '{}' );
1330 ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" )
1332 since both sets of C<'{..}'> brackets are properly nested and evenly balanced.
1333 (In a scalar context just the first element of the array would be returned. In
1334 a void context, C<$text> would be replaced by an empty string.)
1336 Likewise the call in:
1338 @result = extract_bracketed( $text, '{[' );
1340 would return the same result, since all sets of both types of specified
1341 delimiter brackets are correctly nested and balanced.
1343 However, the call in:
1345 @result = extract_bracketed( $text, '{([<' );
1347 would fail, returning:
1349 ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" );
1351 because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and
1352 the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would
1353 return an empty string. In a void context, C<$text> would be unchanged.)
1355 Note that the embedded single-quotes in the string don't help in this
1356 case, since they have not been specified as acceptable delimiters and are
1357 therefore treated as non-delimiter characters (and ignored).
1359 However, if a particular species of quote character is included in the
1360 delimiter specification, then that type of quote will be correctly handled.
1361 for example, if C<$text> is:
1363 $text = '<A HREF=">>>>">link</A>';
1367 @result = extract_bracketed( $text, '<">' );
1371 ( '<A HREF=">>>>">', 'link</A>', "" )
1373 as expected. Without the specification of C<"> as an embedded quoter:
1375 @result = extract_bracketed( $text, '<>' );
1377 the result would be:
1379 ( '<A HREF=">', '>>>">link</A>', "" )
1381 In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like
1382 quoting (i.e. q{string}, qq{string}, etc) can be specified by including the
1383 letter 'q' as a delimiter. Hence:
1385 @result = extract_bracketed( $text, '<q>' );
1387 would correctly match something like this:
1389 $text = '<leftop: conj /and/ conj>';
1391 See also: C<"extract_quotelike"> and C<"extract_codeblock">.
1394 =head2 C<extract_variable>
1396 C<extract_variable> extracts any valid Perl variable or
1397 variable-involved expression, including scalars, arrays, hashes, array
1398 accesses, hash look-ups, method calls through objects, subroutine calls
1399 through subroutine references, etc.
1401 The subroutine takes up to two optional arguments:
1407 A string to be processed (C<$_> if the string is omitted or C<undef>)
1411 A string specifying a pattern to be matched as a prefix (which is to be
1412 skipped). If omitted, optional whitespace is skipped.
1416 On success in a list context, an array of 3 elements is returned. The
1423 the extracted variable, or variablish expression
1427 the remainder of the input text,
1431 the prefix substring (if any),
1435 On failure, all of these values (except the remaining text) are C<undef>.
1437 In a scalar context, C<extract_variable> returns just the complete
1438 substring that matched a variablish expression. C<undef> is returned on
1439 failure. In addition, the original input text has the returned substring
1440 (and any prefix) removed from it.
1442 In a void context, the input text just has the matched substring (and
1443 any specified prefix) removed.
1446 =head2 C<extract_tagged>
1448 C<extract_tagged> extracts and segments text between (balanced)
1451 The subroutine takes up to five optional arguments:
1457 A string to be processed (C<$_> if the string is omitted or C<undef>)
1461 A string specifying a pattern to be matched as the opening tag.
1462 If the pattern string is omitted (or C<undef>) then a pattern
1463 that matches any standard XML tag is used.
1467 A string specifying a pattern to be matched at the closing tag.
1468 If the pattern string is omitted (or C<undef>) then the closing
1469 tag is constructed by inserting a C</> after any leading bracket
1470 characters in the actual opening tag that was matched (I<not> the pattern
1471 that matched the tag). For example, if the opening tag pattern
1472 is specified as C<'{{\w+}}'> and actually matched the opening tag
1473 C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">.
1477 A string specifying a pattern to be matched as a prefix (which is to be
1478 skipped). If omitted, optional whitespace is skipped.
1482 A hash reference containing various parsing options (see below)
1486 The various options that can be specified are:
1490 =item C<reject =E<gt> $listref>
1492 The list reference contains one or more strings specifying patterns
1493 that must I<not> appear within the tagged text.
1495 For example, to extract
1496 an HTML link (which should not contain nested links) use:
1498 extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} );
1500 =item C<ignore =E<gt> $listref>
1502 The list reference contains one or more strings specifying patterns
1503 that are I<not> be be treated as nested tags within the tagged text
1504 (even if they would match the start tag pattern).
1506 For example, to extract an arbitrary XML tag, but ignore "empty" elements:
1508 extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} );
1510 (also see L<"gen_delimited_pat"> below).
1513 =item C<fail =E<gt> $str>
1515 The C<fail> option indicates the action to be taken if a matching end
1516 tag is not encountered (i.e. before the end of the string or some
1517 C<reject> pattern matches). By default, a failure to match a closing
1518 tag causes C<extract_tagged> to immediately fail.
1520 However, if the string value associated with <reject> is "MAX", then
1521 C<extract_tagged> returns the complete text up to the point of failure.
1522 If the string is "PARA", C<extract_tagged> returns only the first paragraph
1523 after the tag (up to the first line that is either empty or contains
1524 only whitespace characters).
1525 If the string is "", the the default behaviour (i.e. failure) is reinstated.
1527 For example, suppose the start tag "/para" introduces a paragraph, which then
1528 continues until the next "/endpara" tag or until another "/para" tag is
1531 $text = "/para line 1\n\nline 3\n/para line 4";
1533 extract_tagged($text, '/para', '/endpara', undef,
1534 {reject => '/para', fail => MAX );
1536 # EXTRACTED: "/para line 1\n\nline 3\n"
1538 Suppose instead, that if no matching "/endpara" tag is found, the "/para"
1539 tag refers only to the immediately following paragraph:
1541 $text = "/para line 1\n\nline 3\n/para line 4";
1543 extract_tagged($text, '/para', '/endpara', undef,
1544 {reject => '/para', fail => MAX );
1546 # EXTRACTED: "/para line 1\n"
1548 Note that the specified C<fail> behaviour applies to nested tags as well.
1552 On success in a list context, an array of 6 elements is returned. The elements are:
1558 the extracted tagged substring (including the outermost tags),
1562 the remainder of the input text,
1566 the prefix substring (if any),
1574 the text between the opening and closing tags
1578 the closing tag (or "" if no closing tag was found)
1582 On failure, all of these values (except the remaining text) are C<undef>.
1584 In a scalar context, C<extract_tagged> returns just the complete
1585 substring that matched a tagged text (including the start and end
1586 tags). C<undef> is returned on failure. In addition, the original input
1587 text has the returned substring (and any prefix) removed from it.
1589 In a void context, the input text just has the matched substring (and
1590 any specified prefix) removed.
1593 =head2 C<gen_extract_tagged>
1595 (Note: This subroutine is only available under Perl5.005)
1597 C<gen_extract_tagged> generates a new anonymous subroutine which
1598 extracts text between (balanced) specified tags. In other words,
1599 it generates a function identical in function to C<extract_tagged>.
1601 The difference between C<extract_tagged> and the anonymous
1602 subroutines generated by
1603 C<gen_extract_tagged>, is that those generated subroutines:
1609 do not have to reparse tag specification or parsing options every time
1610 they are called (whereas C<extract_tagged> has to effectively rebuild
1611 its tag parser on every call);
1615 make use of the new qr// construct to pre-compile the regexes they use
1616 (whereas C<extract_tagged> uses standard string variable interpolation
1617 to create tag-matching patterns).
1621 The subroutine takes up to four optional arguments (the same set as
1622 C<extract_tagged> except for the string to be processed). It returns
1623 a reference to a subroutine which in turn takes a single argument (the text to
1626 In other words, the implementation of C<extract_tagged> is exactly
1632 $extractor = gen_extract_tagged(@_);
1633 return $extractor->($text);
1636 (although C<extract_tagged> is not currently implemented that way, in order
1637 to preserve pre-5.005 compatibility).
1639 Using C<gen_extract_tagged> to create extraction functions for specific tags
1640 is a good idea if those functions are going to be called more than once, since
1641 their performance is typically twice as good as the more general-purpose
1645 =head2 C<extract_quotelike>
1647 C<extract_quotelike> attempts to recognize, extract, and segment any
1648 one of the various Perl quotes and quotelike operators (see
1649 L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket
1650 delimiters (for the quotelike operators), and trailing modifiers are
1651 all caught. For example, in:
1653 extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #'
1655 extract_quotelike ' "You said, \"Use sed\"." '
1657 extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; '
1659 extract_quotelike ' tr/\\\/\\\\/\\\//ds; '
1661 the full Perl quotelike operations are all extracted correctly.
1663 Note too that, when using the /x modifier on a regex, any comment
1664 containing the current pattern delimiter will cause the regex to be
1665 immediately terminated. In other words:
1668 (?i) # CASE INSENSITIVE
1669 [a-z_] # LEADING ALPHABETIC/UNDERSCORE
1670 [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS
1673 will be extracted as if it were:
1676 (?i) # CASE INSENSITIVE
1677 [a-z_] # LEADING ALPHABETIC/'
1679 This behaviour is identical to that of the actual compiler.
1681 C<extract_quotelike> takes two arguments: the text to be processed and
1682 a prefix to be matched at the very beginning of the text. If no prefix
1683 is specified, optional whitespace is the default. If no text is given,
1686 In a list context, an array of 11 elements is returned. The elements are:
1692 the extracted quotelike substring (including trailing modifiers),
1696 the remainder of the input text,
1700 the prefix substring (if any),
1704 the name of the quotelike operator (if any),
1708 the left delimiter of the first block of the operation,
1712 the text of the first block of the operation
1713 (that is, the contents of
1714 a quote, the regex of a match or substitution or the target list of a
1719 the right delimiter of the first block of the operation,
1723 the left delimiter of the second block of the operation
1724 (that is, if it is a C<s>, C<tr>, or C<y>),
1728 the text of the second block of the operation
1729 (that is, the replacement of a substitution or the translation list
1734 the right delimiter of the second block of the operation (if any),
1738 the trailing modifiers on the operation (if any).
1742 For each of the fields marked "(if any)" the default value on success is
1744 On failure, all of these values (except the remaining text) are C<undef>.
1747 In a scalar context, C<extract_quotelike> returns just the complete substring
1748 that matched a quotelike operation (or C<undef> on failure). In a scalar or
1749 void context, the input text has the same substring (and any specified
1754 # Remove the first quotelike literal that appears in text
1756 $quotelike = extract_quotelike($text,'.*?');
1758 # Replace one or more leading whitespace-separated quotelike
1759 # literals in $_ with "<QLL>"
1761 do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@;
1764 # Isolate the search pattern in a quotelike operation from $text
1766 ($op,$pat) = (extract_quotelike $text)[3,5];
1769 print "search pattern: $pat\n";
1773 print "$op is not a pattern matching operation\n";
1777 =head2 C<extract_quotelike> and "here documents"
1779 C<extract_quotelike> can successfully extract "here documents" from an input
1780 string, but with an important caveat in list contexts.
1782 Unlike other types of quote-like literals, a here document is rarely
1783 a contiguous substring. For example, a typical piece of code using
1784 here document might look like this:
1787 This is the message.
1791 Given this as an input string in a scalar context, C<extract_quotelike>
1792 would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG",
1793 leaving the string " || die;\nexit;" in the original variable. In other words,
1794 the two separate pieces of the here document are successfully extracted and
1797 In a list context, C<extract_quotelike> would return the list
1803 "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document,
1804 including fore and aft delimiters),
1808 " || die;\nexit;" (i.e. the remainder of the input text, concatenated),
1812 "" (i.e. the prefix substring -- trivial in this case),
1816 "<<" (i.e. the "name" of the quotelike operator)
1820 "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes),
1824 "This is the message.\n" (i.e. the text of the here document),
1828 "EOMSG" (i.e. the right delimiter of the here document),
1832 "" (a here document has no second left delimiter, second text, second right
1833 delimiter, or trailing modifiers).
1837 However, the matching position of the input variable would be set to
1838 "exit;" (i.e. I<after> the closing delimiter of the here document),
1839 which would cause the earlier " || die;\nexit;" to be skipped in any
1840 sequence of code fragment extractions.
1842 To avoid this problem, when it encounters a here document whilst
1843 extracting from a modifiable string, C<extract_quotelike> silently
1844 rearranges the string to an equivalent piece of Perl:
1847 This is the message.
1852 in which the here document I<is> contiguous. It still leaves the
1853 matching position after the here document, but now the rest of the line
1854 on which the here document starts is not skipped.
1856 To prevent <extract_quotelike> from mucking about with the input in this way
1857 (this is the only case where a list-context C<extract_quotelike> does so),
1858 you can pass the input variable as an interpolated literal:
1860 $quotelike = extract_quotelike("$var");
1863 =head2 C<extract_codeblock>
1865 C<extract_codeblock> attempts to recognize and extract a balanced
1866 bracket delimited substring that may contain unbalanced brackets
1867 inside Perl quotes or quotelike operations. That is, C<extract_codeblock>
1868 is like a combination of C<"extract_bracketed"> and
1869 C<"extract_quotelike">.
1871 C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>:
1872 a text to process, a set of delimiter brackets to look for, and a prefix to
1873 match first. It also takes an optional fourth parameter, which allows the
1874 outermost delimiter brackets to be specified separately (see below).
1876 Omitting the first argument (input text) means process C<$_> instead.
1877 Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used.
1878 Omitting the third argument (prefix argument) implies optional whitespace at the start.
1879 Omitting the fourth argument (outermost delimiter brackets) indicates that the
1880 value of the second argument is to be used for the outermost delimiters.
1882 Once the prefix an dthe outermost opening delimiter bracket have been
1883 recognized, code blocks are extracted by stepping through the input text and
1884 trying the following alternatives in sequence:
1890 Try and match a closing delimiter bracket. If the bracket was the same
1891 species as the last opening bracket, return the substring to that
1892 point. If the bracket was mismatched, return an error.
1896 Try to match a quote or quotelike operator. If found, call
1897 C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return
1898 the error it returned. Otherwise go back to step 1.
1902 Try to match an opening delimiter bracket. If found, call
1903 C<extract_codeblock> recursively to eat the embedded block. If the
1904 recursive call fails, return an error. Otherwise, go back to step 1.
1908 Unconditionally match a bareword or any other single character, and
1909 then go back to step 1.
1916 # Find a while loop in the text
1918 if ($text =~ s/.*?while\s*\{/{/)
1920 $loop = "while " . extract_codeblock($text);
1923 # Remove the first round-bracketed list (which may include
1924 # round- or curly-bracketed code blocks or quotelike operators)
1926 extract_codeblock $text, "(){}", '[^(]*';
1929 The ability to specify a different outermost delimiter bracket is useful
1930 in some circumstances. For example, in the Parse::RecDescent module,
1931 parser actions which are to be performed only on a successful parse
1932 are specified using a C<E<lt>defer:...E<gt>> directive. For example:
1934 sentence: subject verb object
1935 <defer: {$::theVerb = $item{verb}} >
1937 Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code
1938 within the C<E<lt>defer:...E<gt>> directive, but there's a problem.
1940 A deferred action like this:
1942 <defer: {if ($count>10) {$count--}} >
1944 will be incorrectly parsed as:
1946 <defer: {if ($count>
1948 because the "less than" operator is interpreted as a closing delimiter.
1950 But, by extracting the directive using
1951 S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>>
1952 the '>' character is only treated as a delimited at the outermost
1953 level of the code block, so the directive is parsed correctly.
1955 =head2 C<extract_multiple>
1957 The C<extract_multiple> subroutine takes a string to be processed and a
1958 list of extractors (subroutines or regular expressions) to apply to that string.
1960 In an array context C<extract_multiple> returns an array of substrings
1961 of the original string, as extracted by the specified extractors.
1962 In a scalar context, C<extract_multiple> returns the first
1963 substring successfully extracted from the original string. In both
1964 scalar and void contexts the original string has the first successfully
1965 extracted substring removed from it. In all contexts
1966 C<extract_multiple> starts at the current C<pos> of the string, and
1967 sets that C<pos> appropriately after it matches.
1969 Hence, the aim of of a call to C<extract_multiple> in a list context
1970 is to split the processed string into as many non-overlapping fields as
1971 possible, by repeatedly applying each of the specified extractors
1972 to the remainder of the string. Thus C<extract_multiple> is
1973 a generalized form of Perl's C<split> subroutine.
1975 The subroutine takes up to four optional arguments:
1981 A string to be processed (C<$_> if the string is omitted or C<undef>)
1985 A reference to a list of subroutine references and/or qr// objects and/or
1986 literal strings and/or hash references, specifying the extractors
1987 to be used to split the string. If this argument is omitted (or
1991 sub { extract_variable($_[0], '') },
1992 sub { extract_quotelike($_[0],'') },
1993 sub { extract_codeblock($_[0],'{}','') },
2001 An number specifying the maximum number of fields to return. If this
2002 argument is omitted (or C<undef>), split continues as long as possible.
2004 If the third argument is I<N>, then extraction continues until I<N> fields
2005 have been successfully extracted, or until the string has been completely
2008 Note that in scalar and void contexts the value of this argument is
2009 automatically reset to 1 (under C<-w>, a warning is issued if the argument
2014 A value indicating whether unmatched substrings (see below) within the
2015 text should be skipped or returned as fields. If the value is true,
2016 such substrings are skipped. Otherwise, they are returned.
2020 The extraction process works by applying each extractor in
2021 sequence to the text string.
2023 If the extractor is a subroutine it is called in a list context and is
2024 expected to return a list of a single element, namely the extracted
2025 text. It may optionally also return two further arguments: a string
2026 representing the text left after extraction (like $' for a pattern
2027 match), and a string representing any prefix skipped before the
2028 extraction (like $` in a pattern match). Note that this is designed
2029 to facilitate the use of other Text::Balanced subroutines with
2030 C<extract_multiple>. Note too that the value returned by an extractor
2031 subroutine need not bear any relationship to the corresponding substring
2032 of the original text (see examples below).
2034 If the extractor is a precompiled regular expression or a string,
2035 it is matched against the text in a scalar context with a leading
2036 '\G' and the gc modifiers enabled. The extracted value is either
2037 $1 if that variable is defined after the match, or else the
2038 complete match (i.e. $&).
2040 If the extractor is a hash reference, it must contain exactly one element.
2041 The value of that element is one of the
2042 above extractor types (subroutine reference, regular expression, or string).
2043 The key of that element is the name of a class into which the successful
2044 return value of the extractor will be blessed.
2046 If an extractor returns a defined value, that value is immediately
2047 treated as the next extracted field and pushed onto the list of fields.
2048 If the extractor was specified in a hash reference, the field is also
2049 blessed into the appropriate class,
2051 If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is
2052 assumed to have failed to extract.
2053 If none of the extractor subroutines succeeds, then one
2054 character is extracted from the start of the text and the extraction
2055 subroutines reapplied. Characters which are thus removed are accumulated and
2056 eventually become the next field (unless the fourth argument is true, in which
2057 case they are discarded).
2059 For example, the following extracts substrings that are valid Perl variables:
2061 @fields = extract_multiple($text,
2062 [ sub { extract_variable($_[0]) } ],
2065 This example separates a text into fields which are quote delimited,
2066 curly bracketed, and anything else. The delimited and bracketed
2067 parts are also blessed to identify them (the "anything else" is unblessed):
2069 @fields = extract_multiple($text,
2071 { Delim => sub { extract_delimited($_[0],q{'"}) } },
2072 { Brack => sub { extract_bracketed($_[0],'{}') } },
2075 This call extracts the next single substring that is a valid Perl quotelike
2076 operator (and removes it from $text):
2078 $quotelike = extract_multiple($text,
2080 sub { extract_quotelike($_[0]) },
2083 Finally, here is yet another way to do comma-separated value parsing:
2085 @fields = extract_multiple($csv_text,
2087 sub { extract_delimited($_[0],q{'"}) },
2092 The list in the second argument means:
2093 I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">.
2094 The undef third argument means:
2095 I<"...as many times as possible...">,
2096 and the true value in the fourth argument means
2097 I<"...discarding anything else that appears (i.e. the commas)">.
2099 If you wanted the commas preserved as separate fields (i.e. like split
2100 does if your split pattern has capturing parentheses), you would
2101 just make the last parameter undefined (or remove it).
2104 =head2 C<gen_delimited_pat>
2106 The C<gen_delimited_pat> subroutine takes a single (string) argument and
2107 > builds a Friedl-style optimized regex that matches a string delimited
2108 by any one of the characters in the single argument. For example:
2110 gen_delimited_pat(q{'"})
2114 (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\')
2116 Note that the specified delimiters are automatically quotemeta'd.
2118 A typical use of C<gen_delimited_pat> would be to build special purpose tags
2119 for C<extract_tagged>. For example, to properly ignore "empty" XML elements
2120 (which might contain quoted strings):
2122 my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>';
2124 extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} );
2127 C<gen_delimited_pat> may also be called with an optional second argument,
2128 which specifies the "escape" character(s) to be used for each delimiter.
2129 For example to match a Pascal-style string (where ' is the delimiter
2130 and '' is a literal ' within the string):
2132 gen_delimited_pat(q{'},q{'});
2134 Different escape characters can be specified for different delimiters.
2135 For example, to specify that '/' is the escape for single quotes
2136 and '%' is the escape for double quotes:
2138 gen_delimited_pat(q{'"},q{/%});
2140 If more delimiters than escape chars are specified, the last escape char
2141 is used for the remaining delimiters.
2142 If no escape char is specified for a given specified delimiter, '\' is used.
2144 =head2 C<delimited_pat>
2146 Note that C<gen_delimited_pat> was previously called C<delimited_pat>.
2147 That name may still be used, but is now deprecated.
2152 In a list context, all the functions return C<(undef,$original_text)>
2153 on failure. In a scalar context, failure is indicated by returning C<undef>
2154 (in this case the input text is not modified in any way).
2156 In addition, on failure in I<any> context, the C<$@> variable is set.
2157 Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed
2159 Accessing C<$@-E<gt>{pos}> returns the offset into the original string at
2160 which the error was detected (although not necessarily where it occurred!)
2161 Printing C<$@> directly produces the error message, with the offset appended.
2162 On success, the C<$@> variable is guaranteed to be C<undef>.
2164 The available diagnostics are:
2168 =item C<Did not find a suitable bracket: "%s">
2170 The delimiter provided to C<extract_bracketed> was not one of
2171 C<'()[]E<lt>E<gt>{}'>.
2173 =item C<Did not find prefix: /%s/>
2175 A non-optional prefix was specified but wasn't found at the start of the text.
2177 =item C<Did not find opening bracket after prefix: "%s">
2179 C<extract_bracketed> or C<extract_codeblock> was expecting a
2180 particular kind of bracket at the start of the text, and didn't find it.
2182 =item C<No quotelike operator found after prefix: "%s">
2184 C<extract_quotelike> didn't find one of the quotelike operators C<q>,
2185 C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring
2188 =item C<Unmatched closing bracket: "%c">
2190 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered
2191 a closing bracket where none was expected.
2193 =item C<Unmatched opening bracket(s): "%s">
2195 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran
2196 out of characters in the text before closing one or more levels of nested
2199 =item C<Unmatched embedded quote (%s)>
2201 C<extract_bracketed> attempted to match an embedded quoted substring, but
2202 failed to find a closing quote to match it.
2204 =item C<Did not find closing delimiter to match '%s'>
2206 C<extract_quotelike> was unable to find a closing delimiter to match the
2207 one that opened the quote-like operation.
2209 =item C<Mismatched closing bracket: expected "%c" but found "%s">
2211 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found
2212 a valid bracket delimiter, but it was the wrong species. This usually
2213 indicates a nesting error, but may indicate incorrect quoting or escaping.
2215 =item C<No block delimiter found after quotelike "%s">
2217 C<extract_quotelike> or C<extract_codeblock> found one of the
2218 quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y>
2219 without a suitable block after it.
2221 =item C<Did not find leading dereferencer>
2223 C<extract_variable> was expecting one of '$', '@', or '%' at the start of
2224 a variable, but didn't find any of them.
2226 =item C<Bad identifier after dereferencer>
2228 C<extract_variable> found a '$', '@', or '%' indicating a variable, but that
2229 character was not followed by a legal Perl identifier.
2231 =item C<Did not find expected opening bracket at %s>
2233 C<extract_codeblock> failed to find any of the outermost opening brackets
2234 that were specified.
2236 =item C<Improperly nested codeblock at %s>
2238 A nested code block was found that started with a delimiter that was specified
2239 as being only to be used as an outermost bracket.
2241 =item C<Missing second block for quotelike "%s">
2243 C<extract_codeblock> or C<extract_quotelike> found one of the
2244 quotelike operators C<s>, C<tr> or C<y> followed by only one block.
2246 =item C<No match found for opening bracket>
2248 C<extract_codeblock> failed to find a closing bracket to match the outermost
2251 =item C<Did not find opening tag: /%s/>
2253 C<extract_tagged> did not find a suitable opening tag (after any specified
2254 prefix was removed).
2256 =item C<Unable to construct closing tag to match: /%s/>
2258 C<extract_tagged> matched the specified opening tag and tried to
2259 modify the matched text to produce a matching closing tag (because
2260 none was specified). It failed to generate the closing tag, almost
2261 certainly because the opening tag did not start with a
2262 bracket of some kind.
2264 =item C<Found invalid nested tag: %s>
2266 C<extract_tagged> found a nested tag that appeared in the "reject" list
2267 (and the failure mode was not "MAX" or "PARA").
2269 =item C<Found unbalanced nested tag: %s>
2271 C<extract_tagged> found a nested opening tag that was not matched by a
2272 corresponding nested closing tag (and the failure mode was not "MAX" or "PARA").
2274 =item C<Did not find closing tag>
2276 C<extract_tagged> reached the end of the text without finding a closing tag
2277 to match the original opening tag (and the failure mode was not
2288 Damian Conway (damian@conway.org)
2291 =head1 BUGS AND IRRITATIONS
2293 There are undoubtedly serious bugs lurking somewhere in this code, if
2294 only because parts of it give the impression of understanding a great deal
2295 more about Perl than they really do.
2297 Bug reports and other feedback are most welcome.
2302 Copyright (c) 1997-2001, Damian Conway. All Rights Reserved.
2303 This module is free software. It may be used, redistributed
2304 and/or modified under the same terms as Perl itself.