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($$textref);
332 unless ($rdelspec =~ s/\A([[(<{]+)($XMLNAME).*/ quotemeta "$1\/$2". revbracket($1) /oes)
334 _failmsg "Unable to construct closing tag to match: $rdel",
341 $rdelspec = eval "qq{$rdel}";
344 while (pos($$textref) < length($$textref))
346 next if $$textref =~ m/\G\\./gc;
348 if ($$textref =~ m/\G(\n[ \t]*\n)/gc )
350 $parapos = pos($$textref) - length($1)
351 unless defined $parapos;
353 elsif ($$textref =~ m/\G($rdelspec)/gc )
355 $closetagpos = pos($$textref)-length($1);
358 elsif ($ignore && $$textref =~ m/\G(?:$ignore)/gc)
362 elsif ($bad && $$textref =~ m/\G($bad)/gcs)
364 pos($$textref) -= length($1); # CUT OFF WHATEVER CAUSED THE SHORTNESS
365 goto short if ($omode eq 'PARA' || $omode eq 'MAX');
366 _failmsg "Found invalid nested tag: $1", pos $$textref;
369 elsif ($$textref =~ m/\G($ldel)/gc)
372 pos($$textref) -= length($tag); # REWIND TO NESTED TAG
373 unless (_match_tagged(@_)) # MATCH NESTED TAG
375 goto short if $omode eq 'PARA' || $omode eq 'MAX';
376 _failmsg "Found unbalanced nested tag: $tag",
381 else { $$textref =~ m/./gcs }
385 $closetagpos = pos($$textref);
386 goto matched if $omode eq 'MAX';
387 goto failed unless $omode eq 'PARA';
389 if (defined $parapos) { pos($$textref) = $parapos }
390 else { $parapos = pos($$textref) }
393 $startpos, $opentagpos-$startpos, # PREFIX
394 $opentagpos, $textpos-$opentagpos, # OPENING TAG
395 $textpos, $parapos-$textpos, # TEXT
396 $parapos, 0, # NO CLOSING TAG
397 $parapos, length($$textref)-$parapos, # REMAINDER
401 $endpos = pos($$textref);
403 $startpos, $opentagpos-$startpos, # PREFIX
404 $opentagpos, $textpos-$opentagpos, # OPENING TAG
405 $textpos, $closetagpos-$textpos, # TEXT
406 $closetagpos, $endpos-$closetagpos, # CLOSING TAG
407 $endpos, length($$textref)-$endpos, # REMAINDER
411 _failmsg "Did not find closing tag", pos $$textref unless $@;
412 pos($$textref) = $startpos;
416 sub extract_variable (;$$)
418 my $textref = defined $_[0] ? \$_[0] : \$_;
419 return ("","","") unless defined $$textref;
420 my $pre = defined $_[1] ? $_[1] : '\s*';
422 my @match = _match_variable($textref,$pre);
424 return _fail wantarray, $textref unless @match;
426 return _succeed wantarray, $textref,
427 @match[2..3,4..5,0..1]; # MATCH, REMAINDER, PREFIX
430 sub _match_variable($$)
432 my ($textref, $pre) = @_;
433 my $startpos = pos($$textref) = pos($$textref)||0;
434 unless ($$textref =~ m/\G($pre)/gc)
436 _failmsg "Did not find prefix: /$pre/", pos $$textref;
439 my $varpos = pos($$textref);
440 unless ($$textref =~ m/\G(\$#?|[*\@\%]|\\&)+/gc)
442 _failmsg "Did not find leading dereferencer", pos $$textref;
443 pos $$textref = $startpos;
447 unless ($$textref =~ m/\G\s*(?:::|')?(?:[_a-z]\w*(?:::|'))*[_a-z]\w*/gci
448 or _match_codeblock($textref, "", '\{', '\}', '\{', '\}', 0))
450 _failmsg "Bad identifier after dereferencer", pos $$textref;
451 pos $$textref = $startpos;
457 next if _match_codeblock($textref,
458 qr/\s*->\s*(?:[_a-zA-Z]\w+\s*)?/,
459 qr/[({[]/, qr/[)}\]]/,
460 qr/[({[]/, qr/[)}\]]/, 0);
461 next if _match_codeblock($textref,
462 qr/\s*/, qr/[{[]/, qr/[}\]]/,
463 qr/[{[]/, qr/[}\]]/, 0);
464 next if _match_variable($textref,'\s*->\s*');
465 next if $$textref =~ m/\G\s*->\s*\w+(?![{([])/gc;
469 my $endpos = pos($$textref);
470 return ($startpos, $varpos-$startpos,
471 $varpos, $endpos-$varpos,
472 $endpos, length($$textref)-$endpos
476 sub extract_codeblock (;$$$$$)
478 my $textref = defined $_[0] ? \$_[0] : \$_;
479 my $wantarray = wantarray;
480 my $ldel_inner = defined $_[1] ? $_[1] : '{';
481 my $pre = defined $_[2] ? $_[2] : '\s*';
482 my $ldel_outer = defined $_[3] ? $_[3] : $ldel_inner;
484 my $rdel_inner = $ldel_inner;
485 my $rdel_outer = $ldel_outer;
487 for ($ldel_inner, $ldel_outer) { tr/[]()<>{}\0-\377/[[((<<{{/ds }
488 for ($rdel_inner, $rdel_outer) { tr/[]()<>{}\0-\377/]]))>>}}/ds }
489 for ($ldel_inner, $ldel_outer, $rdel_inner, $rdel_outer)
491 $_ = '('.join('|',map { quotemeta $_ } split('',$_)).')'
495 my @match = _match_codeblock($textref, $pre,
496 $ldel_outer, $rdel_outer,
497 $ldel_inner, $rdel_inner,
499 return _fail($wantarray, $textref) unless @match;
500 return _succeed($wantarray, $textref,
501 @match[2..3,4..5,0..1] # MATCH, REMAINDER, PREFIX
506 sub _match_codeblock($$$$$$$)
508 my ($textref, $pre, $ldel_outer, $rdel_outer, $ldel_inner, $rdel_inner, $rd) = @_;
509 my $startpos = pos($$textref) = pos($$textref) || 0;
510 unless ($$textref =~ m/\G($pre)/gc)
512 _failmsg qq{Did not match prefix /$pre/ at"} .
513 substr($$textref,pos($$textref),20) .
518 my $codepos = pos($$textref);
519 unless ($$textref =~ m/\G($ldel_outer)/gc) # OUTERMOST DELIMITER
521 _failmsg qq{Did not find expected opening bracket at "} .
522 substr($$textref,pos($$textref),20) .
525 pos $$textref = $startpos;
529 $closing =~ tr/([<{/)]>}/;
532 while (pos($$textref) < length($$textref))
535 if ($rd && $$textref =~ m#\G(\Q(?)\E|\Q(s?)\E|\Q(s)\E)#gc)
541 if ($$textref =~ m/\G\s*#.*/gc)
546 if ($$textref =~ m/\G\s*($rdel_outer)/gc)
548 unless ($matched = ($closing && $1 eq $closing) )
550 next if $1 eq '>'; # MIGHT BE A "LESS THAN"
551 _failmsg q{Mismatched closing bracket at "} .
552 substr($$textref,pos($$textref),20) .
553 qq{...". Expected '$closing'},
559 if (_match_variable($textref,'\s*') ||
560 _match_quotelike($textref,'\s*',$patvalid,$patvalid) )
567 # NEED TO COVER MANY MORE CASES HERE!!!
568 if ($$textref =~ m#\G\s*( [-+*x/%^&|.]=?
571 | (\*\*|&&|\|\||<<|>>)=?
572 | split|grep|map|return
579 if ( _match_codeblock($textref, '\s*', $ldel_inner, $rdel_inner, $ldel_inner, $rdel_inner, $rd) )
585 if ($$textref =~ m/\G\s*$ldel_outer/gc)
587 _failmsg q{Improperly nested codeblock at "} .
588 substr($$textref,pos($$textref),20) .
595 $$textref =~ m/\G\s*(\w+|[-=>]>|.|\Z)/gc;
597 continue { $@ = undef }
601 _failmsg 'No match found for opening bracket', pos $$textref
606 my $endpos = pos($$textref);
607 return ( $startpos, $codepos-$startpos,
608 $codepos, $endpos-$codepos,
609 $endpos, length($$textref)-$endpos,
615 'none' => '[cgimsox]*',
617 's' => '[cegimsox]*',
627 sub extract_quotelike (;$$)
629 my $textref = $_[0] ? \$_[0] : \$_;
630 my $wantarray = wantarray;
631 my $pre = defined $_[1] ? $_[1] : '\s*';
633 my @match = _match_quotelike($textref,$pre,1,0);
634 return _fail($wantarray, $textref) unless @match;
635 return _succeed($wantarray, $textref,
636 $match[2], $match[18]-$match[2], # MATCH
637 @match[18,19], # REMAINDER
638 @match[0,1], # PREFIX
639 @match[2..17], # THE BITS
640 @match[20,21], # ANY FILLET?
644 sub _match_quotelike($$$$) # ($textref, $prepat, $allow_raw_match)
646 my ($textref, $pre, $rawmatch, $qmark) = @_;
648 my ($textlen,$startpos,
650 $preld1pos,$ld1pos,$str1pos,$rd1pos,
651 $preld2pos,$ld2pos,$str2pos,$rd2pos,
652 $modpos) = ( length($$textref), pos($$textref) = pos($$textref) || 0 );
654 unless ($$textref =~ m/\G($pre)/gc)
656 _failmsg qq{Did not find prefix /$pre/ at "} .
657 substr($$textref, pos($$textref), 20) .
662 $oppos = pos($$textref);
664 my $initial = substr($$textref,$oppos,1);
666 if ($initial && $initial =~ m|^[\"\'\`]|
667 || $rawmatch && $initial =~ m|^/|
668 || $qmark && $initial =~ m|^\?|)
670 unless ($$textref =~ m/ \Q$initial\E [^\\$initial]* (\\.[^\\$initial]*)* \Q$initial\E /gcsx)
672 _failmsg qq{Did not find closing delimiter to match '$initial' at "} .
673 substr($$textref, $oppos, 20) .
676 pos $$textref = $startpos;
679 $modpos= pos($$textref);
682 if ($initial eq '/' || $initial eq '?')
684 $$textref =~ m/\G$mods{none}/gc
687 my $endpos = pos($$textref);
689 $startpos, $oppos-$startpos, # PREFIX
690 $oppos, 0, # NO OPERATOR
691 $oppos, 1, # LEFT DEL
692 $oppos+1, $rd1pos-$oppos-1, # STR/PAT
693 $rd1pos, 1, # RIGHT DEL
694 $modpos, 0, # NO 2ND LDEL
695 $modpos, 0, # NO 2ND STR
696 $modpos, 0, # NO 2ND RDEL
697 $modpos, $endpos-$modpos, # MODIFIERS
698 $endpos, $textlen-$endpos, # REMAINDER
702 unless ($$textref =~ m{\G((?:m|s|qq|qx|qw|q|qr|tr|y)\b(?=\s*\S)|<<)}gc)
704 _failmsg q{No quotelike operator found after prefix at "} .
705 substr($$textref, pos($$textref), 20) .
708 pos $$textref = $startpos;
713 $preld1pos = pos($$textref);
715 $ld1pos = pos($$textref);
717 if ($$textref =~ m{\G([A-Za-z_]\w*)}gc) {
720 elsif ($$textref =~ m{ \G ' ([^'\\]* (?:\\.[^'\\]*)*) '
721 | \G " ([^"\\]* (?:\\.[^"\\]*)*) "
722 | \G ` ([^`\\]* (?:\\.[^`\\]*)*) `
729 my $extrapos = pos($$textref);
730 $$textref =~ m{.*\n}gc;
731 $str1pos = pos($$textref);
732 unless ($$textref =~ m{.*?\n(?=$label\n)}gc) {
733 _failmsg qq{Missing here doc terminator ('$label') after "} .
734 substr($$textref, $startpos, 20) .
737 pos $$textref = $startpos;
740 $rd1pos = pos($$textref);
741 $$textref =~ m{$label\n}gc;
742 $ld2pos = pos($$textref);
744 $startpos, $oppos-$startpos, # PREFIX
745 $oppos, length($op), # OPERATOR
746 $ld1pos, $extrapos-$ld1pos, # LEFT DEL
747 $str1pos, $rd1pos-$str1pos, # STR/PAT
748 $rd1pos, $ld2pos-$rd1pos, # RIGHT DEL
749 $ld2pos, 0, # NO 2ND LDEL
750 $ld2pos, 0, # NO 2ND STR
751 $ld2pos, 0, # NO 2ND RDEL
752 $ld2pos, 0, # NO MODIFIERS
753 $ld2pos, $textlen-$ld2pos, # REMAINDER
754 $extrapos, $str1pos-$extrapos, # FILLETED BIT
758 $$textref =~ m/\G\s*/gc;
759 $ld1pos = pos($$textref);
760 $str1pos = $ld1pos+1;
762 unless ($$textref =~ m/\G(\S)/gc) # SHOULD USE LOOKAHEAD
764 _failmsg "No block delimiter found after quotelike $op",
766 pos $$textref = $startpos;
769 pos($$textref) = $ld1pos; # HAVE TO DO THIS BECAUSE LOOKAHEAD BROKEN
770 my ($ldel1, $rdel1) = ("\Q$1","\Q$1");
771 if ($ldel1 =~ /[[(<{]/)
773 $rdel1 =~ tr/[({</])}>/;
774 _match_bracketed($textref,"",$ldel1,"","",$rdel1)
775 || do { pos $$textref = $startpos; return };
779 $$textref =~ /$ldel1[^\\$ldel1]*(\\.[^\\$ldel1]*)*$ldel1/gcs
780 || do { pos $$textref = $startpos; return };
782 $ld2pos = $rd1pos = pos($$textref)-1;
784 my $second_arg = $op =~ /s|tr|y/ ? 1 : 0;
788 if ($ldel1 =~ /[[(<{]/)
790 unless ($$textref =~ /\G\s*(\S)/gc) # SHOULD USE LOOKAHEAD
792 _failmsg "Missing second block for quotelike $op",
794 pos $$textref = $startpos;
797 $ldel2 = $rdel2 = "\Q$1";
798 $rdel2 =~ tr/[({</])}>/;
802 $ldel2 = $rdel2 = $ldel1;
804 $str2pos = $ld2pos+1;
806 if ($ldel2 =~ /[[(<{]/)
808 pos($$textref)--; # OVERCOME BROKEN LOOKAHEAD
809 _match_bracketed($textref,"",$ldel2,"","",$rdel2)
810 || do { pos $$textref = $startpos; return };
814 $$textref =~ /[^\\$ldel2]*(\\.[^\\$ldel2]*)*$ldel2/gcs
815 || do { pos $$textref = $startpos; return };
817 $rd2pos = pos($$textref)-1;
821 $ld2pos = $str2pos = $rd2pos = $rd1pos;
824 $modpos = pos $$textref;
826 $$textref =~ m/\G($mods{$op})/gc;
827 my $endpos = pos $$textref;
830 $startpos, $oppos-$startpos, # PREFIX
831 $oppos, length($op), # OPERATOR
832 $ld1pos, 1, # LEFT DEL
833 $str1pos, $rd1pos-$str1pos, # STR/PAT
834 $rd1pos, 1, # RIGHT DEL
835 $ld2pos, $second_arg, # 2ND LDEL (MAYBE)
836 $str2pos, $rd2pos-$str2pos, # 2ND STR (MAYBE)
837 $rd2pos, $second_arg, # 2ND RDEL (MAYBE)
838 $modpos, $endpos-$modpos, # MODIFIERS
839 $endpos, $textlen-$endpos, # REMAINDER
845 sub { extract_variable($_[0], '') },
846 sub { extract_quotelike($_[0],'') },
847 sub { extract_codeblock($_[0],'{}','') },
850 sub extract_multiple (;$$$$) # ($text, $functions_ref, $max_fields, $ignoreunknown)
852 my $textref = defined($_[0]) ? \$_[0] : \$_;
854 my ($lastpos, $firstpos);
859 my @func = defined $_[1] ? @{$_[1]} : @{$def_func};
860 my $max = defined $_[2] && $_[2]>0 ? $_[2] : 1_000_000_000;
868 carp "extract_multiple reset maximal count to 1 in scalar context"
869 if $^W && defined($_[2]) && $max > 1;
878 foreach $func ( @func )
880 if (ref($func) eq 'HASH')
882 push @class, (keys %$func)[0];
883 $func = (values %$func)[0];
891 FIELD: while (pos() < length())
894 foreach my $i ( 0..$#func )
899 if (ref($func) eq 'CODE')
900 { ($field) = $func->($_) }
901 elsif (ref($func) eq 'Text::Balanced::Extractor')
902 { $field = $func->extract($_) }
903 elsif( m/\G$func/gc )
904 { $field = defined($1) ? $1 : $& }
906 if (defined($field) && length($field))
908 if (defined($unkpos) && !$igunk)
910 push @fields, substr($_, $unkpos, $lastpos-$unkpos);
911 $firstpos = $unkpos unless defined $firstpos;
913 last FIELD if @fields == $max;
916 ? bless(\$field, $class)
918 $firstpos = $lastpos unless defined $firstpos;
920 last FIELD if @fields == $max;
927 unless $igunk || defined $unkpos;
933 push @fields, substr($_, $unkpos);
934 $firstpos = $unkpos unless defined $firstpos;
940 pos $$textref = $lastpos;
941 return @fields if wantarray;
944 eval { substr($$textref,$firstpos,$lastpos-$firstpos)="";
945 pos $$textref = $firstpos };
950 sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options)
954 my $pre = defined $_[2] ? $_[2] : '\s*';
955 my %options = defined $_[3] ? %{$_[3]} : ();
956 my $omode = defined $options{fail} ? $options{fail} : '';
957 my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}})
958 : defined($options{reject}) ? $options{reject}
961 my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}})
962 : defined($options{ignore}) ? $options{ignore}
966 if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; }
969 for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ }
974 my $textref = defined $_[0] ? \$_[0] : \$_;
975 my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore);
977 return _fail(wantarray, $textref) unless @match;
978 return _succeed wantarray, $textref,
979 $match[2], $match[3]+$match[5]+$match[7], # MATCH
980 @match[8..9,0..1,2..7]; # REM, PRE, BITS
983 bless $closure, 'Text::Balanced::Extractor';
986 package Text::Balanced::Extractor;
988 sub extract($$) # ($self, $text)
993 package Text::Balanced::ErrorMsg;
995 use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" };
1003 Text::Balanced - Extract delimited text sequences from strings.
1008 use Text::Balanced qw (
1021 # Extract the initial substring of $text that is delimited by
1022 # two (unescaped) instances of the first character in $delim.
1024 ($extracted, $remainder) = extract_delimited($text,$delim);
1027 # Extract the initial substring of $text that is bracketed
1028 # with a delimiter(s) specified by $delim (where the string
1029 # in $delim contains one or more of '(){}[]<>').
1031 ($extracted, $remainder) = extract_bracketed($text,$delim);
1034 # Extract the initial substring of $text that is bounded by
1037 ($extracted, $remainder) = extract_tagged($text);
1040 # Extract the initial substring of $text that is bounded by
1041 # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags
1043 ($extracted, $remainder) =
1044 extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]});
1047 # Extract the initial substring of $text that represents a
1048 # Perl "quote or quote-like operation"
1050 ($extracted, $remainder) = extract_quotelike($text);
1053 # Extract the initial substring of $text that represents a block
1054 # of Perl code, bracketed by any of character(s) specified by $delim
1055 # (where the string $delim contains one or more of '(){}[]<>').
1057 ($extracted, $remainder) = extract_codeblock($text,$delim);
1060 # Extract the initial substrings of $text that would be extracted by
1061 # one or more sequential applications of the specified functions
1062 # or regular expressions
1064 @extracted = extract_multiple($text,
1065 [ \&extract_bracketed,
1066 \&extract_quotelike,
1067 \&some_other_extractor_sub,
1072 # Create a string representing an optimized pattern (a la Friedl)
1073 # that matches a substring delimited by any of the specified characters
1074 # (in this case: any type of quote or a slash)
1076 $patstring = gen_delimited_pat(q{'"`/});
1079 # Generate a reference to an anonymous sub that is just like extract_tagged
1080 # but pre-compiled and optimized for a specific pair of tags, and consequently
1081 # much faster (i.e. 3 times faster). It uses qr// for better performance on
1082 # repeated calls, so it only works under Perl 5.005 or later.
1084 $extract_head = gen_extract_tagged('<HEAD>','</HEAD>');
1086 ($extracted, $remainder) = $extract_head->($text);
1091 The various C<extract_...> subroutines may be used to extract a
1092 delimited string (possibly after skipping a specified prefix string).
1093 The search for the string always begins at the current C<pos>
1094 location of the string's variable (or at index zero, if no C<pos>
1095 position is defined).
1097 =head2 General behaviour in list contexts
1099 In a list context, all the subroutines return a list, the first three
1100 elements of which are always:
1106 The extracted string, including the specified delimiters.
1107 If the extraction fails an empty string is returned.
1111 The remainder of the input string (i.e. the characters after the
1112 extracted string). On failure, the entire string is returned.
1116 The skipped prefix (i.e. the characters before the extracted string).
1117 On failure, the empty string is returned.
1121 Note that in a list context, the contents of the original input text (the first
1122 argument) are not modified in any way.
1124 However, if the input text was passed in a variable, that variable's
1125 C<pos> value is updated to point at the first character after the
1126 extracted text. That means that in a list context the various
1127 subroutines can be used much like regular expressions. For example:
1129 while ( $next = (extract_quotelike($text))[0] )
1131 # process next quote-like (in $next)
1135 =head2 General behaviour in scalar and void contexts
1137 In a scalar context, the extracted string is returned, having first been
1138 removed from the input text. Thus, the following code also processes
1139 each quote-like operation, but actually removes them from $text:
1141 while ( $next = extract_quotelike($text) )
1143 # process next quote-like (in $next)
1146 Note that if the input text is a read-only string (i.e. a literal),
1147 no attempt is made to remove the extracted text.
1149 In a void context the behaviour of the extraction subroutines is
1150 exactly the same as in a scalar context, except (of course) that the
1151 extracted substring is not returned.
1153 =head2 A note about prefixes
1155 Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.)
1156 This can bite you if you're expecting a prefix specification like
1157 '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix
1158 pattern will only succeed if the <H1> tag is on the current line, since
1159 . normally doesn't match newlines.
1161 To overcome this limitation, you need to turn on /s matching within
1162 the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)'
1165 =head2 C<extract_delimited>
1167 The C<extract_delimited> function formalizes the common idiom
1168 of extracting a single-character-delimited substring from the start of
1169 a string. For example, to extract a single-quote delimited string, the
1170 following code is typically used:
1172 ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s;
1175 but with C<extract_delimited> it can be simplified to:
1177 ($extracted,$remainder) = extract_delimited($text, "'");
1179 C<extract_delimited> takes up to four scalars (the input text, the
1180 delimiters, a prefix pattern to be skipped, and any escape characters)
1181 and extracts the initial substring of the text that
1182 is appropriately delimited. If the delimiter string has multiple
1183 characters, the first one encountered in the text is taken to delimit
1185 The third argument specifies a prefix pattern that is to be skipped
1186 (but must be present!) before the substring is extracted.
1187 The final argument specifies the escape character to be used for each
1190 All arguments are optional. If the escape characters are not specified,
1191 every delimiter is escaped with a backslash (C<\>).
1192 If the prefix is not specified, the
1193 pattern C<'\s*'> - optional whitespace - is used. If the delimiter set
1194 is also not specified, the set C</["'`]/> is used. If the text to be processed
1195 is not specified either, C<$_> is used.
1197 In list context, C<extract_delimited> returns a array of three
1198 elements, the extracted substring (I<including the surrounding
1199 delimiters>), the remainder of the text, and the skipped prefix (if
1200 any). If a suitable delimited substring is not found, the first
1201 element of the array is the empty string, the second is the complete
1202 original text, and the prefix returned in the third element is an
1205 In a scalar context, just the extracted substring is returned. In
1206 a void context, the extracted substring (and any prefix) are simply
1207 removed from the beginning of the first argument.
1211 # Remove a single-quoted substring from the very beginning of $text:
1213 $substring = extract_delimited($text, "'", '');
1215 # Remove a single-quoted Pascalish substring (i.e. one in which
1216 # doubling the quote character escapes it) from the very
1217 # beginning of $text:
1219 $substring = extract_delimited($text, "'", '', "'");
1221 # Extract a single- or double- quoted substring from the
1222 # beginning of $text, optionally after some whitespace
1223 # (note the list context to protect $text from modification):
1225 ($substring) = extract_delimited $text, q{"'};
1228 # Delete the substring delimited by the first '/' in $text:
1230 $text = join '', (extract_delimited($text,'/','[^/]*')[2,1];
1232 Note that this last example is I<not> the same as deleting the first
1233 quote-like pattern. For instance, if C<$text> contained the string:
1235 "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }"
1237 then after the deletion it would contain:
1239 "if ('.$UNIXCMD/s) { $cmd = $1; }"
1243 "if ('./cmd' =~ ms) { $cmd = $1; }"
1246 See L<"extract_quotelike"> for a (partial) solution to this problem.
1249 =head2 C<extract_bracketed>
1251 Like C<"extract_delimited">, the C<extract_bracketed> function takes
1252 up to three optional scalar arguments: a string to extract from, a delimiter
1253 specifier, and a prefix pattern. As before, a missing prefix defaults to
1254 optional whitespace and a missing text defaults to C<$_>. However, a missing
1255 delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below).
1257 C<extract_bracketed> extracts a balanced-bracket-delimited
1258 substring (using any one (or more) of the user-specified delimiter
1259 brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also
1260 respect quoted unbalanced brackets (see below).
1262 A "delimiter bracket" is a bracket in list of delimiters passed as
1263 C<extract_bracketed>'s second argument. Delimiter brackets are
1264 specified by giving either the left or right (or both!) versions
1265 of the required bracket(s). Note that the order in which
1266 two or more delimiter brackets are specified is not significant.
1268 A "balanced-bracket-delimited substring" is a substring bounded by
1269 matched brackets, such that any other (left or right) delimiter
1270 bracket I<within> the substring is also matched by an opposite
1271 (right or left) delimiter bracket I<at the same level of nesting>. Any
1272 type of bracket not in the delimiter list is treated as an ordinary
1275 In other words, each type of bracket specified as a delimiter must be
1276 balanced and correctly nested within the substring, and any other kind of
1277 ("non-delimiter") bracket in the substring is ignored.
1279 For example, given the string:
1281 $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }";
1283 then a call to C<extract_bracketed> in a list context:
1285 @result = extract_bracketed( $text, '{}' );
1289 ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" )
1291 since both sets of C<'{..}'> brackets are properly nested and evenly balanced.
1292 (In a scalar context just the first element of the array would be returned. In
1293 a void context, C<$text> would be replaced by an empty string.)
1295 Likewise the call in:
1297 @result = extract_bracketed( $text, '{[' );
1299 would return the same result, since all sets of both types of specified
1300 delimiter brackets are correctly nested and balanced.
1302 However, the call in:
1304 @result = extract_bracketed( $text, '{([<' );
1306 would fail, returning:
1308 ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" );
1310 because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and
1311 the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would
1312 return an empty string. In a void context, C<$text> would be unchanged.)
1314 Note that the embedded single-quotes in the string don't help in this
1315 case, since they have not been specified as acceptable delimiters and are
1316 therefore treated as non-delimiter characters (and ignored).
1318 However, if a particular species of quote character is included in the
1319 delimiter specification, then that type of quote will be correctly handled.
1320 for example, if C<$text> is:
1322 $text = '<A HREF=">>>>">link</A>';
1326 @result = extract_bracketed( $text, '<">' );
1330 ( '<A HREF=">>>>">', 'link</A>', "" )
1332 as expected. Without the specification of C<"> as an embedded quoter:
1334 @result = extract_bracketed( $text, '<>' );
1336 the result would be:
1338 ( '<A HREF=">', '>>>">link</A>', "" )
1340 In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like
1341 quoting (i.e. q{string}, qq{string}, etc) can be specified by including the
1342 letter 'q' as a delimiter. Hence:
1344 @result = extract_bracketed( $text, '<q>' );
1346 would correctly match something like this:
1348 $text = '<leftop: conj /and/ conj>';
1350 See also: C<"extract_quotelike"> and C<"extract_codeblock">.
1353 =head2 C<extract_tagged>
1355 C<extract_tagged> extracts and segments text between (balanced)
1358 The subroutine takes up to five optional arguments:
1364 A string to be processed (C<$_> if the string is omitted or C<undef>)
1368 A string specifying a pattern to be matched as the opening tag.
1369 If the pattern string is omitted (or C<undef>) then a pattern
1370 that matches any standard HTML/XML tag is used.
1374 A string specifying a pattern to be matched at the closing tag.
1375 If the pattern string is omitted (or C<undef>) then the closing
1376 tag is constructed by inserting a C</> after any leading bracket
1377 characters in the actual opening tag that was matched (I<not> the pattern
1378 that matched the tag). For example, if the opening tag pattern
1379 is specified as C<'{{\w+}}'> and actually matched the opening tag
1380 C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">.
1384 A string specifying a pattern to be matched as a prefix (which is to be
1385 skipped). If omitted, optional whitespace is skipped.
1389 A hash reference containing various parsing options (see below)
1393 The various options that can be specified are:
1397 =item C<reject =E<gt> $listref>
1399 The list reference contains one or more strings specifying patterns
1400 that must I<not> appear within the tagged text.
1402 For example, to extract
1403 an HTML link (which should not contain nested links) use:
1405 extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} );
1407 =item C<ignore =E<gt> $listref>
1409 The list reference contains one or more strings specifying patterns
1410 that are I<not> be be treated as nested tags within the tagged text
1411 (even if they would match the start tag pattern).
1413 For example, to extract an arbitrary XML tag, but ignore "empty" elements:
1415 extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} );
1417 (also see L<"gen_delimited_pat"> below).
1420 =item C<fail =E<gt> $str>
1422 The C<fail> option indicates the action to be taken if a matching end
1423 tag is not encountered (i.e. before the end of the string or some
1424 C<reject> pattern matches). By default, a failure to match a closing
1425 tag causes C<extract_tagged> to immediately fail.
1427 However, if the string value associated with <reject> is "MAX", then
1428 C<extract_tagged> returns the complete text up to the point of failure.
1429 If the string is "PARA", C<extract_tagged> returns only the first paragraph
1430 after the tag (up to the first line that is either empty or contains
1431 only whitespace characters).
1432 If the string is "", the the default behaviour (i.e. failure) is reinstated.
1434 For example, suppose the start tag "/para" introduces a paragraph, which then
1435 continues until the next "/endpara" tag or until another "/para" tag is
1438 $text = "/para line 1\n\nline 3\n/para line 4";
1440 extract_tagged($text, '/para', '/endpara', undef,
1441 {reject => '/para', fail => MAX );
1443 # EXTRACTED: "/para line 1\n\nline 3\n"
1445 Suppose instead, that if no matching "/endpara" tag is found, the "/para"
1446 tag refers only to the immediately following paragraph:
1448 $text = "/para line 1\n\nline 3\n/para line 4";
1450 extract_tagged($text, '/para', '/endpara', undef,
1451 {reject => '/para', fail => MAX );
1453 # EXTRACTED: "/para line 1\n"
1455 Note that the specified C<fail> behaviour applies to nested tags as well.
1459 On success in a list context, an array of 6 elements is returned. The elements are:
1465 the extracted tagged substring (including the outermost tags),
1469 the remainder of the input text,
1473 the prefix substring (if any),
1481 the text between the opening and closing tags
1485 the closing tag (or "" if no closing tag was found)
1489 On failure, all of these values (except the remaining text) are C<undef>.
1491 In a scalar context, C<extract_tagged> returns just the complete
1492 substring that matched a tagged text (including the start and end
1493 tags). C<undef> is returned on failure. In addition, the original input
1494 text has the returned substring (and any prefix) removed from it.
1496 In a void context, the input text just has the matched substring (and
1497 any specified prefix) removed.
1500 =head2 C<gen_extract_tagged>
1502 (Note: This subroutine is only available under Perl5.005)
1504 C<gen_extract_tagged> generates a new anonymous subroutine which
1505 extracts text between (balanced) specified tags. In other words,
1506 it generates a function identical in function to C<extract_tagged>.
1508 The difference between C<extract_tagged> and the anonymous
1509 subroutines generated by
1510 C<gen_extract_tagged>, is that those generated subroutines:
1516 do not have to reparse tag specification or parsing options every time
1517 they are called (whereas C<extract_tagged> has to effectively rebuild
1518 its tag parser on every call);
1522 make use of the new qr// construct to pre-compile the regexes they use
1523 (whereas C<extract_tagged> uses standard string variable interpolation
1524 to create tag-matching patterns).
1528 The subroutine takes up to four optional arguments (the same set as
1529 C<extract_tagged> except for the string to be processed). It returns
1530 a reference to a subroutine which in turn takes a single argument (the text to
1533 In other words, the implementation of C<extract_tagged> is exactly
1539 $extractor = gen_extract_tagged(@_);
1540 return $extractor->($text);
1543 (although C<extract_tagged> is not currently implemented that way, in order
1544 to preserve pre-5.005 compatibility).
1546 Using C<gen_extract_tagged> to create extraction functions for specific tags
1547 is a good idea if those functions are going to be called more than once, since
1548 their performance is typically twice as good as the more general-purpose
1552 =head2 C<extract_quotelike>
1554 C<extract_quotelike> attempts to recognize, extract, and segment any
1555 one of the various Perl quotes and quotelike operators (see
1556 L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket
1557 delimiters (for the quotelike operators), and trailing modifiers are
1558 all caught. For example, in:
1560 extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #'
1562 extract_quotelike ' "You said, \"Use sed\"." '
1564 extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; '
1566 extract_quotelike ' tr/\\\/\\\\/\\\//ds; '
1568 the full Perl quotelike operations are all extracted correctly.
1570 Note too that, when using the /x modifier on a regex, any comment
1571 containing the current pattern delimiter will cause the regex to be
1572 immediately terminated. In other words:
1575 (?i) # CASE INSENSITIVE
1576 [a-z_] # LEADING ALPHABETIC/UNDERSCORE
1577 [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS
1580 will be extracted as if it were:
1583 (?i) # CASE INSENSITIVE
1584 [a-z_] # LEADING ALPHABETIC/'
1586 This behaviour is identical to that of the actual compiler.
1588 C<extract_quotelike> takes two arguments: the text to be processed and
1589 a prefix to be matched at the very beginning of the text. If no prefix
1590 is specified, optional whitespace is the default. If no text is given,
1593 In a list context, an array of 11 elements is returned. The elements are:
1599 the extracted quotelike substring (including trailing modifiers),
1603 the remainder of the input text,
1607 the prefix substring (if any),
1611 the name of the quotelike operator (if any),
1615 the left delimiter of the first block of the operation,
1619 the text of the first block of the operation
1620 (that is, the contents of
1621 a quote, the regex of a match or substitution or the target list of a
1626 the right delimiter of the first block of the operation,
1630 the left delimiter of the second block of the operation
1631 (that is, if it is a C<s>, C<tr>, or C<y>),
1635 the text of the second block of the operation
1636 (that is, the replacement of a substitution or the translation list
1641 the right delimiter of the second block of the operation (if any),
1645 the trailing modifiers on the operation (if any).
1649 For each of the fields marked "(if any)" the default value on success is
1651 On failure, all of these values (except the remaining text) are C<undef>.
1654 In a scalar context, C<extract_quotelike> returns just the complete substring
1655 that matched a quotelike operation (or C<undef> on failure). In a scalar or
1656 void context, the input text has the same substring (and any specified
1661 # Remove the first quotelike literal that appears in text
1663 $quotelike = extract_quotelike($text,'.*?');
1665 # Replace one or more leading whitespace-separated quotelike
1666 # literals in $_ with "<QLL>"
1668 do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@;
1671 # Isolate the search pattern in a quotelike operation from $text
1673 ($op,$pat) = (extract_quotelike $text)[3,5];
1676 print "search pattern: $pat\n";
1680 print "$op is not a pattern matching operation\n";
1684 =head2 C<extract_quotelike> and "here documents"
1686 C<extract_quotelike> can successfully extract "here documents" from an input
1687 string, but with an important caveat in list contexts.
1689 Unlike other types of quote-like literals, a here document is rarely
1690 a contiguous substring. For example, a typical piece of code using
1691 here document might look like this:
1694 This is the message.
1698 Given this as an input string in a scalar context, C<extract_quotelike>
1699 would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG",
1700 leaving the string " || die;\nexit;" in the original variable. In other words,
1701 the two separate pieces of the here document are successfully extracted and
1704 In a list context, C<extract_quotelike> would return the list
1710 "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document,
1711 including fore and aft delimiters),
1715 " || die;\nexit;" (i.e. the remainder of the input text, concatenated),
1719 "" (i.e. the prefix substring -- trivial in this case),
1723 "<<" (i.e. the "name" of the quotelike operator)
1727 "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes),
1731 "This is the message.\n" (i.e. the text of the here document),
1735 "EOMSG" (i.e. the right delimiter of the here document),
1739 "" (a here document has no second left delimiter, second text, second right
1740 delimiter, or trailing modifiers).
1744 However, the matching position of the input variable would be set to
1745 "exit;" (i.e. I<after> the closing delimiter of the here document),
1746 which would cause the earlier " || die;\nexit;" to be skipped in any
1747 sequence of code fragment extractions.
1749 To avoid this problem, when it encounters a here document whilst
1750 extracting from a modifiable string, C<extract_quotelike> silently
1751 rearranges the string to an equivalent piece of Perl:
1754 This is the message.
1759 in which the here document I<is> contiguous. It still leaves the
1760 matching position after the here document, but now the rest of the line
1761 on which the here document starts is not skipped.
1763 To prevent <extract_quotelike> from mucking about with the input in this way
1764 (this is the only case where a list-context C<extract_quotelike> does so),
1765 you can pass the input variable as an interpolated literal:
1767 $quotelike = extract_quotelike("$var");
1770 =head2 C<extract_codeblock>
1772 C<extract_codeblock> attempts to recognize and extract a balanced
1773 bracket delimited substring that may contain unbalanced brackets
1774 inside Perl quotes or quotelike operations. That is, C<extract_codeblock>
1775 is like a combination of C<"extract_bracketed"> and
1776 C<"extract_quotelike">.
1778 C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>:
1779 a text to process, a set of delimiter brackets to look for, and a prefix to
1780 match first. It also takes an optional fourth parameter, which allows the
1781 outermost delimiter brackets to be specified separately (see below).
1783 Omitting the first argument (input text) means process C<$_> instead.
1784 Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used.
1785 Omitting the third argument (prefix argument) implies optional whitespace at the start.
1786 Omitting the fourth argument (outermost delimiter brackets) indicates that the
1787 value of the second argument is to be used for the outermost delimiters.
1789 Once the prefix an dthe outermost opening delimiter bracket have been
1790 recognized, code blocks are extracted by stepping through the input text and
1791 trying the following alternatives in sequence:
1797 Try and match a closing delimiter bracket. If the bracket was the same
1798 species as the last opening bracket, return the substring to that
1799 point. If the bracket was mismatched, return an error.
1803 Try to match a quote or quotelike operator. If found, call
1804 C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return
1805 the error it returned. Otherwise go back to step 1.
1809 Try to match an opening delimiter bracket. If found, call
1810 C<extract_codeblock> recursively to eat the embedded block. If the
1811 recursive call fails, return an error. Otherwise, go back to step 1.
1815 Unconditionally match a bareword or any other single character, and
1816 then go back to step 1.
1823 # Find a while loop in the text
1825 if ($text =~ s/.*?while\s*\{/{/)
1827 $loop = "while " . extract_codeblock($text);
1830 # Remove the first round-bracketed list (which may include
1831 # round- or curly-bracketed code blocks or quotelike operators)
1833 extract_codeblock $text, "(){}", '[^(]*';
1836 The ability to specify a different outermost delimiter bracket is useful
1837 in some circumstances. For example, in the Parse::RecDescent module,
1838 parser actions which are to be performed only on a successful parse
1839 are specified using a C<E<lt>defer:...E<gt>> directive. For example:
1841 sentence: subject verb object
1842 <defer: {$::theVerb = $item{verb}} >
1844 Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code
1845 within the C<E<lt>defer:...E<gt>> directive, but there's a problem.
1847 A deferred action like this:
1849 <defer: {if ($count>10) {$count--}} >
1851 will be incorrectly parsed as:
1853 <defer: {if ($count>
1855 because the "less than" operator is interpreted as a closing delimiter.
1857 But, by extracting the directive using
1858 S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>>
1859 the '>' character is only treated as a delimited at the outermost
1860 level of the code block, so the directive is parsed correctly.
1862 =head2 C<extract_multiple>
1864 The C<extract_multiple> subroutine takes a string to be processed and a
1865 list of extractors (subroutines or regular expressions) to apply to that string.
1867 In an array context C<extract_multiple> returns an array of substrings
1868 of the original string, as extracted by the specified extractors.
1869 In a scalar context, C<extract_multiple> returns the first
1870 substring successfully extracted from the original string. In both
1871 scalar and void contexts the original string has the first successfully
1872 extracted substring removed from it. In all contexts
1873 C<extract_multiple> starts at the current C<pos> of the string, and
1874 sets that C<pos> appropriately after it matches.
1876 Hence, the aim of of a call to C<extract_multiple> in a list context
1877 is to split the processed string into as many non-overlapping fields as
1878 possible, by repeatedly applying each of the specified extractors
1879 to the remainder of the string. Thus C<extract_multiple> is
1880 a generalized form of Perl's C<split> subroutine.
1882 The subroutine takes up to four optional arguments:
1888 A string to be processed (C<$_> if the string is omitted or C<undef>)
1892 A reference to a list of subroutine references and/or qr// objects and/or
1893 literal strings and/or hash references, specifying the extractors
1894 to be used to split the string. If this argument is omitted (or
1898 sub { extract_variable($_[0], '') },
1899 sub { extract_quotelike($_[0],'') },
1900 sub { extract_codeblock($_[0],'{}','') },
1908 An number specifying the maximum number of fields to return. If this
1909 argument is omitted (or C<undef>), split continues as long as possible.
1911 If the third argument is I<N>, then extraction continues until I<N> fields
1912 have been successfully extracted, or until the string has been completely
1915 Note that in scalar and void contexts the value of this argument is
1916 automatically reset to 1 (under C<-w>, a warning is issued if the argument
1921 A value indicating whether unmatched substrings (see below) within the
1922 text should be skipped or returned as fields. If the value is true,
1923 such substrings are skipped. Otherwise, they are returned.
1927 The extraction process works by applying each extractor in
1928 sequence to the text string. If the extractor is a subroutine it
1930 context and is expected to return a list of a single element, namely
1932 Note that the value returned by an extractor subroutine need not bear any
1933 relationship to the corresponding substring of the original text (see
1936 If the extractor is a precompiled regular expression or a string,
1937 it is matched against the text in a scalar context with a leading
1938 '\G' and the gc modifiers enabled. The extracted value is either
1939 $1 if that variable is defined after the match, or else the
1940 complete match (i.e. $&).
1942 If the extractor is a hash reference, it must contain exactly one element.
1943 The value of that element is one of the
1944 above extractor types (subroutine reference, regular expression, or string).
1945 The key of that element is the name of a class into which the successful
1946 return value of the extractor will be blessed.
1948 If an extractor returns a defined value, that value is immediately
1949 treated as the next extracted field and pushed onto the list of fields.
1950 If the extractor was specified in a hash reference, the field is also
1951 blessed into the appropriate class,
1953 If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is
1954 assumed to have failed to extract.
1955 If none of the extractor subroutines succeeds, then one
1956 character is extracted from the start of the text and the extraction
1957 subroutines reapplied. Characters which are thus removed are accumulated and
1958 eventually become the next field (unless the fourth argument is true, in which
1959 case they are disgarded).
1961 For example, the following extracts substrings that are valid Perl variables:
1963 @fields = extract_multiple($text,
1964 [ sub { extract_variable($_[0]) } ],
1967 This example separates a text into fields which are quote delimited,
1968 curly bracketed, and anything else. The delimited and bracketed
1969 parts are also blessed to identify them (the "anything else" is unblessed):
1971 @fields = extract_multiple($text,
1973 { Delim => sub { extract_delimited($_[0],q{'"}) } },
1974 { Brack => sub { extract_bracketed($_[0],'{}') } },
1977 This call extracts the next single substring that is a valid Perl quotelike
1978 operator (and removes it from $text):
1980 $quotelike = extract_multiple($text,
1982 sub { extract_quotelike($_[0]) },
1985 Finally, here is yet another way to do comma-separated value parsing:
1987 @fields = extract_multiple($csv_text,
1989 sub { extract_delimited($_[0],q{'"}) },
1994 The list in the second argument means:
1995 I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">.
1996 The undef third argument means:
1997 I<"...as many times as possible...">,
1998 and the true value in the fourth argument means
1999 I<"...discarding anything else that appears (i.e. the commas)">.
2001 If you wanted the commas preserved as separate fields (i.e. like split
2002 does if your split pattern has capturing parentheses), you would
2003 just make the last parameter undefined (or remove it).
2006 =head2 C<gen_delimited_pat>
2008 The C<gen_delimited_pat> subroutine takes a single (string) argument and
2009 > builds a Friedl-style optimized regex that matches a string delimited
2010 by any one of the characters in the single argument. For example:
2012 gen_delimited_pat(q{'"})
2016 (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\')
2018 Note that the specified delimiters are automatically quotemeta'd.
2020 A typical use of C<gen_delimited_pat> would be to build special purpose tags
2021 for C<extract_tagged>. For example, to properly ignore "empty" XML elements
2022 (which might contain quoted strings):
2024 my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>';
2026 extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} );
2029 C<gen_delimited_pat> may also be called with an optional second argument,
2030 which specifies the "escape" character(s) to be used for each delimiter.
2031 For example to match a Pascal-style string (where ' is the delimiter
2032 and '' is a literal ' within the string):
2034 gen_delimited_pat(q{'},q{'});
2036 Different escape characters can be specified for different delimiters.
2037 For example, to specify that '/' is the escape for single quotes
2038 and '%' is the escape for double quotes:
2040 gen_delimited_pat(q{'"},q{/%});
2042 If more delimiters than escape chars are specified, the last escape char
2043 is used for the remaining delimiters.
2044 If no escape char is specified for a given specified delimiter, '\' is used.
2047 C<gen_delimited_pat> was previously called
2048 C<delimited_pat>. That name may still be used, but is now deprecated.
2053 In a list context, all the functions return C<(undef,$original_text)>
2054 on failure. In a scalar context, failure is indicated by returning C<undef>
2055 (in this case the input text is not modified in any way).
2057 In addition, on failure in I<any> context, the C<$@> variable is set.
2058 Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed
2060 Accessing C<$@-E<gt>{pos}> returns the offset into the original string at
2061 which the error was detected (although not necessarily where it occurred!)
2062 Printing C<$@> directly produces the error message, with the offset appended.
2063 On success, the C<$@> variable is guaranteed to be C<undef>.
2065 The available diagnostics are:
2069 =item C<Did not find a suitable bracket: "%s">
2071 The delimiter provided to C<extract_bracketed> was not one of
2072 C<'()[]E<lt>E<gt>{}'>.
2074 =item C<Did not find prefix: /%s/>
2076 A non-optional prefix was specified but wasn't found at the start of the text.
2078 =item C<Did not find opening bracket after prefix: "%s">
2080 C<extract_bracketed> or C<extract_codeblock> was expecting a
2081 particular kind of bracket at the start of the text, and didn't find it.
2083 =item C<No quotelike operator found after prefix: "%s">
2085 C<extract_quotelike> didn't find one of the quotelike operators C<q>,
2086 C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring
2089 =item C<Unmatched closing bracket: "%c">
2091 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered
2092 a closing bracket where none was expected.
2094 =item C<Unmatched opening bracket(s): "%s">
2096 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran
2097 out of characters in the text before closing one or more levels of nested
2100 =item C<Unmatched embedded quote (%s)>
2102 C<extract_bracketed> attempted to match an embedded quoted substring, but
2103 failed to find a closing quote to match it.
2105 =item C<Did not find closing delimiter to match '%s'>
2107 C<extract_quotelike> was unable to find a closing delimiter to match the
2108 one that opened the quote-like operation.
2110 =item C<Mismatched closing bracket: expected "%c" but found "%s">
2112 C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found
2113 a valid bracket delimiter, but it was the wrong species. This usually
2114 indicates a nesting error, but may indicate incorrect quoting or escaping.
2116 =item C<No block delimiter found after quotelike "%s">
2118 C<extract_quotelike> or C<extract_codeblock> found one of the
2119 quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y>
2120 without a suitable block after it.
2122 =item C<Did not find leading dereferencer>
2124 C<extract_variable> was expecting one of '$', '@', or '%' at the start of
2125 a variable, but didn't find any of them.
2127 =item C<Bad identifier after dereferencer>
2129 C<extract_variable> found a '$', '@', or '%' indicating a variable, but that
2130 character was not followed by a legal Perl identifier.
2132 =item C<Did not find expected opening bracket at %s>
2134 C<extract_codeblock> failed to find any of the outermost opening brackets
2135 that were specified.
2137 =item C<Improperly nested codeblock at %s>
2139 A nested code block was found that started with a delimiter that was specified
2140 as being only to be used as an outermost bracket.
2142 =item C<Missing second block for quotelike "%s">
2144 C<extract_codeblock> or C<extract_quotelike> found one of the
2145 quotelike operators C<s>, C<tr> or C<y> followed by only one block.
2147 =item C<No match found for opening bracket>
2149 C<extract_codeblock> failed to find a closing bracket to match the outermost
2152 =item C<Did not find opening tag: /%s/>
2154 C<extract_tagged> did not find a suitable opening tag (after any specified
2155 prefix was removed).
2157 =item C<Unable to construct closing tag to match: /%s/>
2159 C<extract_tagged> matched the specified opening tag and tried to
2160 modify the matched text to produce a matching closing tag (because
2161 none was specified). It failed to generate the closing tag, almost
2162 certainly because the opening tag did not start with a
2163 bracket of some kind.
2165 =item C<Found invalid nested tag: %s>
2167 C<extract_tagged> found a nested tag that appeared in the "reject" list
2168 (and the failure mode was not "MAX" or "PARA").
2170 =item C<Found unbalanced nested tag: %s>
2172 C<extract_tagged> found a nested opening tag that was not matched by a
2173 corresponding nested closing tag (and the failure mode was not "MAX" or "PARA").
2175 =item C<Did not find closing tag>
2177 C<extract_tagged> reached the end of the text without finding a closing tag
2178 to match the original opening tag (and the failure mode was not
2189 Damian Conway (damian@conway.org)
2192 =head1 BUGS AND IRRITATIONS
2194 There are undoubtedly serious bugs lurking somewhere in this code, if
2195 only because parts of it give the impression of understanding a great deal
2196 more about Perl than they really do.
2198 Bug reports and other feedback are most welcome.
2203 Copyright (c) 1997-2001, Damian Conway. All Rights Reserved.
2204 This module is free software. It may be used, redistributed
2205 and/or modified under the same terms as Perl itself.