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 };
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($$$$$$$);
38 sub _match_quotelike($$$$);
40 # HANDLE RETURN VALUES IN VARIOUS CONTEXTS
43 my ($message, $pos) = @_;
44 $@ = bless { error=>$message, pos=>$pos }, "Text::Balanced::ErrorMsg";
49 my ($wantarray, $textref, $message, $pos) = @_;
50 _failmsg $message, $pos if $message;
51 return ("",$$textref,"") if $wantarray;
58 my ($wantarray,$textref) = splice @_, 0, 2;
59 my ($extrapos, $extralen) = @_>18 ? splice(@_, -2, 2) : (0,0);
60 my ($startlen) = $_[5];
61 my $remainderpos = $_[2];
65 while (my ($from, $len) = splice @_, 0, 2)
67 push @res, substr($$textref,$from,$len);
69 if ($extralen) { # CORRECT FILLET
70 my $extra = substr($res[0], $extrapos-$startlen, $extralen, "\n");
71 $res[1] = "$extra$res[1]";
72 eval { substr($$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($$textref);
332 unless ($rdelspec =~ s/\A([[(<{]+)($XMLNAME).*/ quotemeta "$1\/$2". revbracket($1) /oes)
334 _failmsg "Unable to construct closing tag to match: $rdel",
341 $rdelspec = eval "qq{$rdel}" || do {
343 for (qw,~ ! ^ & * ) _ + - = } ] : " ; ' > . ? / | ',)
344 { next if $rdel =~ /\Q$_/; $del = $_; last }
347 croak "Can't interpolate right delimiter $rdel"
349 eval "qq$del$rdel$del";
353 while (pos($$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 ($$textref =~ m{.*?\n(?=$label\n)}gc) {
753 _failmsg qq{Missing here doc terminator ('$label') after "} .
754 substr($$textref, $startpos, 20) .
757 pos $$textref = $startpos;
760 $rd1pos = pos($$textref);
761 $$textref =~ m{$label\n}gc;
762 $ld2pos = pos($$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($$textref) = $ld1pos; # HAVE TO DO THIS BECAUSE LOOKAHEAD BROKEN
790 my ($ldel1, $rdel1) = ("\Q$1","\Q$1");
791 if ($ldel1 =~ /[[(<{]/)
793 $rdel1 =~ tr/[({</])}>/;
794 _match_bracketed($textref,"",$ldel1,"","",$rdel1)
795 || do { pos $$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 _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->($$textref);
923 # print "[$field|$rem]" if $field;
925 elsif (ref($func) eq 'Text::Balanced::Extractor')
926 { @bits = $field = $func->extract($$textref) }
927 elsif( $$textref =~ m/\G$func/gc )
928 { @bits = $field = defined($1) ? $1 : $& }
930 if (defined($field) && length($field))
933 $unkpos = pos $$textref
934 if length($pref) && !defined($unkpos);
937 push @fields, substr($$textref, $unkpos, $lastpos-$unkpos).$pref;
938 $firstpos = $unkpos unless defined $firstpos;
940 last FIELD if @fields == $max;
944 ? bless (\$field, $class)
946 $firstpos = $lastpos unless defined $firstpos;
947 $lastpos = pos $$textref;
948 last FIELD if @fields == $max;
952 if ($$textref =~ /\G(.)/gcs)
954 $unkpos = pos($$textref)-1
955 unless $igunk || defined $unkpos;
961 push @fields, substr($$textref, $unkpos);
962 $firstpos = $unkpos unless defined $firstpos;
963 $lastpos = length $$textref;
968 pos $$textref = $lastpos;
969 return @fields if wantarray;
972 eval { substr($$textref,$firstpos,$lastpos-$firstpos)="";
973 pos $$textref = $firstpos };
978 sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options)
982 my $pre = defined $_[2] ? $_[2] : '\s*';
983 my %options = defined $_[3] ? %{$_[3]} : ();
984 my $omode = defined $options{fail} ? $options{fail} : '';
985 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
986 : defined($options{reject}) ? $options{reject}
989 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
990 : defined($options{ignore}) ? $options{ignore}
994 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
997 for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ }
1002 my $textref = defined $_[0] ? \$_[0] : \$_;
1003 my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
1005 return _fail(wantarray, $textref) unless @match;
1006 return _succeed wantarray, $textref,
1007 $match[2], $match[3]+$match[5]+$match[7], # MATCH
1008 @match[8..9,0..1,2..7]; # REM, PRE, BITS
1011 bless $closure, 'Text::Balanced::Extractor';
1014 package Text::Balanced::Extractor;
1016 sub extract($$) # ($self, $text)
1021 package Text::Balanced::ErrorMsg;
1023 use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" };
1031 Text::Balanced - Extract delimited text sequences from strings.
1036 use Text::Balanced qw (
1049 # Extract the initial substring of $text that is delimited by
1050 # two (unescaped) instances of the first character in $delim.
1052 ($extracted, $remainder) = extract_delimited($text,$delim);
1055 # Extract the initial substring of $text that is bracketed
1056 # with a delimiter(s) specified by $delim (where the string
1057 # in $delim contains one or more of '(){}[]<>').
1059 ($extracted, $remainder) = extract_bracketed($text,$delim);
1062 # Extract the initial substring of $text that is bounded by
1065 ($extracted, $remainder) = extract_tagged($text);
1068 # Extract the initial substring of $text that is bounded by
1069 # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags
1071 ($extracted, $remainder) =
1072 extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]});
1075 # Extract the initial substring of $text that represents a
1076 # Perl "quote or quote-like operation"
1078 ($extracted, $remainder) = extract_quotelike($text);
1081 # Extract the initial substring of $text that represents a block
1082 # of Perl code, bracketed by any of character(s) specified by $delim
1083 # (where the string $delim contains one or more of '(){}[]<>').
1085 ($extracted, $remainder) = extract_codeblock($text,$delim);
1088 # Extract the initial substrings of $text that would be extracted by
1089 # one or more sequential applications of the specified functions
1090 # or regular expressions
1092 @extracted = extract_multiple($text,
1093 [ \&extract_bracketed,
1094 \&extract_quotelike,
1095 \&some_other_extractor_sub,
1100 # Create a string representing an optimized pattern (a la Friedl)
1101 # that matches a substring delimited by any of the specified characters
1102 # (in this case: any type of quote or a slash)
1104 $patstring = gen_delimited_pat(q{'"`/});
1107 # Generate a reference to an anonymous sub that is just like extract_tagged
1108 # but pre-compiled and optimized for a specific pair of tags, and consequently
1109 # much faster (i.e. 3 times faster). It uses qr// for better performance on
1110 # repeated calls, so it only works under Perl 5.005 or later.
1112 $extract_head = gen_extract_tagged('<HEAD>','</HEAD>');
1114 ($extracted, $remainder) = $extract_head->($text);
1119 The various C<extract_...> subroutines may be used to
1120 extract a delimited substring, possibly after skipping a
1121 specified prefix string. By default, that prefix is
1122 optional whitespace (C</\s*/>), but you can change it to whatever
1123 you wish (see below).
1125 The substring to be extracted must appear at the
1126 current C<pos> location of the string's variable
1127 (or at index zero, if no C<pos> position is defined).
1128 In other words, the C<extract_...> subroutines I<don't>
1129 extract the first occurance of a substring anywhere
1130 in a string (like an unanchored regex would). Rather,
1131 they extract an occurance of the substring appearing
1132 immediately at the current matching position in the
1133 string (like a C<\G>-anchored regex would).
1137 =head2 General behaviour in list contexts
1139 In a list context, all the subroutines return a list, the first three
1140 elements of which are always:
1146 The extracted string, including the specified delimiters.
1147 If the extraction fails an empty string is returned.
1151 The remainder of the input string (i.e. the characters after the
1152 extracted string). On failure, the entire string is returned.
1156 The skipped prefix (i.e. the characters before the extracted string).
1157 On failure, the empty string is returned.
1161 Note that in a list context, the contents of the original input text (the first
1162 argument) are not modified in any way.
1164 However, if the input text was passed in a variable, that variable's
1165 C<pos> value is updated to point at the first character after the
1166 extracted text. That means that in a list context the various
1167 subroutines can be used much like regular expressions. For example:
1169 while ( $next = (extract_quotelike($text))[0] )
1171 # process next quote-like (in $next)
1175 =head2 General behaviour in scalar and void contexts
1177 In a scalar context, the extracted string is returned, having first been
1178 removed from the input text. Thus, the following code also processes
1179 each quote-like operation, but actually removes them from $text:
1181 while ( $next = extract_quotelike($text) )
1183 # process next quote-like (in $next)
1186 Note that if the input text is a read-only string (i.e. a literal),
1187 no attempt is made to remove the extracted text.
1189 In a void context the behaviour of the extraction subroutines is
1190 exactly the same as in a scalar context, except (of course) that the
1191 extracted substring is not returned.
1193 =head2 A note about prefixes
1195 Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.)
1196 This can bite you if you're expecting a prefix specification like
1197 '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix
1198 pattern will only succeed if the <H1> tag is on the current line, since
1199 . normally doesn't match newlines.
1201 To overcome this limitation, you need to turn on /s matching within
1202 the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)'
1205 =head2 C<extract_delimited>
1207 The C<extract_delimited> function formalizes the common idiom
1208 of extracting a single-character-delimited substring from the start of
1209 a string. For example, to extract a single-quote delimited string, the
1210 following code is typically used:
1212 ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s;
1215 but with C<extract_delimited> it can be simplified to:
1217 ($extracted,$remainder) = extract_delimited($text, "'");
1219 C<extract_delimited> takes up to four scalars (the input text, the
1220 delimiters, a prefix pattern to be skipped, and any escape characters)
1221 and extracts the initial substring of the text that
1222 is appropriately delimited. If the delimiter string has multiple
1223 characters, the first one encountered in the text is taken to delimit
1225 The third argument specifies a prefix pattern that is to be skipped
1226 (but must be present!) before the substring is extracted.
1227 The final argument specifies the escape character to be used for each
1230 All arguments are optional. If the escape characters are not specified,
1231 every delimiter is escaped with a backslash (C<\>).
1232 If the prefix is not specified, the
1233 pattern C<'\s*'> - optional whitespace - is used. If the delimiter set
1234 is also not specified, the set C</["'`]/> is used. If the text to be processed
1235 is not specified either, C<$_> is used.
1237 In list context, C<extract_delimited> returns a array of three
1238 elements, the extracted substring (I<including the surrounding
1239 delimiters>), the remainder of the text, and the skipped prefix (if
1240 any). If a suitable delimited substring is not found, the first
1241 element of the array is the empty string, the second is the complete
1242 original text, and the prefix returned in the third element is an
1245 In a scalar context, just the extracted substring is returned. In
1246 a void context, the extracted substring (and any prefix) are simply
1247 removed from the beginning of the first argument.
1251 # Remove a single-quoted substring from the very beginning of $text:
1253 $substring = extract_delimited($text, "'", '');
1255 # Remove a single-quoted Pascalish substring (i.e. one in which
1256 # doubling the quote character escapes it) from the very
1257 # beginning of $text:
1259 $substring = extract_delimited($text, "'", '', "'");
1261 # Extract a single- or double- quoted substring from the
1262 # beginning of $text, optionally after some whitespace
1263 # (note the list context to protect $text from modification):
1265 ($substring) = extract_delimited $text, q{"'};
1268 # Delete the substring delimited by the first '/' in $text:
1270 $text = join '', (extract_delimited($text,'/','[^/]*')[2,1];
1272 Note that this last example is I<not> the same as deleting the first
1273 quote-like pattern. For instance, if C<$text> contained the string:
1275 "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }"
1277 then after the deletion it would contain:
1279 "if ('.$UNIXCMD/s) { $cmd = $1; }"
1283 "if ('./cmd' =~ ms) { $cmd = $1; }"
1286 See L<"extract_quotelike"> for a (partial) solution to this problem.
1289 =head2 C<extract_bracketed>
1291 Like C<"extract_delimited">, the C<extract_bracketed> function takes
1292 up to three optional scalar arguments: a string to extract from, a delimiter
1293 specifier, and a prefix pattern. As before, a missing prefix defaults to
1294 optional whitespace and a missing text defaults to C<$_>. However, a missing
1295 delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below).
1297 C<extract_bracketed> extracts a balanced-bracket-delimited
1298 substring (using any one (or more) of the user-specified delimiter
1299 brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also
1300 respect quoted unbalanced brackets (see below).
1302 A "delimiter bracket" is a bracket in list of delimiters passed as
1303 C<extract_bracketed>'s second argument. Delimiter brackets are
1304 specified by giving either the left or right (or both!) versions
1305 of the required bracket(s). Note that the order in which
1306 two or more delimiter brackets are specified is not significant.
1308 A "balanced-bracket-delimited substring" is a substring bounded by
1309 matched brackets, such that any other (left or right) delimiter
1310 bracket I<within> the substring is also matched by an opposite
1311 (right or left) delimiter bracket I<at the same level of nesting>. Any
1312 type of bracket not in the delimiter list is treated as an ordinary
1315 In other words, each type of bracket specified as a delimiter must be
1316 balanced and correctly nested within the substring, and any other kind of
1317 ("non-delimiter") bracket in the substring is ignored.
1319 For example, given the string:
1321 $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }";
1323 then a call to C<extract_bracketed> in a list context:
1325 @result = extract_bracketed( $text, '{}' );
1329 ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" )
1331 since both sets of C<'{..}'> brackets are properly nested and evenly balanced.
1332 (In a scalar context just the first element of the array would be returned. In
1333 a void context, C<$text> would be replaced by an empty string.)
1335 Likewise the call in:
1337 @result = extract_bracketed( $text, '{[' );
1339 would return the same result, since all sets of both types of specified
1340 delimiter brackets are correctly nested and balanced.
1342 However, the call in:
1344 @result = extract_bracketed( $text, '{([<' );
1346 would fail, returning:
1348 ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" );
1350 because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and
1351 the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would
1352 return an empty string. In a void context, C<$text> would be unchanged.)
1354 Note that the embedded single-quotes in the string don't help in this
1355 case, since they have not been specified as acceptable delimiters and are
1356 therefore treated as non-delimiter characters (and ignored).
1358 However, if a particular species of quote character is included in the
1359 delimiter specification, then that type of quote will be correctly handled.
1360 for example, if C<$text> is:
1362 $text = '<A HREF=">>>>">link</A>';
1366 @result = extract_bracketed( $text, '<">' );
1370 ( '<A HREF=">>>>">', 'link</A>', "" )
1372 as expected. Without the specification of C<"> as an embedded quoter:
1374 @result = extract_bracketed( $text, '<>' );
1376 the result would be:
1378 ( '<A HREF=">', '>>>">link</A>', "" )
1380 In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like
1381 quoting (i.e. q{string}, qq{string}, etc) can be specified by including the
1382 letter 'q' as a delimiter. Hence:
1384 @result = extract_bracketed( $text, '<q>' );
1386 would correctly match something like this:
1388 $text = '<leftop: conj /and/ conj>';
1390 See also: C<"extract_quotelike"> and C<"extract_codeblock">.
1393 =head2 C<extract_variable>
1395 C<extract_variable> extracts any valid Perl variable or
1396 variable-involved expression, including scalars, arrays, hashes, array
1397 accesses, hash look-ups, method calls through objects, subroutine calles
1398 through subroutine references, etc.
1400 The subroutine takes up to two optional arguments:
1406 A string to be processed (C<$_> if the string is omitted or C<undef>)
1410 A string specifying a pattern to be matched as a prefix (which is to be
1411 skipped). If omitted, optional whitespace is skipped.
1415 On success in a list context, an array of 3 elements is returned. The
1422 the extracted variable, or variablish expression
1426 the remainder of the input text,
1430 the prefix substring (if any),
1434 On failure, all of these values (except the remaining text) are C<undef>.
1436 In a scalar context, C<extract_variable> returns just the complete
1437 substring that matched a variablish expression. C<undef> is returned on
1438 failure. In addition, the original input text has the returned substring
1439 (and any prefix) removed from it.
1441 In a void context, the input text just has the matched substring (and
1442 any specified prefix) removed.
1445 =head2 C<extract_tagged>
1447 C<extract_tagged> extracts and segments text between (balanced)
1450 The subroutine takes up to five optional arguments:
1456 A string to be processed (C<$_> if the string is omitted or C<undef>)
1460 A string specifying a pattern to be matched as the opening tag.
1461 If the pattern string is omitted (or C<undef>) then a pattern
1462 that matches any standard XML tag is used.
1466 A string specifying a pattern to be matched at the closing tag.
1467 If the pattern string is omitted (or C<undef>) then the closing
1468 tag is constructed by inserting a C</> after any leading bracket
1469 characters in the actual opening tag that was matched (I<not> the pattern
1470 that matched the tag). For example, if the opening tag pattern
1471 is specified as C<'{{\w+}}'> and actually matched the opening tag
1472 C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">.
1476 A string specifying a pattern to be matched as a prefix (which is to be
1477 skipped). If omitted, optional whitespace is skipped.
1481 A hash reference containing various parsing options (see below)
1485 The various options that can be specified are:
1489 =item C<reject =E<gt> $listref>
1491 The list reference contains one or more strings specifying patterns
1492 that must I<not> appear within the tagged text.
1494 For example, to extract
1495 an HTML link (which should not contain nested links) use:
1497 extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} );
1499 =item C<ignore =E<gt> $listref>
1501 The list reference contains one or more strings specifying patterns
1502 that are I<not> be be treated as nested tags within the tagged text
1503 (even if they would match the start tag pattern).
1505 For example, to extract an arbitrary XML tag, but ignore "empty" elements:
1507 extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} );
1509 (also see L<"gen_delimited_pat"> below).
1512 =item C<fail =E<gt> $str>
1514 The C<fail> option indicates the action to be taken if a matching end
1515 tag is not encountered (i.e. before the end of the string or some
1516 C<reject> pattern matches). By default, a failure to match a closing
1517 tag causes C<extract_tagged> to immediately fail.
1519 However, if the string value associated with <reject> is "MAX", then
1520 C<extract_tagged> returns the complete text up to the point of failure.
1521 If the string is "PARA", C<extract_tagged> returns only the first paragraph
1522 after the tag (up to the first line that is either empty or contains
1523 only whitespace characters).
1524 If the string is "", the the default behaviour (i.e. failure) is reinstated.
1526 For example, suppose the start tag "/para" introduces a paragraph, which then
1527 continues until the next "/endpara" tag or until another "/para" tag is
1530 $text = "/para line 1\n\nline 3\n/para line 4";
1532 extract_tagged($text, '/para', '/endpara', undef,
1533 {reject => '/para', fail => MAX );
1535 # EXTRACTED: "/para line 1\n\nline 3\n"
1537 Suppose instead, that if no matching "/endpara" tag is found, the "/para"
1538 tag refers only to the immediately following paragraph:
1540 $text = "/para line 1\n\nline 3\n/para line 4";
1542 extract_tagged($text, '/para', '/endpara', undef,
1543 {reject => '/para', fail => MAX );
1545 # EXTRACTED: "/para line 1\n"
1547 Note that the specified C<fail> behaviour applies to nested tags as well.
1551 On success in a list context, an array of 6 elements is returned. The elements are:
1557 the extracted tagged substring (including the outermost tags),
1561 the remainder of the input text,
1565 the prefix substring (if any),
1573 the text between the opening and closing tags
1577 the closing tag (or "" if no closing tag was found)
1581 On failure, all of these values (except the remaining text) are C<undef>.
1583 In a scalar context, C<extract_tagged> returns just the complete
1584 substring that matched a tagged text (including the start and end
1585 tags). C<undef> is returned on failure. In addition, the original input
1586 text has the returned substring (and any prefix) removed from it.
1588 In a void context, the input text just has the matched substring (and
1589 any specified prefix) removed.
1592 =head2 C<gen_extract_tagged>
1594 (Note: This subroutine is only available under Perl5.005)
1596 C<gen_extract_tagged> generates a new anonymous subroutine which
1597 extracts text between (balanced) specified tags. In other words,
1598 it generates a function identical in function to C<extract_tagged>.
1600 The difference between C<extract_tagged> and the anonymous
1601 subroutines generated by
1602 C<gen_extract_tagged>, is that those generated subroutines:
1608 do not have to reparse tag specification or parsing options every time
1609 they are called (whereas C<extract_tagged> has to effectively rebuild
1610 its tag parser on every call);
1614 make use of the new qr// construct to pre-compile the regexes they use
1615 (whereas C<extract_tagged> uses standard string variable interpolation
1616 to create tag-matching patterns).
1620 The subroutine takes up to four optional arguments (the same set as
1621 C<extract_tagged> except for the string to be processed). It returns
1622 a reference to a subroutine which in turn takes a single argument (the text to
1625 In other words, the implementation of C<extract_tagged> is exactly
1631 $extractor = gen_extract_tagged(@_);
1632 return $extractor->($text);
1635 (although C<extract_tagged> is not currently implemented that way, in order
1636 to preserve pre-5.005 compatibility).
1638 Using C<gen_extract_tagged> to create extraction functions for specific tags
1639 is a good idea if those functions are going to be called more than once, since
1640 their performance is typically twice as good as the more general-purpose
1644 =head2 C<extract_quotelike>
1646 C<extract_quotelike> attempts to recognize, extract, and segment any
1647 one of the various Perl quotes and quotelike operators (see
1648 L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket
1649 delimiters (for the quotelike operators), and trailing modifiers are
1650 all caught. For example, in:
1652 extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #'
1654 extract_quotelike ' "You said, \"Use sed\"." '
1656 extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; '
1658 extract_quotelike ' tr/\\\/\\\\/\\\//ds; '
1660 the full Perl quotelike operations are all extracted correctly.
1662 Note too that, when using the /x modifier on a regex, any comment
1663 containing the current pattern delimiter will cause the regex to be
1664 immediately terminated. In other words:
1667 (?i) # CASE INSENSITIVE
1668 [a-z_] # LEADING ALPHABETIC/UNDERSCORE
1669 [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS
1672 will be extracted as if it were:
1675 (?i) # CASE INSENSITIVE
1676 [a-z_] # LEADING ALPHABETIC/'
1678 This behaviour is identical to that of the actual compiler.
1680 C<extract_quotelike> takes two arguments: the text to be processed and
1681 a prefix to be matched at the very beginning of the text. If no prefix
1682 is specified, optional whitespace is the default. If no text is given,
1685 In a list context, an array of 11 elements is returned. The elements are:
1691 the extracted quotelike substring (including trailing modifiers),
1695 the remainder of the input text,
1699 the prefix substring (if any),
1703 the name of the quotelike operator (if any),
1707 the left delimiter of the first block of the operation,
1711 the text of the first block of the operation
1712 (that is, the contents of
1713 a quote, the regex of a match or substitution or the target list of a
1718 the right delimiter of the first block of the operation,
1722 the left delimiter of the second block of the operation
1723 (that is, if it is a C<s>, C<tr>, or C<y>),
1727 the text of the second block of the operation
1728 (that is, the replacement of a substitution or the translation list
1733 the right delimiter of the second block of the operation (if any),
1737 the trailing modifiers on the operation (if any).
1741 For each of the fields marked "(if any)" the default value on success is
1743 On failure, all of these values (except the remaining text) are C<undef>.
1746 In a scalar context, C<extract_quotelike> returns just the complete substring
1747 that matched a quotelike operation (or C<undef> on failure). In a scalar or
1748 void context, the input text has the same substring (and any specified
1753 # Remove the first quotelike literal that appears in text
1755 $quotelike = extract_quotelike($text,'.*?');
1757 # Replace one or more leading whitespace-separated quotelike
1758 # literals in $_ with "<QLL>"
1760 do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@;
1763 # Isolate the search pattern in a quotelike operation from $text
1765 ($op,$pat) = (extract_quotelike $text)[3,5];
1768 print "search pattern: $pat\n";
1772 print "$op is not a pattern matching operation\n";
1776 =head2 C<extract_quotelike> and "here documents"
1778 C<extract_quotelike> can successfully extract "here documents" from an input
1779 string, but with an important caveat in list contexts.
1781 Unlike other types of quote-like literals, a here document is rarely
1782 a contiguous substring. For example, a typical piece of code using
1783 here document might look like this:
1786 This is the message.
1790 Given this as an input string in a scalar context, C<extract_quotelike>
1791 would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG",
1792 leaving the string " || die;\nexit;" in the original variable. In other words,
1793 the two separate pieces of the here document are successfully extracted and
1796 In a list context, C<extract_quotelike> would return the list
1802 "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document,
1803 including fore and aft delimiters),
1807 " || die;\nexit;" (i.e. the remainder of the input text, concatenated),
1811 "" (i.e. the prefix substring -- trivial in this case),
1815 "<<" (i.e. the "name" of the quotelike operator)
1819 "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes),
1823 "This is the message.\n" (i.e. the text of the here document),
1827 "EOMSG" (i.e. the right delimiter of the here document),
1831 "" (a here document has no second left delimiter, second text, second right
1832 delimiter, or trailing modifiers).
1836 However, the matching position of the input variable would be set to
1837 "exit;" (i.e. I<after> the closing delimiter of the here document),
1838 which would cause the earlier " || die;\nexit;" to be skipped in any
1839 sequence of code fragment extractions.
1841 To avoid this problem, when it encounters a here document whilst
1842 extracting from a modifiable string, C<extract_quotelike> silently
1843 rearranges the string to an equivalent piece of Perl:
1846 This is the message.
1851 in which the here document I<is> contiguous. It still leaves the
1852 matching position after the here document, but now the rest of the line
1853 on which the here document starts is not skipped.
1855 To prevent <extract_quotelike> from mucking about with the input in this way
1856 (this is the only case where a list-context C<extract_quotelike> does so),
1857 you can pass the input variable as an interpolated literal:
1859 $quotelike = extract_quotelike("$var");
1862 =head2 C<extract_codeblock>
1864 C<extract_codeblock> attempts to recognize and extract a balanced
1865 bracket delimited substring that may contain unbalanced brackets
1866 inside Perl quotes or quotelike operations. That is, C<extract_codeblock>
1867 is like a combination of C<"extract_bracketed"> and
1868 C<"extract_quotelike">.
1870 C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>:
1871 a text to process, a set of delimiter brackets to look for, and a prefix to
1872 match first. It also takes an optional fourth parameter, which allows the
1873 outermost delimiter brackets to be specified separately (see below).
1875 Omitting the first argument (input text) means process C<$_> instead.
1876 Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used.
1877 Omitting the third argument (prefix argument) implies optional whitespace at the start.
1878 Omitting the fourth argument (outermost delimiter brackets) indicates that the
1879 value of the second argument is to be used for the outermost delimiters.
1881 Once the prefix an dthe outermost opening delimiter bracket have been
1882 recognized, code blocks are extracted by stepping through the input text and
1883 trying the following alternatives in sequence:
1889 Try and match a closing delimiter bracket. If the bracket was the same
1890 species as the last opening bracket, return the substring to that
1891 point. If the bracket was mismatched, return an error.
1895 Try to match a quote or quotelike operator. If found, call
1896 C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return
1897 the error it returned. Otherwise go back to step 1.
1901 Try to match an opening delimiter bracket. If found, call
1902 C<extract_codeblock> recursively to eat the embedded block. If the
1903 recursive call fails, return an error. Otherwise, go back to step 1.
1907 Unconditionally match a bareword or any other single character, and
1908 then go back to step 1.
1915 # Find a while loop in the text
1917 if ($text =~ s/.*?while\s*\{/{/)
1919 $loop = "while " . extract_codeblock($text);
1922 # Remove the first round-bracketed list (which may include
1923 # round- or curly-bracketed code blocks or quotelike operators)
1925 extract_codeblock $text, "(){}", '[^(]*';
1928 The ability to specify a different outermost delimiter bracket is useful
1929 in some circumstances. For example, in the Parse::RecDescent module,
1930 parser actions which are to be performed only on a successful parse
1931 are specified using a C<E<lt>defer:...E<gt>> directive. For example:
1933 sentence: subject verb object
1934 <defer: {$::theVerb = $item{verb}} >
1936 Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code
1937 within the C<E<lt>defer:...E<gt>> directive, but there's a problem.
1939 A deferred action like this:
1941 <defer: {if ($count>10) {$count--}} >
1943 will be incorrectly parsed as:
1945 <defer: {if ($count>
1947 because the "less than" operator is interpreted as a closing delimiter.
1949 But, by extracting the directive using
1950 S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>>
1951 the '>' character is only treated as a delimited at the outermost
1952 level of the code block, so the directive is parsed correctly.
1954 =head2 C<extract_multiple>
1956 The C<extract_multiple> subroutine takes a string to be processed and a
1957 list of extractors (subroutines or regular expressions) to apply to that string.
1959 In an array context C<extract_multiple> returns an array of substrings
1960 of the original string, as extracted by the specified extractors.
1961 In a scalar context, C<extract_multiple> returns the first
1962 substring successfully extracted from the original string. In both
1963 scalar and void contexts the original string has the first successfully
1964 extracted substring removed from it. In all contexts
1965 C<extract_multiple> starts at the current C<pos> of the string, and
1966 sets that C<pos> appropriately after it matches.
1968 Hence, the aim of of a call to C<extract_multiple> in a list context
1969 is to split the processed string into as many non-overlapping fields as
1970 possible, by repeatedly applying each of the specified extractors
1971 to the remainder of the string. Thus C<extract_multiple> is
1972 a generalized form of Perl's C<split> subroutine.
1974 The subroutine takes up to four optional arguments:
1980 A string to be processed (C<$_> if the string is omitted or C<undef>)
1984 A reference to a list of subroutine references and/or qr// objects and/or
1985 literal strings and/or hash references, specifying the extractors
1986 to be used to split the string. If this argument is omitted (or
1990 sub { extract_variable($_[0], '') },
1991 sub { extract_quotelike($_[0],'') },
1992 sub { extract_codeblock($_[0],'{}','') },
2000 An number specifying the maximum number of fields to return. If this
2001 argument is omitted (or C<undef>), split continues as long as possible.
2003 If the third argument is I<N>, then extraction continues until I<N> fields
2004 have been successfully extracted, or until the string has been completely
2007 Note that in scalar and void contexts the value of this argument is
2008 automatically reset to 1 (under C<-w>, a warning is issued if the argument
2013 A value indicating whether unmatched substrings (see below) within the
2014 text should be skipped or returned as fields. If the value is true,
2015 such substrings are skipped. Otherwise, they are returned.
2019 The extraction process works by applying each extractor in
2020 sequence to the text string.
2022 If the extractor is a subroutine it is called in a list context and is
2023 expected to return a list of a single element, namely the extracted
2024 text. It may optionally also return two further arguments: a string
2025 representing the text left after extraction (like $' for a pattern
2026 match), and a string representing any prefix skipped before the
2027 extraction (like $` in a pattern match). Note that this is designed
2028 to facilitate the use of other Text::Balanced subroutines with
2029 C<extract_multiple>. Note too that the value returned by an extractor
2030 subroutine need not bear any relationship to the corresponding substring
2031 of the original text (see examples below).
2033 If the extractor is a precompiled regular expression or a string,
2034 it is matched against the text in a scalar context with a leading
2035 '\G' and the gc modifiers enabled. The extracted value is either
2036 $1 if that variable is defined after the match, or else the
2037 complete match (i.e. $&).
2039 If the extractor is a hash reference, it must contain exactly one element.
2040 The value of that element is one of the
2041 above extractor types (subroutine reference, regular expression, or string).
2042 The key of that element is the name of a class into which the successful
2043 return value of the extractor will be blessed.
2045 If an extractor returns a defined value, that value is immediately
2046 treated as the next extracted field and pushed onto the list of fields.
2047 If the extractor was specified in a hash reference, the field is also
2048 blessed into the appropriate class,
2050 If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is
2051 assumed to have failed to extract.
2052 If none of the extractor subroutines succeeds, then one
2053 character is extracted from the start of the text and the extraction
2054 subroutines reapplied. Characters which are thus removed are accumulated and
2055 eventually become the next field (unless the fourth argument is true, in which
2056 case they are disgarded).
2058 For example, the following extracts substrings that are valid Perl variables:
2060 @fields = extract_multiple($text,
2061 [ sub { extract_variable($_[0]) } ],
2064 This example separates a text into fields which are quote delimited,
2065 curly bracketed, and anything else. The delimited and bracketed
2066 parts are also blessed to identify them (the "anything else" is unblessed):
2068 @fields = extract_multiple($text,
2070 { Delim => sub { extract_delimited($_[0],q{'"}) } },
2071 { Brack => sub { extract_bracketed($_[0],'{}') } },
2074 This call extracts the next single substring that is a valid Perl quotelike
2075 operator (and removes it from $text):
2077 $quotelike = extract_multiple($text,
2079 sub { extract_quotelike($_[0]) },
2082 Finally, here is yet another way to do comma-separated value parsing:
2084 @fields = extract_multiple($csv_text,
2086 sub { extract_delimited($_[0],q{'"}) },
2091 The list in the second argument means:
2092 I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">.
2093 The undef third argument means:
2094 I<"...as many times as possible...">,
2095 and the true value in the fourth argument means
2096 I<"...discarding anything else that appears (i.e. the commas)">.
2098 If you wanted the commas preserved as separate fields (i.e. like split
2099 does if your split pattern has capturing parentheses), you would
2100 just make the last parameter undefined (or remove it).
2103 =head2 C<gen_delimited_pat>
2105 The C<gen_delimited_pat> subroutine takes a single (string) argument and
2106 > builds a Friedl-style optimized regex that matches a string delimited
2107 by any one of the characters in the single argument. For example:
2109 gen_delimited_pat(q{'"})
2113 (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\')
2115 Note that the specified delimiters are automatically quotemeta'd.
2117 A typical use of C<gen_delimited_pat> would be to build special purpose tags
2118 for C<extract_tagged>. For example, to properly ignore "empty" XML elements
2119 (which might contain quoted strings):
2121 my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>';
2123 extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} );
2126 C<gen_delimited_pat> may also be called with an optional second argument,
2127 which specifies the "escape" character(s) to be used for each delimiter.
2128 For example to match a Pascal-style string (where ' is the delimiter
2129 and '' is a literal ' within the string):
2131 gen_delimited_pat(q{'},q{'});
2133 Different escape characters can be specified for different delimiters.
2134 For example, to specify that '/' is the escape for single quotes
2135 and '%' is the escape for double quotes:
2137 gen_delimited_pat(q{'"},q{/%});
2139 If more delimiters than escape chars are specified, the last escape char
2140 is used for the remaining delimiters.
2141 If no escape char is specified for a given specified delimiter, '\' is used.
2144 C<gen_delimited_pat> was previously called
2145 C<delimited_pat>. That name may still be used, but is now deprecated.
2150 In a list context, all the functions return C<(undef,$original_text)>
2151 on failure. In a scalar context, failure is indicated by returning C<undef>
2152 (in this case the input text is not modified in any way).
2154 In addition, on failure in I<any> context, the C<$@> variable is set.
2155 Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed
2157 Accessing C<$@-E<gt>{pos}> returns the offset into the original string at
2158 which the error was detected (although not necessarily where it occurred!)
2159 Printing C<$@> directly produces the error message, with the offset appended.
2160 On success, the C<$@> variable is guaranteed to be C<undef>.
2162 The available diagnostics are:
2166 =item C<Did not find a suitable bracket: "%s">
2168 The delimiter provided to C<extract_bracketed> was not one of
2169 C<'()[]E<lt>E<gt>{}'>.
2171 =item C<Did not find prefix: /%s/>
2173 A non-optional prefix was specified but wasn't found at the start of the text.
2175 =item C<Did not find opening bracket after prefix: "%s">
2177 C<extract_bracketed> or C<extract_codeblock> was expecting a
2178 particular kind of bracket at the start of the text, and didn't find it.
2180 =item C<No quotelike operator found after prefix: "%s">
2182 C<extract_quotelike> didn't find one of the quotelike operators C<q>,
2183 C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring
2186 =item C<Unmatched closing bracket: "%c">
2188 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered
2189 a closing bracket where none was expected.
2191 =item C<Unmatched opening bracket(s): "%s">
2193 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran
2194 out of characters in the text before closing one or more levels of nested
2197 =item C<Unmatched embedded quote (%s)>
2199 C<extract_bracketed> attempted to match an embedded quoted substring, but
2200 failed to find a closing quote to match it.
2202 =item C<Did not find closing delimiter to match '%s'>
2204 C<extract_quotelike> was unable to find a closing delimiter to match the
2205 one that opened the quote-like operation.
2207 =item C<Mismatched closing bracket: expected "%c" but found "%s">
2209 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found
2210 a valid bracket delimiter, but it was the wrong species. This usually
2211 indicates a nesting error, but may indicate incorrect quoting or escaping.
2213 =item C<No block delimiter found after quotelike "%s">
2215 C<extract_quotelike> or C<extract_codeblock> found one of the
2216 quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y>
2217 without a suitable block after it.
2219 =item C<Did not find leading dereferencer>
2221 C<extract_variable> was expecting one of '$', '@', or '%' at the start of
2222 a variable, but didn't find any of them.
2224 =item C<Bad identifier after dereferencer>
2226 C<extract_variable> found a '$', '@', or '%' indicating a variable, but that
2227 character was not followed by a legal Perl identifier.
2229 =item C<Did not find expected opening bracket at %s>
2231 C<extract_codeblock> failed to find any of the outermost opening brackets
2232 that were specified.
2234 =item C<Improperly nested codeblock at %s>
2236 A nested code block was found that started with a delimiter that was specified
2237 as being only to be used as an outermost bracket.
2239 =item C<Missing second block for quotelike "%s">
2241 C<extract_codeblock> or C<extract_quotelike> found one of the
2242 quotelike operators C<s>, C<tr> or C<y> followed by only one block.
2244 =item C<No match found for opening bracket>
2246 C<extract_codeblock> failed to find a closing bracket to match the outermost
2249 =item C<Did not find opening tag: /%s/>
2251 C<extract_tagged> did not find a suitable opening tag (after any specified
2252 prefix was removed).
2254 =item C<Unable to construct closing tag to match: /%s/>
2256 C<extract_tagged> matched the specified opening tag and tried to
2257 modify the matched text to produce a matching closing tag (because
2258 none was specified). It failed to generate the closing tag, almost
2259 certainly because the opening tag did not start with a
2260 bracket of some kind.
2262 =item C<Found invalid nested tag: %s>
2264 C<extract_tagged> found a nested tag that appeared in the "reject" list
2265 (and the failure mode was not "MAX" or "PARA").
2267 =item C<Found unbalanced nested tag: %s>
2269 C<extract_tagged> found a nested opening tag that was not matched by a
2270 corresponding nested closing tag (and the failure mode was not "MAX" or "PARA").
2272 =item C<Did not find closing tag>
2274 C<extract_tagged> reached the end of the text without finding a closing tag
2275 to match the original opening tag (and the failure mode was not
2286 Damian Conway (damian@conway.org)
2289 =head1 BUGS AND IRRITATIONS
2291 There are undoubtedly serious bugs lurking somewhere in this code, if
2292 only because parts of it give the impression of understanding a great deal
2293 more about Perl than they really do.
2295 Bug reports and other feedback are most welcome.
2300 Copyright (c) 1997-2001, Damian Conway. All Rights Reserved.
2301 This module is free software. It may be used, redistributed
2302 and/or modified under the same terms as Perl itself.