Commit | Line | Data |
3270c621 |
1 | # EXTRACT VARIOUSLY DELIMITED TEXT SEQUENCES FROM STRINGS. |
2 | # FOR FULL DOCUMENTATION SEE Balanced.pod |
3 | |
4 | use 5.005; |
5 | use strict; |
6 | |
7 | package Text::Balanced; |
8 | |
9 | use Exporter; |
10 | use SelfLoader; |
11 | use vars qw { $VERSION @ISA %EXPORT_TAGS }; |
12 | |
a7602084 |
13 | $VERSION = '1.89'; |
3270c621 |
14 | @ISA = qw ( Exporter ); |
15 | |
16 | %EXPORT_TAGS = ( ALL => [ qw( |
17 | &extract_delimited |
18 | &extract_bracketed |
19 | &extract_quotelike |
20 | &extract_codeblock |
21 | &extract_variable |
22 | &extract_tagged |
23 | &extract_multiple |
24 | |
25 | &gen_delimited_pat |
26 | &gen_extract_tagged |
27 | |
28 | &delimited_pat |
29 | ) ] ); |
30 | |
31 | Exporter::export_ok_tags('ALL'); |
32 | |
24e08cba |
33 | ## |
34 | ## These shenanagins are to avoid using $& in perl5.6+ |
35 | ## |
36 | my $GetMatchedText = ($] < 5.006) ? eval 'sub { $& } ' |
37 | : eval 'sub { |
38 | substr($_[0], $-[0], $+[0] - $-[0]) |
39 | }'; |
40 | |
41 | |
3270c621 |
42 | # PROTOTYPES |
43 | |
44 | sub _match_bracketed($$$$$$); |
45 | sub _match_variable($$); |
46 | sub _match_codeblock($$$$$$$); |
47 | sub _match_quotelike($$$$); |
48 | |
49 | # HANDLE RETURN VALUES IN VARIOUS CONTEXTS |
50 | |
51 | sub _failmsg { |
52 | my ($message, $pos) = @_; |
53 | $@ = bless { error=>$message, pos=>$pos }, "Text::Balanced::ErrorMsg"; |
54 | } |
55 | |
56 | sub _fail |
57 | { |
58 | my ($wantarray, $textref, $message, $pos) = @_; |
59 | _failmsg $message, $pos if $message; |
60 | return ("",$$textref,"") if $wantarray; |
61 | return undef; |
62 | } |
63 | |
64 | sub _succeed |
65 | { |
66 | $@ = undef; |
67 | my ($wantarray,$textref) = splice @_, 0, 2; |
68 | my ($extrapos, $extralen) = @_>18 ? splice(@_, -2, 2) : (0,0); |
69 | my ($startlen) = $_[5]; |
70 | my $remainderpos = $_[2]; |
71 | if ($wantarray) |
72 | { |
73 | my @res; |
74 | while (my ($from, $len) = splice @_, 0, 2) |
75 | { |
76 | push @res, substr($$textref,$from,$len); |
77 | } |
78 | if ($extralen) { # CORRECT FILLET |
79 | my $extra = substr($res[0], $extrapos-$startlen, $extralen, "\n"); |
80 | $res[1] = "$extra$res[1]"; |
81 | eval { substr($$textref,$remainderpos,0) = $extra; |
82 | substr($$textref,$extrapos,$extralen,"\n")} ; |
83 | #REARRANGE HERE DOC AND FILLET IF POSSIBLE |
84 | pos($$textref) = $remainderpos-$extralen+1; # RESET \G |
85 | } |
86 | else { |
87 | pos($$textref) = $remainderpos; # RESET \G |
88 | } |
89 | return @res; |
90 | } |
91 | else |
92 | { |
93 | my $match = substr($$textref,$_[0],$_[1]); |
94 | substr($match,$extrapos-$_[0]-$startlen,$extralen,"") if $extralen; |
95 | my $extra = $extralen |
96 | ? substr($$textref, $extrapos, $extralen)."\n" : ""; |
97 | eval {substr($$textref,$_[4],$_[1]+$_[5])=$extra} ; #CHOP OUT PREFIX & MATCH, IF POSSIBLE |
98 | pos($$textref) = $_[4]; # RESET \G |
99 | return $match; |
100 | } |
101 | } |
102 | |
103 | # BUILD A PATTERN MATCHING A SIMPLE DELIMITED STRING |
104 | |
105 | sub gen_delimited_pat($;$) # ($delimiters;$escapes) |
106 | { |
107 | my ($dels, $escs) = @_; |
108 | return "" unless $dels =~ /\S/; |
109 | $escs = '\\' unless $escs; |
110 | $escs .= substr($escs,-1) x (length($dels)-length($escs)); |
111 | my @pat = (); |
112 | my $i; |
113 | for ($i=0; $i<length $dels; $i++) |
114 | { |
115 | my $del = quotemeta substr($dels,$i,1); |
116 | my $esc = quotemeta substr($escs,$i,1); |
117 | if ($del eq $esc) |
118 | { |
119 | push @pat, "$del(?:[^$del]*(?:(?:$del$del)[^$del]*)*)$del"; |
120 | } |
121 | else |
122 | { |
123 | push @pat, "$del(?:[^$esc$del]*(?:$esc.[^$esc$del]*)*)$del"; |
124 | } |
125 | } |
126 | my $pat = join '|', @pat; |
127 | return "(?:$pat)"; |
128 | } |
129 | |
130 | *delimited_pat = \&gen_delimited_pat; |
131 | |
132 | |
133 | # THE EXTRACTION FUNCTIONS |
134 | |
135 | sub extract_delimited (;$$$$) |
136 | { |
137 | my $textref = defined $_[0] ? \$_[0] : \$_; |
138 | my $wantarray = wantarray; |
139 | my $del = defined $_[1] ? $_[1] : qq{\'\"\`}; |
140 | my $pre = defined $_[2] ? $_[2] : '\s*'; |
141 | my $esc = defined $_[3] ? $_[3] : qq{\\}; |
142 | my $pat = gen_delimited_pat($del, $esc); |
143 | my $startpos = pos $$textref || 0; |
144 | return _fail($wantarray, $textref, "Not a delimited pattern", 0) |
145 | unless $$textref =~ m/\G($pre)($pat)/gc; |
146 | my $prelen = length($1); |
147 | my $matchpos = $startpos+$prelen; |
148 | my $endpos = pos $$textref; |
149 | return _succeed $wantarray, $textref, |
150 | $matchpos, $endpos-$matchpos, # MATCH |
151 | $endpos, length($$textref)-$endpos, # REMAINDER |
152 | $startpos, $prelen; # PREFIX |
153 | } |
154 | |
155 | sub extract_bracketed (;$$$) |
156 | { |
157 | my $textref = defined $_[0] ? \$_[0] : \$_; |
158 | my $ldel = defined $_[1] ? $_[1] : '{([<'; |
159 | my $pre = defined $_[2] ? $_[2] : '\s*'; |
160 | my $wantarray = wantarray; |
161 | my $qdel = ""; |
162 | my $quotelike; |
163 | $ldel =~ s/'//g and $qdel .= q{'}; |
164 | $ldel =~ s/"//g and $qdel .= q{"}; |
165 | $ldel =~ s/`//g and $qdel .= q{`}; |
166 | $ldel =~ s/q//g and $quotelike = 1; |
167 | $ldel =~ tr/[](){}<>\0-\377/[[(({{<</ds; |
168 | my $rdel = $ldel; |
169 | unless ($rdel =~ tr/[({</])}>/) |
170 | { |
171 | return _fail $wantarray, $textref, |
172 | "Did not find a suitable bracket in delimiter: \"$_[1]\"", |
173 | 0; |
174 | } |
175 | my $posbug = pos; |
176 | $ldel = join('|', map { quotemeta $_ } split('', $ldel)); |
177 | $rdel = join('|', map { quotemeta $_ } split('', $rdel)); |
178 | pos = $posbug; |
179 | |
180 | my $startpos = pos $$textref || 0; |
181 | my @match = _match_bracketed($textref,$pre, $ldel, $qdel, $quotelike, $rdel); |
182 | |
183 | return _fail ($wantarray, $textref) unless @match; |
184 | |
185 | return _succeed ( $wantarray, $textref, |
186 | $match[2], $match[5]+2, # MATCH |
187 | @match[8,9], # REMAINDER |
188 | @match[0,1], # PREFIX |
189 | ); |
190 | } |
191 | |
192 | sub _match_bracketed($$$$$$) # $textref, $pre, $ldel, $qdel, $quotelike, $rdel |
193 | { |
194 | my ($textref, $pre, $ldel, $qdel, $quotelike, $rdel) = @_; |
195 | my ($startpos, $ldelpos, $endpos) = (pos $$textref = pos $$textref||0); |
196 | unless ($$textref =~ m/\G$pre/gc) |
197 | { |
198 | _failmsg "Did not find prefix: /$pre/", $startpos; |
199 | return; |
200 | } |
201 | |
202 | $ldelpos = pos $$textref; |
203 | |
204 | unless ($$textref =~ m/\G($ldel)/gc) |
205 | { |
206 | _failmsg "Did not find opening bracket after prefix: \"$pre\"", |
207 | pos $$textref; |
208 | pos $$textref = $startpos; |
209 | return; |
210 | } |
211 | |
212 | my @nesting = ( $1 ); |
213 | my $textlen = length $$textref; |
214 | while (pos $$textref < $textlen) |
215 | { |
216 | next if $$textref =~ m/\G\\./gcs; |
217 | |
218 | if ($$textref =~ m/\G($ldel)/gc) |
219 | { |
220 | push @nesting, $1; |
221 | } |
222 | elsif ($$textref =~ m/\G($rdel)/gc) |
223 | { |
224 | my ($found, $brackettype) = ($1, $1); |
225 | if ($#nesting < 0) |
226 | { |
227 | _failmsg "Unmatched closing bracket: \"$found\"", |
228 | pos $$textref; |
229 | pos $$textref = $startpos; |
230 | return; |
231 | } |
232 | my $expected = pop(@nesting); |
233 | $expected =~ tr/({[</)}]>/; |
234 | if ($expected ne $brackettype) |
235 | { |
236 | _failmsg qq{Mismatched closing bracket: expected "$expected" but found "$found"}, |
237 | pos $$textref; |
238 | pos $$textref = $startpos; |
239 | return; |
240 | } |
241 | last if $#nesting < 0; |
242 | } |
243 | elsif ($qdel && $$textref =~ m/\G([$qdel])/gc) |
244 | { |
9686a75b |
245 | $$textref =~ m/\G[^\\$1]*(?:\\.[^\\$1]*)*(\Q$1\E)/gsc and next; |
3270c621 |
246 | _failmsg "Unmatched embedded quote ($1)", |
247 | pos $$textref; |
248 | pos $$textref = $startpos; |
249 | return; |
250 | } |
251 | elsif ($quotelike && _match_quotelike($textref,"",1,0)) |
252 | { |
253 | next; |
254 | } |
255 | |
256 | else { $$textref =~ m/\G(?:[a-zA-Z0-9]+|.)/gcs } |
257 | } |
258 | if ($#nesting>=0) |
259 | { |
260 | _failmsg "Unmatched opening bracket(s): " |
261 | . join("..",@nesting)."..", |
262 | pos $$textref; |
263 | pos $$textref = $startpos; |
264 | return; |
265 | } |
266 | |
267 | $endpos = pos $$textref; |
268 | |
269 | return ( |
270 | $startpos, $ldelpos-$startpos, # PREFIX |
271 | $ldelpos, 1, # OPENING BRACKET |
272 | $ldelpos+1, $endpos-$ldelpos-2, # CONTENTS |
273 | $endpos-1, 1, # CLOSING BRACKET |
274 | $endpos, length($$textref)-$endpos, # REMAINDER |
275 | ); |
276 | } |
277 | |
278 | sub revbracket($) |
279 | { |
280 | my $brack = reverse $_[0]; |
281 | $brack =~ tr/[({</])}>/; |
282 | return $brack; |
283 | } |
284 | |
285 | my $XMLNAME = q{[a-zA-Z_:][a-zA-Z0-9_:.-]*}; |
286 | |
287 | sub extract_tagged (;$$$$$) # ($text, $opentag, $closetag, $pre, \%options) |
288 | { |
289 | my $textref = defined $_[0] ? \$_[0] : \$_; |
290 | my $ldel = $_[1]; |
291 | my $rdel = $_[2]; |
292 | my $pre = defined $_[3] ? $_[3] : '\s*'; |
293 | my %options = defined $_[4] ? %{$_[4]} : (); |
294 | my $omode = defined $options{fail} ? $options{fail} : ''; |
295 | my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}}) |
296 | : defined($options{reject}) ? $options{reject} |
297 | : '' |
298 | ; |
299 | my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}}) |
300 | : defined($options{ignore}) ? $options{ignore} |
301 | : '' |
302 | ; |
303 | |
304 | if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; } |
305 | $@ = undef; |
306 | |
307 | my @match = _match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore); |
308 | |
309 | return _fail(wantarray, $textref) unless @match; |
310 | return _succeed wantarray, $textref, |
311 | $match[2], $match[3]+$match[5]+$match[7], # MATCH |
312 | @match[8..9,0..1,2..7]; # REM, PRE, BITS |
313 | } |
314 | |
315 | sub _match_tagged # ($$$$$$$) |
316 | { |
317 | my ($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore) = @_; |
318 | my $rdelspec; |
319 | |
320 | my ($startpos, $opentagpos, $textpos, $parapos, $closetagpos, $endpos) = ( pos($$textref) = pos($$textref)||0 ); |
321 | |
322 | unless ($$textref =~ m/\G($pre)/gc) |
323 | { |
324 | _failmsg "Did not find prefix: /$pre/", pos $$textref; |
325 | goto failed; |
326 | } |
327 | |
328 | $opentagpos = pos($$textref); |
329 | |
330 | unless ($$textref =~ m/\G$ldel/gc) |
331 | { |
332 | _failmsg "Did not find opening tag: /$ldel/", pos $$textref; |
333 | goto failed; |
334 | } |
335 | |
336 | $textpos = pos($$textref); |
337 | |
338 | if (!defined $rdel) |
339 | { |
24e08cba |
340 | $rdelspec = &$GetMatchedText($$textref); |
341 | |
3270c621 |
342 | unless ($rdelspec =~ s/\A([[(<{]+)($XMLNAME).*/ quotemeta "$1\/$2". revbracket($1) /oes) |
343 | { |
344 | _failmsg "Unable to construct closing tag to match: $rdel", |
345 | pos $$textref; |
346 | goto failed; |
347 | } |
348 | } |
349 | else |
350 | { |
351 | $rdelspec = eval "qq{$rdel}"; |
352 | } |
353 | |
354 | while (pos($$textref) < length($$textref)) |
355 | { |
356 | next if $$textref =~ m/\G\\./gc; |
357 | |
358 | if ($$textref =~ m/\G(\n[ \t]*\n)/gc ) |
359 | { |
360 | $parapos = pos($$textref) - length($1) |
361 | unless defined $parapos; |
362 | } |
363 | elsif ($$textref =~ m/\G($rdelspec)/gc ) |
364 | { |
365 | $closetagpos = pos($$textref)-length($1); |
366 | goto matched; |
367 | } |
368 | elsif ($ignore && $$textref =~ m/\G(?:$ignore)/gc) |
369 | { |
370 | next; |
371 | } |
372 | elsif ($bad && $$textref =~ m/\G($bad)/gcs) |
373 | { |
374 | pos($$textref) -= length($1); # CUT OFF WHATEVER CAUSED THE SHORTNESS |
375 | goto short if ($omode eq 'PARA' || $omode eq 'MAX'); |
376 | _failmsg "Found invalid nested tag: $1", pos $$textref; |
377 | goto failed; |
378 | } |
379 | elsif ($$textref =~ m/\G($ldel)/gc) |
380 | { |
381 | my $tag = $1; |
382 | pos($$textref) -= length($tag); # REWIND TO NESTED TAG |
383 | unless (_match_tagged(@_)) # MATCH NESTED TAG |
384 | { |
385 | goto short if $omode eq 'PARA' || $omode eq 'MAX'; |
386 | _failmsg "Found unbalanced nested tag: $tag", |
387 | pos $$textref; |
388 | goto failed; |
389 | } |
390 | } |
391 | else { $$textref =~ m/./gcs } |
392 | } |
393 | |
394 | short: |
395 | $closetagpos = pos($$textref); |
396 | goto matched if $omode eq 'MAX'; |
397 | goto failed unless $omode eq 'PARA'; |
398 | |
399 | if (defined $parapos) { pos($$textref) = $parapos } |
400 | else { $parapos = pos($$textref) } |
401 | |
402 | return ( |
403 | $startpos, $opentagpos-$startpos, # PREFIX |
404 | $opentagpos, $textpos-$opentagpos, # OPENING TAG |
405 | $textpos, $parapos-$textpos, # TEXT |
406 | $parapos, 0, # NO CLOSING TAG |
407 | $parapos, length($$textref)-$parapos, # REMAINDER |
408 | ); |
409 | |
410 | matched: |
411 | $endpos = pos($$textref); |
412 | return ( |
413 | $startpos, $opentagpos-$startpos, # PREFIX |
414 | $opentagpos, $textpos-$opentagpos, # OPENING TAG |
415 | $textpos, $closetagpos-$textpos, # TEXT |
416 | $closetagpos, $endpos-$closetagpos, # CLOSING TAG |
417 | $endpos, length($$textref)-$endpos, # REMAINDER |
418 | ); |
419 | |
420 | failed: |
421 | _failmsg "Did not find closing tag", pos $$textref unless $@; |
422 | pos($$textref) = $startpos; |
423 | return; |
424 | } |
425 | |
426 | sub extract_variable (;$$) |
427 | { |
428 | my $textref = defined $_[0] ? \$_[0] : \$_; |
429 | return ("","","") unless defined $$textref; |
430 | my $pre = defined $_[1] ? $_[1] : '\s*'; |
431 | |
432 | my @match = _match_variable($textref,$pre); |
433 | |
434 | return _fail wantarray, $textref unless @match; |
435 | |
436 | return _succeed wantarray, $textref, |
437 | @match[2..3,4..5,0..1]; # MATCH, REMAINDER, PREFIX |
438 | } |
439 | |
440 | sub _match_variable($$) |
441 | { |
a7602084 |
442 | # $# |
443 | # $^ |
444 | # $$ |
3270c621 |
445 | my ($textref, $pre) = @_; |
446 | my $startpos = pos($$textref) = pos($$textref)||0; |
447 | unless ($$textref =~ m/\G($pre)/gc) |
448 | { |
449 | _failmsg "Did not find prefix: /$pre/", pos $$textref; |
450 | return; |
451 | } |
452 | my $varpos = pos($$textref); |
a7602084 |
453 | unless ($$textref =~ m{\G\$\s*(\d+|[][&`'+*./|,";%=~:?!\@<>()-]|\^[a-z]?)}gci) |
3270c621 |
454 | { |
a7602084 |
455 | unless ($$textref =~ m/\G((\$#?|[*\@\%]|\\&)+)/gc) |
456 | { |
3270c621 |
457 | _failmsg "Did not find leading dereferencer", pos $$textref; |
458 | pos $$textref = $startpos; |
459 | return; |
a7602084 |
460 | } |
461 | my $deref = $1; |
3270c621 |
462 | |
a7602084 |
463 | unless ($$textref =~ m/\G\s*(?:::|')?(?:[_a-z]\w*(?:::|'))*[_a-z]\w*/gci |
464 | or _match_codeblock($textref, "", '\{', '\}', '\{', '\}', 0) |
465 | or $deref eq '$#' or $deref eq '$$' ) |
466 | { |
3270c621 |
467 | _failmsg "Bad identifier after dereferencer", pos $$textref; |
468 | pos $$textref = $startpos; |
469 | return; |
a7602084 |
470 | } |
3270c621 |
471 | } |
472 | |
473 | while (1) |
474 | { |
475 | next if _match_codeblock($textref, |
2f250b7c |
476 | qr/\s*->\s*(?:[_a-zA-Z]\w+\s*)?/, |
3270c621 |
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 $$textref =~ m/\G\s*->\s*\w+(?![{([])/gc; |
484 | last; |
485 | } |
486 | |
487 | my $endpos = pos($$textref); |
488 | return ($startpos, $varpos-$startpos, |
489 | $varpos, $endpos-$varpos, |
490 | $endpos, length($$textref)-$endpos |
491 | ); |
492 | } |
493 | |
494 | sub extract_codeblock (;$$$$$) |
495 | { |
496 | my $textref = defined $_[0] ? \$_[0] : \$_; |
497 | my $wantarray = wantarray; |
498 | my $ldel_inner = defined $_[1] ? $_[1] : '{'; |
499 | my $pre = defined $_[2] ? $_[2] : '\s*'; |
500 | my $ldel_outer = defined $_[3] ? $_[3] : $ldel_inner; |
501 | my $rd = $_[4]; |
502 | my $rdel_inner = $ldel_inner; |
503 | my $rdel_outer = $ldel_outer; |
504 | my $posbug = pos; |
505 | for ($ldel_inner, $ldel_outer) { tr/[]()<>{}\0-\377/[[((<<{{/ds } |
506 | for ($rdel_inner, $rdel_outer) { tr/[]()<>{}\0-\377/]]))>>}}/ds } |
507 | for ($ldel_inner, $ldel_outer, $rdel_inner, $rdel_outer) |
508 | { |
509 | $_ = '('.join('|',map { quotemeta $_ } split('',$_)).')' |
510 | } |
511 | pos = $posbug; |
512 | |
513 | my @match = _match_codeblock($textref, $pre, |
514 | $ldel_outer, $rdel_outer, |
515 | $ldel_inner, $rdel_inner, |
516 | $rd); |
517 | return _fail($wantarray, $textref) unless @match; |
518 | return _succeed($wantarray, $textref, |
519 | @match[2..3,4..5,0..1] # MATCH, REMAINDER, PREFIX |
520 | ); |
521 | |
522 | } |
523 | |
524 | sub _match_codeblock($$$$$$$) |
525 | { |
526 | my ($textref, $pre, $ldel_outer, $rdel_outer, $ldel_inner, $rdel_inner, $rd) = @_; |
527 | my $startpos = pos($$textref) = pos($$textref) || 0; |
528 | unless ($$textref =~ m/\G($pre)/gc) |
529 | { |
530 | _failmsg qq{Did not match prefix /$pre/ at"} . |
531 | substr($$textref,pos($$textref),20) . |
532 | q{..."}, |
533 | pos $$textref; |
534 | return; |
535 | } |
536 | my $codepos = pos($$textref); |
537 | unless ($$textref =~ m/\G($ldel_outer)/gc) # OUTERMOST DELIMITER |
538 | { |
539 | _failmsg qq{Did not find expected opening bracket at "} . |
540 | substr($$textref,pos($$textref),20) . |
541 | q{..."}, |
542 | pos $$textref; |
543 | pos $$textref = $startpos; |
544 | return; |
545 | } |
546 | my $closing = $1; |
547 | $closing =~ tr/([<{/)]>}/; |
548 | my $matched; |
549 | my $patvalid = 1; |
550 | while (pos($$textref) < length($$textref)) |
551 | { |
552 | $matched = ''; |
553 | if ($rd && $$textref =~ m#\G(\Q(?)\E|\Q(s?)\E|\Q(s)\E)#gc) |
554 | { |
555 | $patvalid = 0; |
556 | next; |
557 | } |
558 | |
559 | if ($$textref =~ m/\G\s*#.*/gc) |
560 | { |
561 | next; |
562 | } |
563 | |
564 | if ($$textref =~ m/\G\s*($rdel_outer)/gc) |
565 | { |
566 | unless ($matched = ($closing && $1 eq $closing) ) |
567 | { |
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'}, |
572 | pos $$textref; |
573 | } |
574 | last; |
575 | } |
576 | |
577 | if (_match_variable($textref,'\s*') || |
578 | _match_quotelike($textref,'\s*',$patvalid,$patvalid) ) |
579 | { |
580 | $patvalid = 0; |
581 | next; |
582 | } |
583 | |
584 | |
585 | # NEED TO COVER MANY MORE CASES HERE!!! |
586 | if ($$textref =~ m#\G\s*( [-+*x/%^&|.]=? |
55a1c97c |
587 | | [!=]~ |
3270c621 |
588 | | =(?!>) |
589 | | (\*\*|&&|\|\||<<|>>)=? |
3270c621 |
590 | | split|grep|map|return |
591 | )#gcx) |
592 | { |
593 | $patvalid = 1; |
594 | next; |
595 | } |
596 | |
597 | if ( _match_codeblock($textref, '\s*', $ldel_inner, $rdel_inner, $ldel_inner, $rdel_inner, $rd) ) |
598 | { |
599 | $patvalid = 1; |
600 | next; |
601 | } |
602 | |
603 | if ($$textref =~ m/\G\s*$ldel_outer/gc) |
604 | { |
605 | _failmsg q{Improperly nested codeblock at "} . |
606 | substr($$textref,pos($$textref),20) . |
607 | q{..."}, |
608 | pos $$textref; |
609 | last; |
610 | } |
611 | |
612 | $patvalid = 0; |
613 | $$textref =~ m/\G\s*(\w+|[-=>]>|.|\Z)/gc; |
614 | } |
615 | continue { $@ = undef } |
616 | |
617 | unless ($matched) |
618 | { |
619 | _failmsg 'No match found for opening bracket', pos $$textref |
620 | unless $@; |
621 | return; |
622 | } |
623 | |
624 | my $endpos = pos($$textref); |
625 | return ( $startpos, $codepos-$startpos, |
626 | $codepos, $endpos-$codepos, |
627 | $endpos, length($$textref)-$endpos, |
628 | ); |
629 | } |
630 | |
631 | |
632 | my %mods = ( |
633 | 'none' => '[cgimsox]*', |
634 | 'm' => '[cgimsox]*', |
635 | 's' => '[cegimsox]*', |
636 | 'tr' => '[cds]*', |
637 | 'y' => '[cds]*', |
638 | 'qq' => '', |
639 | 'qx' => '', |
640 | 'qw' => '', |
641 | 'qr' => '[imsx]*', |
642 | 'q' => '', |
643 | ); |
644 | |
645 | sub extract_quotelike (;$$) |
646 | { |
647 | my $textref = $_[0] ? \$_[0] : \$_; |
648 | my $wantarray = wantarray; |
649 | my $pre = defined $_[1] ? $_[1] : '\s*'; |
650 | |
651 | my @match = _match_quotelike($textref,$pre,1,0); |
652 | return _fail($wantarray, $textref) unless @match; |
653 | return _succeed($wantarray, $textref, |
654 | $match[2], $match[18]-$match[2], # MATCH |
655 | @match[18,19], # REMAINDER |
656 | @match[0,1], # PREFIX |
657 | @match[2..17], # THE BITS |
658 | @match[20,21], # ANY FILLET? |
659 | ); |
660 | }; |
661 | |
662 | sub _match_quotelike($$$$) # ($textref, $prepat, $allow_raw_match) |
663 | { |
664 | my ($textref, $pre, $rawmatch, $qmark) = @_; |
665 | |
666 | my ($textlen,$startpos, |
667 | $oppos, |
668 | $preld1pos,$ld1pos,$str1pos,$rd1pos, |
669 | $preld2pos,$ld2pos,$str2pos,$rd2pos, |
670 | $modpos) = ( length($$textref), pos($$textref) = pos($$textref) || 0 ); |
671 | |
672 | unless ($$textref =~ m/\G($pre)/gc) |
673 | { |
674 | _failmsg qq{Did not find prefix /$pre/ at "} . |
675 | substr($$textref, pos($$textref), 20) . |
676 | q{..."}, |
677 | pos $$textref; |
678 | return; |
679 | } |
680 | $oppos = pos($$textref); |
681 | |
682 | my $initial = substr($$textref,$oppos,1); |
683 | |
684 | if ($initial && $initial =~ m|^[\"\'\`]| |
685 | || $rawmatch && $initial =~ m|^/| |
686 | || $qmark && $initial =~ m|^\?|) |
687 | { |
9686a75b |
688 | unless ($$textref =~ m/ \Q$initial\E [^\\$initial]* (\\.[^\\$initial]*)* \Q$initial\E /gcsx) |
3270c621 |
689 | { |
690 | _failmsg qq{Did not find closing delimiter to match '$initial' at "} . |
691 | substr($$textref, $oppos, 20) . |
692 | q{..."}, |
693 | pos $$textref; |
694 | pos $$textref = $startpos; |
695 | return; |
696 | } |
697 | $modpos= pos($$textref); |
698 | $rd1pos = $modpos-1; |
699 | |
700 | if ($initial eq '/' || $initial eq '?') |
701 | { |
702 | $$textref =~ m/\G$mods{none}/gc |
703 | } |
704 | |
705 | my $endpos = pos($$textref); |
706 | return ( |
707 | $startpos, $oppos-$startpos, # PREFIX |
708 | $oppos, 0, # NO OPERATOR |
709 | $oppos, 1, # LEFT DEL |
710 | $oppos+1, $rd1pos-$oppos-1, # STR/PAT |
711 | $rd1pos, 1, # RIGHT DEL |
712 | $modpos, 0, # NO 2ND LDEL |
713 | $modpos, 0, # NO 2ND STR |
714 | $modpos, 0, # NO 2ND RDEL |
715 | $modpos, $endpos-$modpos, # MODIFIERS |
716 | $endpos, $textlen-$endpos, # REMAINDER |
717 | ); |
718 | } |
719 | |
720 | unless ($$textref =~ m{\G((?:m|s|qq|qx|qw|q|qr|tr|y)\b(?=\s*\S)|<<)}gc) |
721 | { |
722 | _failmsg q{No quotelike operator found after prefix at "} . |
723 | substr($$textref, pos($$textref), 20) . |
724 | q{..."}, |
725 | pos $$textref; |
726 | pos $$textref = $startpos; |
727 | return; |
728 | } |
729 | |
730 | my $op = $1; |
731 | $preld1pos = pos($$textref); |
732 | if ($op eq '<<') { |
733 | $ld1pos = pos($$textref); |
734 | my $label; |
735 | if ($$textref =~ m{\G([A-Za-z_]\w*)}gc) { |
736 | $label = $1; |
737 | } |
738 | elsif ($$textref =~ m{ \G ' ([^'\\]* (?:\\.[^'\\]*)*) ' |
739 | | \G " ([^"\\]* (?:\\.[^"\\]*)*) " |
740 | | \G ` ([^`\\]* (?:\\.[^`\\]*)*) ` |
9686a75b |
741 | }gcsx) { |
3270c621 |
742 | $label = $+; |
743 | } |
744 | else { |
745 | $label = ""; |
746 | } |
747 | my $extrapos = pos($$textref); |
748 | $$textref =~ m{.*\n}gc; |
749 | $str1pos = pos($$textref); |
750 | unless ($$textref =~ m{.*?\n(?=$label\n)}gc) { |
751 | _failmsg qq{Missing here doc terminator ('$label') after "} . |
752 | substr($$textref, $startpos, 20) . |
753 | q{..."}, |
754 | pos $$textref; |
755 | pos $$textref = $startpos; |
756 | return; |
757 | } |
758 | $rd1pos = pos($$textref); |
759 | $$textref =~ m{$label\n}gc; |
760 | $ld2pos = pos($$textref); |
761 | return ( |
762 | $startpos, $oppos-$startpos, # PREFIX |
763 | $oppos, length($op), # OPERATOR |
764 | $ld1pos, $extrapos-$ld1pos, # LEFT DEL |
765 | $str1pos, $rd1pos-$str1pos, # STR/PAT |
766 | $rd1pos, $ld2pos-$rd1pos, # RIGHT DEL |
767 | $ld2pos, 0, # NO 2ND LDEL |
768 | $ld2pos, 0, # NO 2ND STR |
769 | $ld2pos, 0, # NO 2ND RDEL |
770 | $ld2pos, 0, # NO MODIFIERS |
771 | $ld2pos, $textlen-$ld2pos, # REMAINDER |
772 | $extrapos, $str1pos-$extrapos, # FILLETED BIT |
773 | ); |
774 | } |
775 | |
776 | $$textref =~ m/\G\s*/gc; |
777 | $ld1pos = pos($$textref); |
778 | $str1pos = $ld1pos+1; |
779 | |
780 | unless ($$textref =~ m/\G(\S)/gc) # SHOULD USE LOOKAHEAD |
781 | { |
782 | _failmsg "No block delimiter found after quotelike $op", |
783 | pos $$textref; |
784 | pos $$textref = $startpos; |
785 | return; |
786 | } |
787 | pos($$textref) = $ld1pos; # HAVE TO DO THIS BECAUSE LOOKAHEAD BROKEN |
788 | my ($ldel1, $rdel1) = ("\Q$1","\Q$1"); |
789 | if ($ldel1 =~ /[[(<{]/) |
790 | { |
791 | $rdel1 =~ tr/[({</])}>/; |
792 | _match_bracketed($textref,"",$ldel1,"","",$rdel1) |
793 | || do { pos $$textref = $startpos; return }; |
794 | } |
795 | else |
796 | { |
9686a75b |
797 | $$textref =~ /$ldel1[^\\$ldel1]*(\\.[^\\$ldel1]*)*$ldel1/gcs |
3270c621 |
798 | || do { pos $$textref = $startpos; return }; |
799 | } |
800 | $ld2pos = $rd1pos = pos($$textref)-1; |
801 | |
802 | my $second_arg = $op =~ /s|tr|y/ ? 1 : 0; |
803 | if ($second_arg) |
804 | { |
805 | my ($ldel2, $rdel2); |
806 | if ($ldel1 =~ /[[(<{]/) |
807 | { |
808 | unless ($$textref =~ /\G\s*(\S)/gc) # SHOULD USE LOOKAHEAD |
809 | { |
810 | _failmsg "Missing second block for quotelike $op", |
811 | pos $$textref; |
812 | pos $$textref = $startpos; |
813 | return; |
814 | } |
815 | $ldel2 = $rdel2 = "\Q$1"; |
816 | $rdel2 =~ tr/[({</])}>/; |
817 | } |
818 | else |
819 | { |
820 | $ldel2 = $rdel2 = $ldel1; |
821 | } |
822 | $str2pos = $ld2pos+1; |
823 | |
824 | if ($ldel2 =~ /[[(<{]/) |
825 | { |
826 | pos($$textref)--; # OVERCOME BROKEN LOOKAHEAD |
827 | _match_bracketed($textref,"",$ldel2,"","",$rdel2) |
828 | || do { pos $$textref = $startpos; return }; |
829 | } |
830 | else |
831 | { |
9686a75b |
832 | $$textref =~ /[^\\$ldel2]*(\\.[^\\$ldel2]*)*$ldel2/gcs |
3270c621 |
833 | || do { pos $$textref = $startpos; return }; |
834 | } |
835 | $rd2pos = pos($$textref)-1; |
836 | } |
837 | else |
838 | { |
839 | $ld2pos = $str2pos = $rd2pos = $rd1pos; |
840 | } |
841 | |
842 | $modpos = pos $$textref; |
843 | |
844 | $$textref =~ m/\G($mods{$op})/gc; |
845 | my $endpos = pos $$textref; |
846 | |
847 | return ( |
848 | $startpos, $oppos-$startpos, # PREFIX |
849 | $oppos, length($op), # OPERATOR |
850 | $ld1pos, 1, # LEFT DEL |
851 | $str1pos, $rd1pos-$str1pos, # STR/PAT |
852 | $rd1pos, 1, # RIGHT DEL |
853 | $ld2pos, $second_arg, # 2ND LDEL (MAYBE) |
854 | $str2pos, $rd2pos-$str2pos, # 2ND STR (MAYBE) |
855 | $rd2pos, $second_arg, # 2ND RDEL (MAYBE) |
856 | $modpos, $endpos-$modpos, # MODIFIERS |
857 | $endpos, $textlen-$endpos, # REMAINDER |
858 | ); |
859 | } |
860 | |
861 | my $def_func = |
862 | [ |
863 | sub { extract_variable($_[0], '') }, |
864 | sub { extract_quotelike($_[0],'') }, |
865 | sub { extract_codeblock($_[0],'{}','') }, |
866 | ]; |
867 | |
868 | sub extract_multiple (;$$$$) # ($text, $functions_ref, $max_fields, $ignoreunknown) |
869 | { |
870 | my $textref = defined($_[0]) ? \$_[0] : \$_; |
871 | my $posbug = pos; |
872 | my ($lastpos, $firstpos); |
873 | my @fields = (); |
874 | |
a7602084 |
875 | #for ($$textref) |
3270c621 |
876 | { |
877 | my @func = defined $_[1] ? @{$_[1]} : @{$def_func}; |
878 | my $max = defined $_[2] && $_[2]>0 ? $_[2] : 1_000_000_000; |
879 | my $igunk = $_[3]; |
880 | |
a7602084 |
881 | pos $$textref ||= 0; |
3270c621 |
882 | |
883 | unless (wantarray) |
884 | { |
885 | use Carp; |
886 | carp "extract_multiple reset maximal count to 1 in scalar context" |
887 | if $^W && defined($_[2]) && $max > 1; |
888 | $max = 1 |
889 | } |
890 | |
891 | my $unkpos; |
892 | my $func; |
893 | my $class; |
894 | |
895 | my @class; |
896 | foreach $func ( @func ) |
897 | { |
898 | if (ref($func) eq 'HASH') |
899 | { |
900 | push @class, (keys %$func)[0]; |
901 | $func = (values %$func)[0]; |
902 | } |
903 | else |
904 | { |
905 | push @class, undef; |
906 | } |
907 | } |
908 | |
a7602084 |
909 | FIELD: while (pos($$textref) < length($$textref)) |
3270c621 |
910 | { |
911 | my $field; |
a7602084 |
912 | my @bits; |
3270c621 |
913 | foreach my $i ( 0..$#func ) |
914 | { |
a7602084 |
915 | my $pref; |
3270c621 |
916 | $func = $func[$i]; |
917 | $class = $class[$i]; |
a7602084 |
918 | $lastpos = pos $$textref; |
3270c621 |
919 | if (ref($func) eq 'CODE') |
a7602084 |
920 | { ($field,undef,$pref) = @bits = $func->($$textref) } |
3270c621 |
921 | elsif (ref($func) eq 'Text::Balanced::Extractor') |
a7602084 |
922 | { @bits = $field = $func->extract($$textref) } |
923 | elsif( $$textref =~ m/\G$func/gc ) |
24e08cba |
924 | { @bits = $field = defined($1) ? $1 : &$GetMatchedText($$textref) } |
925 | # substr() on previous line is "$&", without the pain |
a7602084 |
926 | $pref ||= ""; |
3270c621 |
927 | if (defined($field) && length($field)) |
928 | { |
a7602084 |
929 | if (!$igunk) { |
930 | $unkpos = pos $$textref |
931 | if length($pref) && !defined($unkpos); |
932 | if (defined $unkpos) |
933 | { |
934 | push @fields, substr($$textref, $unkpos, $lastpos-$unkpos).$pref; |
935 | $firstpos = $unkpos unless defined $firstpos; |
936 | undef $unkpos; |
937 | last FIELD if @fields == $max; |
938 | } |
3270c621 |
939 | } |
a7602084 |
940 | push @fields, $class |
941 | ? bless (\$field, $class) |
3270c621 |
942 | : $field; |
943 | $firstpos = $lastpos unless defined $firstpos; |
a7602084 |
944 | $lastpos = pos $$textref; |
3270c621 |
945 | last FIELD if @fields == $max; |
946 | next FIELD; |
947 | } |
948 | } |
a7602084 |
949 | if ($$textref =~ /\G(.)/gcs) |
3270c621 |
950 | { |
a7602084 |
951 | $unkpos = pos($$textref)-1 |
3270c621 |
952 | unless $igunk || defined $unkpos; |
953 | } |
954 | } |
955 | |
956 | if (defined $unkpos) |
957 | { |
a7602084 |
958 | push @fields, substr($$textref, $unkpos); |
3270c621 |
959 | $firstpos = $unkpos unless defined $firstpos; |
a7602084 |
960 | $lastpos = length $$textref; |
3270c621 |
961 | } |
962 | last; |
963 | } |
964 | |
965 | pos $$textref = $lastpos; |
966 | return @fields if wantarray; |
967 | |
968 | $firstpos ||= 0; |
969 | eval { substr($$textref,$firstpos,$lastpos-$firstpos)=""; |
970 | pos $$textref = $firstpos }; |
971 | return $fields[0]; |
972 | } |
973 | |
974 | |
975 | sub gen_extract_tagged # ($opentag, $closetag, $pre, \%options) |
976 | { |
977 | my $ldel = $_[0]; |
978 | my $rdel = $_[1]; |
979 | my $pre = defined $_[2] ? $_[2] : '\s*'; |
980 | my %options = defined $_[3] ? %{$_[3]} : (); |
981 | my $omode = defined $options{fail} ? $options{fail} : ''; |
982 | my $bad = ref($options{reject}) eq 'ARRAY' ? join('|', @{$options{reject}}) |
983 | : defined($options{reject}) ? $options{reject} |
984 | : '' |
985 | ; |
986 | my $ignore = ref($options{ignore}) eq 'ARRAY' ? join('|', @{$options{ignore}}) |
987 | : defined($options{ignore}) ? $options{ignore} |
988 | : '' |
989 | ; |
990 | |
991 | if (!defined $ldel) { $ldel = '<\w+(?:' . gen_delimited_pat(q{'"}) . '|[^>])*>'; } |
992 | |
993 | my $posbug = pos; |
994 | for ($ldel, $pre, $bad, $ignore) { $_ = qr/$_/ if $_ } |
995 | pos = $posbug; |
996 | |
997 | my $closure = sub |
998 | { |
999 | my $textref = defined $_[0] ? \$_[0] : \$_; |
1000 | my @match = Text::Balanced::_match_tagged($textref, $pre, $ldel, $rdel, $omode, $bad, $ignore); |
1001 | |
1002 | return _fail(wantarray, $textref) unless @match; |
1003 | return _succeed wantarray, $textref, |
1004 | $match[2], $match[3]+$match[5]+$match[7], # MATCH |
1005 | @match[8..9,0..1,2..7]; # REM, PRE, BITS |
1006 | }; |
1007 | |
1008 | bless $closure, 'Text::Balanced::Extractor'; |
1009 | } |
1010 | |
1011 | package Text::Balanced::Extractor; |
1012 | |
1013 | sub extract($$) # ($self, $text) |
1014 | { |
1015 | &{$_[0]}($_[1]); |
1016 | } |
1017 | |
1018 | package Text::Balanced::ErrorMsg; |
1019 | |
1020 | use overload '""' => sub { "$_[0]->{error}, detected at offset $_[0]->{pos}" }; |
1021 | |
1022 | 1; |
55a1c97c |
1023 | |
1024 | __END__ |
1025 | |
1026 | =head1 NAME |
1027 | |
1028 | Text::Balanced - Extract delimited text sequences from strings. |
1029 | |
1030 | |
1031 | =head1 SYNOPSIS |
1032 | |
1033 | use Text::Balanced qw ( |
1034 | extract_delimited |
1035 | extract_bracketed |
1036 | extract_quotelike |
1037 | extract_codeblock |
1038 | extract_variable |
1039 | extract_tagged |
1040 | extract_multiple |
1041 | |
1042 | gen_delimited_pat |
1043 | gen_extract_tagged |
1044 | ); |
1045 | |
1046 | # Extract the initial substring of $text that is delimited by |
1047 | # two (unescaped) instances of the first character in $delim. |
1048 | |
1049 | ($extracted, $remainder) = extract_delimited($text,$delim); |
1050 | |
1051 | |
1052 | # Extract the initial substring of $text that is bracketed |
1053 | # with a delimiter(s) specified by $delim (where the string |
1054 | # in $delim contains one or more of '(){}[]<>'). |
1055 | |
1056 | ($extracted, $remainder) = extract_bracketed($text,$delim); |
1057 | |
1058 | |
1059 | # Extract the initial substring of $text that is bounded by |
1060 | # an HTML/XML tag. |
1061 | |
1062 | ($extracted, $remainder) = extract_tagged($text); |
1063 | |
1064 | |
1065 | # Extract the initial substring of $text that is bounded by |
1066 | # a C<BEGIN>...C<END> pair. Don't allow nested C<BEGIN> tags |
1067 | |
1068 | ($extracted, $remainder) = |
1069 | extract_tagged($text,"BEGIN","END",undef,{bad=>["BEGIN"]}); |
1070 | |
1071 | |
1072 | # Extract the initial substring of $text that represents a |
1073 | # Perl "quote or quote-like operation" |
1074 | |
1075 | ($extracted, $remainder) = extract_quotelike($text); |
1076 | |
1077 | |
1078 | # Extract the initial substring of $text that represents a block |
1079 | # of Perl code, bracketed by any of character(s) specified by $delim |
1080 | # (where the string $delim contains one or more of '(){}[]<>'). |
1081 | |
1082 | ($extracted, $remainder) = extract_codeblock($text,$delim); |
1083 | |
1084 | |
1085 | # Extract the initial substrings of $text that would be extracted by |
1086 | # one or more sequential applications of the specified functions |
1087 | # or regular expressions |
1088 | |
1089 | @extracted = extract_multiple($text, |
1090 | [ \&extract_bracketed, |
1091 | \&extract_quotelike, |
1092 | \&some_other_extractor_sub, |
1093 | qr/[xyz]*/, |
1094 | 'literal', |
1095 | ]); |
1096 | |
1097 | # Create a string representing an optimized pattern (a la Friedl) |
1098 | # that matches a substring delimited by any of the specified characters |
1099 | # (in this case: any type of quote or a slash) |
1100 | |
1101 | $patstring = gen_delimited_pat(q{'"`/}); |
1102 | |
1103 | |
1104 | # Generate a reference to an anonymous sub that is just like extract_tagged |
1105 | # but pre-compiled and optimized for a specific pair of tags, and consequently |
1106 | # much faster (i.e. 3 times faster). It uses qr// for better performance on |
1107 | # repeated calls, so it only works under Perl 5.005 or later. |
1108 | |
1109 | $extract_head = gen_extract_tagged('<HEAD>','</HEAD>'); |
1110 | |
1111 | ($extracted, $remainder) = $extract_head->($text); |
1112 | |
1113 | |
1114 | =head1 DESCRIPTION |
1115 | |
1116 | The various C<extract_...> subroutines may be used to extract a |
1117 | delimited string (possibly after skipping a specified prefix string). |
1118 | The search for the string always begins at the current C<pos> |
1119 | location of the string's variable (or at index zero, if no C<pos> |
1120 | position is defined). |
1121 | |
1122 | =head2 General behaviour in list contexts |
1123 | |
1124 | In a list context, all the subroutines return a list, the first three |
1125 | elements of which are always: |
1126 | |
1127 | =over 4 |
1128 | |
1129 | =item [0] |
1130 | |
1131 | The extracted string, including the specified delimiters. |
1132 | If the extraction fails an empty string is returned. |
1133 | |
1134 | =item [1] |
1135 | |
1136 | The remainder of the input string (i.e. the characters after the |
1137 | extracted string). On failure, the entire string is returned. |
1138 | |
1139 | =item [2] |
1140 | |
1141 | The skipped prefix (i.e. the characters before the extracted string). |
1142 | On failure, the empty string is returned. |
1143 | |
1144 | =back |
1145 | |
1146 | Note that in a list context, the contents of the original input text (the first |
1147 | argument) are not modified in any way. |
1148 | |
1149 | However, if the input text was passed in a variable, that variable's |
1150 | C<pos> value is updated to point at the first character after the |
1151 | extracted text. That means that in a list context the various |
1152 | subroutines can be used much like regular expressions. For example: |
1153 | |
1154 | while ( $next = (extract_quotelike($text))[0] ) |
1155 | { |
1156 | # process next quote-like (in $next) |
1157 | } |
1158 | |
1159 | |
1160 | =head2 General behaviour in scalar and void contexts |
1161 | |
1162 | In a scalar context, the extracted string is returned, having first been |
1163 | removed from the input text. Thus, the following code also processes |
1164 | each quote-like operation, but actually removes them from $text: |
1165 | |
1166 | while ( $next = extract_quotelike($text) ) |
1167 | { |
1168 | # process next quote-like (in $next) |
1169 | } |
1170 | |
1171 | Note that if the input text is a read-only string (i.e. a literal), |
1172 | no attempt is made to remove the extracted text. |
1173 | |
1174 | In a void context the behaviour of the extraction subroutines is |
1175 | exactly the same as in a scalar context, except (of course) that the |
1176 | extracted substring is not returned. |
1177 | |
1178 | =head2 A note about prefixes |
1179 | |
1180 | Prefix patterns are matched without any trailing modifiers (C</gimsox> etc.) |
1181 | This can bite you if you're expecting a prefix specification like |
1182 | '.*?(?=<H1>)' to skip everything up to the first <H1> tag. Such a prefix |
1183 | pattern will only succeed if the <H1> tag is on the current line, since |
1184 | . normally doesn't match newlines. |
1185 | |
1186 | To overcome this limitation, you need to turn on /s matching within |
1187 | the prefix pattern, using the C<(?s)> directive: '(?s).*?(?=<H1>)' |
1188 | |
1189 | |
1190 | =head2 C<extract_delimited> |
1191 | |
1192 | The C<extract_delimited> function formalizes the common idiom |
1193 | of extracting a single-character-delimited substring from the start of |
1194 | a string. For example, to extract a single-quote delimited string, the |
1195 | following code is typically used: |
1196 | |
1197 | ($remainder = $text) =~ s/\A('(\\.|[^'])*')//s; |
1198 | $extracted = $1; |
1199 | |
1200 | but with C<extract_delimited> it can be simplified to: |
1201 | |
1202 | ($extracted,$remainder) = extract_delimited($text, "'"); |
1203 | |
1204 | C<extract_delimited> takes up to four scalars (the input text, the |
1205 | delimiters, a prefix pattern to be skipped, and any escape characters) |
1206 | and extracts the initial substring of the text that |
1207 | is appropriately delimited. If the delimiter string has multiple |
1208 | characters, the first one encountered in the text is taken to delimit |
1209 | the substring. |
1210 | The third argument specifies a prefix pattern that is to be skipped |
1211 | (but must be present!) before the substring is extracted. |
1212 | The final argument specifies the escape character to be used for each |
1213 | delimiter. |
1214 | |
1215 | All arguments are optional. If the escape characters are not specified, |
1216 | every delimiter is escaped with a backslash (C<\>). |
1217 | If the prefix is not specified, the |
1218 | pattern C<'\s*'> - optional whitespace - is used. If the delimiter set |
1219 | is also not specified, the set C</["'`]/> is used. If the text to be processed |
1220 | is not specified either, C<$_> is used. |
1221 | |
d1be9408 |
1222 | In list context, C<extract_delimited> returns an array of three |
55a1c97c |
1223 | elements, the extracted substring (I<including the surrounding |
1224 | delimiters>), the remainder of the text, and the skipped prefix (if |
1225 | any). If a suitable delimited substring is not found, the first |
1226 | element of the array is the empty string, the second is the complete |
1227 | original text, and the prefix returned in the third element is an |
1228 | empty string. |
1229 | |
1230 | In a scalar context, just the extracted substring is returned. In |
1231 | a void context, the extracted substring (and any prefix) are simply |
1232 | removed from the beginning of the first argument. |
1233 | |
1234 | Examples: |
1235 | |
1236 | # Remove a single-quoted substring from the very beginning of $text: |
1237 | |
1238 | $substring = extract_delimited($text, "'", ''); |
1239 | |
1240 | # Remove a single-quoted Pascalish substring (i.e. one in which |
1241 | # doubling the quote character escapes it) from the very |
1242 | # beginning of $text: |
1243 | |
1244 | $substring = extract_delimited($text, "'", '', "'"); |
1245 | |
1246 | # Extract a single- or double- quoted substring from the |
1247 | # beginning of $text, optionally after some whitespace |
1248 | # (note the list context to protect $text from modification): |
1249 | |
1250 | ($substring) = extract_delimited $text, q{"'}; |
1251 | |
1252 | |
1253 | # Delete the substring delimited by the first '/' in $text: |
1254 | |
1255 | $text = join '', (extract_delimited($text,'/','[^/]*')[2,1]; |
1256 | |
1257 | Note that this last example is I<not> the same as deleting the first |
1258 | quote-like pattern. For instance, if C<$text> contained the string: |
1259 | |
1260 | "if ('./cmd' =~ m/$UNIXCMD/s) { $cmd = $1; }" |
1261 | |
1262 | then after the deletion it would contain: |
1263 | |
1264 | "if ('.$UNIXCMD/s) { $cmd = $1; }" |
1265 | |
1266 | not: |
1267 | |
1268 | "if ('./cmd' =~ ms) { $cmd = $1; }" |
1269 | |
1270 | |
1271 | See L<"extract_quotelike"> for a (partial) solution to this problem. |
1272 | |
1273 | |
1274 | =head2 C<extract_bracketed> |
1275 | |
1276 | Like C<"extract_delimited">, the C<extract_bracketed> function takes |
1277 | up to three optional scalar arguments: a string to extract from, a delimiter |
1278 | specifier, and a prefix pattern. As before, a missing prefix defaults to |
1279 | optional whitespace and a missing text defaults to C<$_>. However, a missing |
1280 | delimiter specifier defaults to C<'{}()[]E<lt>E<gt>'> (see below). |
1281 | |
1282 | C<extract_bracketed> extracts a balanced-bracket-delimited |
1283 | substring (using any one (or more) of the user-specified delimiter |
1284 | brackets: '(..)', '{..}', '[..]', or '<..>'). Optionally it will also |
1285 | respect quoted unbalanced brackets (see below). |
1286 | |
1287 | A "delimiter bracket" is a bracket in list of delimiters passed as |
1288 | C<extract_bracketed>'s second argument. Delimiter brackets are |
1289 | specified by giving either the left or right (or both!) versions |
1290 | of the required bracket(s). Note that the order in which |
1291 | two or more delimiter brackets are specified is not significant. |
1292 | |
1293 | A "balanced-bracket-delimited substring" is a substring bounded by |
1294 | matched brackets, such that any other (left or right) delimiter |
1295 | bracket I<within> the substring is also matched by an opposite |
1296 | (right or left) delimiter bracket I<at the same level of nesting>. Any |
1297 | type of bracket not in the delimiter list is treated as an ordinary |
1298 | character. |
1299 | |
1300 | In other words, each type of bracket specified as a delimiter must be |
1301 | balanced and correctly nested within the substring, and any other kind of |
1302 | ("non-delimiter") bracket in the substring is ignored. |
1303 | |
1304 | For example, given the string: |
1305 | |
1306 | $text = "{ an '[irregularly :-(] {} parenthesized >:-)' string }"; |
1307 | |
1308 | then a call to C<extract_bracketed> in a list context: |
1309 | |
1310 | @result = extract_bracketed( $text, '{}' ); |
1311 | |
1312 | would return: |
1313 | |
1314 | ( "{ an '[irregularly :-(] {} parenthesized >:-)' string }" , "" , "" ) |
1315 | |
1316 | since both sets of C<'{..}'> brackets are properly nested and evenly balanced. |
1317 | (In a scalar context just the first element of the array would be returned. In |
1318 | a void context, C<$text> would be replaced by an empty string.) |
1319 | |
1320 | Likewise the call in: |
1321 | |
1322 | @result = extract_bracketed( $text, '{[' ); |
1323 | |
1324 | would return the same result, since all sets of both types of specified |
1325 | delimiter brackets are correctly nested and balanced. |
1326 | |
1327 | However, the call in: |
1328 | |
1329 | @result = extract_bracketed( $text, '{([<' ); |
1330 | |
1331 | would fail, returning: |
1332 | |
1333 | ( undef , "{ an '[irregularly :-(] {} parenthesized >:-)' string }" ); |
1334 | |
1335 | because the embedded pairs of C<'(..)'>s and C<'[..]'>s are "cross-nested" and |
1336 | the embedded C<'E<gt>'> is unbalanced. (In a scalar context, this call would |
1337 | return an empty string. In a void context, C<$text> would be unchanged.) |
1338 | |
1339 | Note that the embedded single-quotes in the string don't help in this |
1340 | case, since they have not been specified as acceptable delimiters and are |
1341 | therefore treated as non-delimiter characters (and ignored). |
1342 | |
1343 | However, if a particular species of quote character is included in the |
1344 | delimiter specification, then that type of quote will be correctly handled. |
1345 | for example, if C<$text> is: |
1346 | |
1347 | $text = '<A HREF=">>>>">link</A>'; |
1348 | |
1349 | then |
1350 | |
1351 | @result = extract_bracketed( $text, '<">' ); |
1352 | |
1353 | returns: |
1354 | |
1355 | ( '<A HREF=">>>>">', 'link</A>', "" ) |
1356 | |
1357 | as expected. Without the specification of C<"> as an embedded quoter: |
1358 | |
1359 | @result = extract_bracketed( $text, '<>' ); |
1360 | |
1361 | the result would be: |
1362 | |
1363 | ( '<A HREF=">', '>>>">link</A>', "" ) |
1364 | |
1365 | In addition to the quote delimiters C<'>, C<">, and C<`>, full Perl quote-like |
1366 | quoting (i.e. q{string}, qq{string}, etc) can be specified by including the |
1367 | letter 'q' as a delimiter. Hence: |
1368 | |
1369 | @result = extract_bracketed( $text, '<q>' ); |
1370 | |
1371 | would correctly match something like this: |
1372 | |
1373 | $text = '<leftop: conj /and/ conj>'; |
1374 | |
1375 | See also: C<"extract_quotelike"> and C<"extract_codeblock">. |
1376 | |
1377 | |
1378 | =head2 C<extract_tagged> |
1379 | |
1380 | C<extract_tagged> extracts and segments text between (balanced) |
1381 | specified tags. |
1382 | |
1383 | The subroutine takes up to five optional arguments: |
1384 | |
1385 | =over 4 |
1386 | |
1387 | =item 1. |
1388 | |
1389 | A string to be processed (C<$_> if the string is omitted or C<undef>) |
1390 | |
1391 | =item 2. |
1392 | |
1393 | A string specifying a pattern to be matched as the opening tag. |
1394 | If the pattern string is omitted (or C<undef>) then a pattern |
1395 | that matches any standard HTML/XML tag is used. |
1396 | |
1397 | =item 3. |
1398 | |
1399 | A string specifying a pattern to be matched at the closing tag. |
1400 | If the pattern string is omitted (or C<undef>) then the closing |
1401 | tag is constructed by inserting a C</> after any leading bracket |
1402 | characters in the actual opening tag that was matched (I<not> the pattern |
1403 | that matched the tag). For example, if the opening tag pattern |
1404 | is specified as C<'{{\w+}}'> and actually matched the opening tag |
1405 | C<"{{DATA}}">, then the constructed closing tag would be C<"{{/DATA}}">. |
1406 | |
1407 | =item 4. |
1408 | |
1409 | A string specifying a pattern to be matched as a prefix (which is to be |
1410 | skipped). If omitted, optional whitespace is skipped. |
1411 | |
1412 | =item 5. |
1413 | |
1414 | A hash reference containing various parsing options (see below) |
1415 | |
1416 | =back |
1417 | |
1418 | The various options that can be specified are: |
1419 | |
1420 | =over 4 |
1421 | |
1422 | =item C<reject =E<gt> $listref> |
1423 | |
1424 | The list reference contains one or more strings specifying patterns |
1425 | that must I<not> appear within the tagged text. |
1426 | |
1427 | For example, to extract |
1428 | an HTML link (which should not contain nested links) use: |
1429 | |
1430 | extract_tagged($text, '<A>', '</A>', undef, {reject => ['<A>']} ); |
1431 | |
1432 | =item C<ignore =E<gt> $listref> |
1433 | |
1434 | The list reference contains one or more strings specifying patterns |
1435 | that are I<not> be be treated as nested tags within the tagged text |
1436 | (even if they would match the start tag pattern). |
1437 | |
1438 | For example, to extract an arbitrary XML tag, but ignore "empty" elements: |
1439 | |
1440 | extract_tagged($text, undef, undef, undef, {ignore => ['<[^>]*/>']} ); |
1441 | |
1442 | (also see L<"gen_delimited_pat"> below). |
1443 | |
1444 | |
1445 | =item C<fail =E<gt> $str> |
1446 | |
1447 | The C<fail> option indicates the action to be taken if a matching end |
1448 | tag is not encountered (i.e. before the end of the string or some |
1449 | C<reject> pattern matches). By default, a failure to match a closing |
1450 | tag causes C<extract_tagged> to immediately fail. |
1451 | |
1452 | However, if the string value associated with <reject> is "MAX", then |
1453 | C<extract_tagged> returns the complete text up to the point of failure. |
1454 | If the string is "PARA", C<extract_tagged> returns only the first paragraph |
1455 | after the tag (up to the first line that is either empty or contains |
1456 | only whitespace characters). |
d1be9408 |
1457 | If the string is "", the default behaviour (i.e. failure) is reinstated. |
55a1c97c |
1458 | |
1459 | For example, suppose the start tag "/para" introduces a paragraph, which then |
1460 | continues until the next "/endpara" tag or until another "/para" tag is |
1461 | encountered: |
1462 | |
1463 | $text = "/para line 1\n\nline 3\n/para line 4"; |
1464 | |
1465 | extract_tagged($text, '/para', '/endpara', undef, |
1466 | {reject => '/para', fail => MAX ); |
1467 | |
1468 | # EXTRACTED: "/para line 1\n\nline 3\n" |
1469 | |
1470 | Suppose instead, that if no matching "/endpara" tag is found, the "/para" |
1471 | tag refers only to the immediately following paragraph: |
1472 | |
1473 | $text = "/para line 1\n\nline 3\n/para line 4"; |
1474 | |
1475 | extract_tagged($text, '/para', '/endpara', undef, |
1476 | {reject => '/para', fail => MAX ); |
1477 | |
1478 | # EXTRACTED: "/para line 1\n" |
1479 | |
1480 | Note that the specified C<fail> behaviour applies to nested tags as well. |
1481 | |
1482 | =back |
1483 | |
1484 | On success in a list context, an array of 6 elements is returned. The elements are: |
1485 | |
1486 | =over 4 |
1487 | |
1488 | =item [0] |
1489 | |
1490 | the extracted tagged substring (including the outermost tags), |
1491 | |
1492 | =item [1] |
1493 | |
1494 | the remainder of the input text, |
1495 | |
1496 | =item [2] |
1497 | |
1498 | the prefix substring (if any), |
1499 | |
1500 | =item [3] |
1501 | |
1502 | the opening tag |
1503 | |
1504 | =item [4] |
1505 | |
1506 | the text between the opening and closing tags |
1507 | |
1508 | =item [5] |
1509 | |
1510 | the closing tag (or "" if no closing tag was found) |
1511 | |
1512 | =back |
1513 | |
1514 | On failure, all of these values (except the remaining text) are C<undef>. |
1515 | |
1516 | In a scalar context, C<extract_tagged> returns just the complete |
1517 | substring that matched a tagged text (including the start and end |
1518 | tags). C<undef> is returned on failure. In addition, the original input |
1519 | text has the returned substring (and any prefix) removed from it. |
1520 | |
1521 | In a void context, the input text just has the matched substring (and |
1522 | any specified prefix) removed. |
1523 | |
1524 | |
1525 | =head2 C<gen_extract_tagged> |
1526 | |
1527 | (Note: This subroutine is only available under Perl5.005) |
1528 | |
1529 | C<gen_extract_tagged> generates a new anonymous subroutine which |
1530 | extracts text between (balanced) specified tags. In other words, |
1531 | it generates a function identical in function to C<extract_tagged>. |
1532 | |
1533 | The difference between C<extract_tagged> and the anonymous |
1534 | subroutines generated by |
1535 | C<gen_extract_tagged>, is that those generated subroutines: |
1536 | |
1537 | =over 4 |
1538 | |
1539 | =item * |
1540 | |
1541 | do not have to reparse tag specification or parsing options every time |
1542 | they are called (whereas C<extract_tagged> has to effectively rebuild |
1543 | its tag parser on every call); |
1544 | |
1545 | =item * |
1546 | |
1547 | make use of the new qr// construct to pre-compile the regexes they use |
1548 | (whereas C<extract_tagged> uses standard string variable interpolation |
1549 | to create tag-matching patterns). |
1550 | |
1551 | =back |
1552 | |
1553 | The subroutine takes up to four optional arguments (the same set as |
1554 | C<extract_tagged> except for the string to be processed). It returns |
1555 | a reference to a subroutine which in turn takes a single argument (the text to |
1556 | be extracted from). |
1557 | |
1558 | In other words, the implementation of C<extract_tagged> is exactly |
1559 | equivalent to: |
1560 | |
1561 | sub extract_tagged |
1562 | { |
1563 | my $text = shift; |
1564 | $extractor = gen_extract_tagged(@_); |
1565 | return $extractor->($text); |
1566 | } |
1567 | |
1568 | (although C<extract_tagged> is not currently implemented that way, in order |
1569 | to preserve pre-5.005 compatibility). |
1570 | |
1571 | Using C<gen_extract_tagged> to create extraction functions for specific tags |
1572 | is a good idea if those functions are going to be called more than once, since |
1573 | their performance is typically twice as good as the more general-purpose |
1574 | C<extract_tagged>. |
1575 | |
1576 | |
1577 | =head2 C<extract_quotelike> |
1578 | |
1579 | C<extract_quotelike> attempts to recognize, extract, and segment any |
1580 | one of the various Perl quotes and quotelike operators (see |
1581 | L<perlop(3)>) Nested backslashed delimiters, embedded balanced bracket |
1582 | delimiters (for the quotelike operators), and trailing modifiers are |
1583 | all caught. For example, in: |
1584 | |
1585 | extract_quotelike 'q # an octothorpe: \# (not the end of the q!) #' |
1586 | |
1587 | extract_quotelike ' "You said, \"Use sed\"." ' |
1588 | |
1589 | extract_quotelike ' s{([A-Z]{1,8}\.[A-Z]{3})} /\L$1\E/; ' |
1590 | |
1591 | extract_quotelike ' tr/\\\/\\\\/\\\//ds; ' |
1592 | |
1593 | the full Perl quotelike operations are all extracted correctly. |
1594 | |
1595 | Note too that, when using the /x modifier on a regex, any comment |
1596 | containing the current pattern delimiter will cause the regex to be |
1597 | immediately terminated. In other words: |
1598 | |
1599 | 'm / |
1600 | (?i) # CASE INSENSITIVE |
1601 | [a-z_] # LEADING ALPHABETIC/UNDERSCORE |
1602 | [a-z0-9]* # FOLLOWED BY ANY NUMBER OF ALPHANUMERICS |
1603 | /x' |
1604 | |
1605 | will be extracted as if it were: |
1606 | |
1607 | 'm / |
1608 | (?i) # CASE INSENSITIVE |
1609 | [a-z_] # LEADING ALPHABETIC/' |
1610 | |
1611 | This behaviour is identical to that of the actual compiler. |
1612 | |
1613 | C<extract_quotelike> takes two arguments: the text to be processed and |
1614 | a prefix to be matched at the very beginning of the text. If no prefix |
1615 | is specified, optional whitespace is the default. If no text is given, |
1616 | C<$_> is used. |
1617 | |
1618 | In a list context, an array of 11 elements is returned. The elements are: |
1619 | |
1620 | =over 4 |
1621 | |
1622 | =item [0] |
1623 | |
1624 | the extracted quotelike substring (including trailing modifiers), |
1625 | |
1626 | =item [1] |
1627 | |
1628 | the remainder of the input text, |
1629 | |
1630 | =item [2] |
1631 | |
1632 | the prefix substring (if any), |
1633 | |
1634 | =item [3] |
1635 | |
1636 | the name of the quotelike operator (if any), |
1637 | |
1638 | =item [4] |
1639 | |
1640 | the left delimiter of the first block of the operation, |
1641 | |
1642 | =item [5] |
1643 | |
1644 | the text of the first block of the operation |
1645 | (that is, the contents of |
1646 | a quote, the regex of a match or substitution or the target list of a |
1647 | translation), |
1648 | |
1649 | =item [6] |
1650 | |
1651 | the right delimiter of the first block of the operation, |
1652 | |
1653 | =item [7] |
1654 | |
1655 | the left delimiter of the second block of the operation |
d1be9408 |
1656 | (that is, if it is an C<s>, C<tr>, or C<y>), |
55a1c97c |
1657 | |
1658 | =item [8] |
1659 | |
1660 | the text of the second block of the operation |
1661 | (that is, the replacement of a substitution or the translation list |
1662 | of a translation), |
1663 | |
1664 | =item [9] |
1665 | |
1666 | the right delimiter of the second block of the operation (if any), |
1667 | |
1668 | =item [10] |
1669 | |
1670 | the trailing modifiers on the operation (if any). |
1671 | |
1672 | =back |
1673 | |
1674 | For each of the fields marked "(if any)" the default value on success is |
1675 | an empty string. |
1676 | On failure, all of these values (except the remaining text) are C<undef>. |
1677 | |
1678 | |
1679 | In a scalar context, C<extract_quotelike> returns just the complete substring |
1680 | that matched a quotelike operation (or C<undef> on failure). In a scalar or |
1681 | void context, the input text has the same substring (and any specified |
1682 | prefix) removed. |
1683 | |
1684 | Examples: |
1685 | |
1686 | # Remove the first quotelike literal that appears in text |
1687 | |
1688 | $quotelike = extract_quotelike($text,'.*?'); |
1689 | |
1690 | # Replace one or more leading whitespace-separated quotelike |
1691 | # literals in $_ with "<QLL>" |
1692 | |
1693 | do { $_ = join '<QLL>', (extract_quotelike)[2,1] } until $@; |
1694 | |
1695 | |
1696 | # Isolate the search pattern in a quotelike operation from $text |
1697 | |
1698 | ($op,$pat) = (extract_quotelike $text)[3,5]; |
1699 | if ($op =~ /[ms]/) |
1700 | { |
1701 | print "search pattern: $pat\n"; |
1702 | } |
1703 | else |
1704 | { |
1705 | print "$op is not a pattern matching operation\n"; |
1706 | } |
1707 | |
1708 | |
1709 | =head2 C<extract_quotelike> and "here documents" |
1710 | |
1711 | C<extract_quotelike> can successfully extract "here documents" from an input |
1712 | string, but with an important caveat in list contexts. |
1713 | |
1714 | Unlike other types of quote-like literals, a here document is rarely |
1715 | a contiguous substring. For example, a typical piece of code using |
1716 | here document might look like this: |
1717 | |
1718 | <<'EOMSG' || die; |
1719 | This is the message. |
1720 | EOMSG |
1721 | exit; |
1722 | |
1723 | Given this as an input string in a scalar context, C<extract_quotelike> |
1724 | would correctly return the string "<<'EOMSG'\nThis is the message.\nEOMSG", |
1725 | leaving the string " || die;\nexit;" in the original variable. In other words, |
1726 | the two separate pieces of the here document are successfully extracted and |
1727 | concatenated. |
1728 | |
1729 | In a list context, C<extract_quotelike> would return the list |
1730 | |
1731 | =over 4 |
1732 | |
1733 | =item [0] |
1734 | |
1735 | "<<'EOMSG'\nThis is the message.\nEOMSG\n" (i.e. the full extracted here document, |
1736 | including fore and aft delimiters), |
1737 | |
1738 | =item [1] |
1739 | |
1740 | " || die;\nexit;" (i.e. the remainder of the input text, concatenated), |
1741 | |
1742 | =item [2] |
1743 | |
1744 | "" (i.e. the prefix substring -- trivial in this case), |
1745 | |
1746 | =item [3] |
1747 | |
1748 | "<<" (i.e. the "name" of the quotelike operator) |
1749 | |
1750 | =item [4] |
1751 | |
1752 | "'EOMSG'" (i.e. the left delimiter of the here document, including any quotes), |
1753 | |
1754 | =item [5] |
1755 | |
1756 | "This is the message.\n" (i.e. the text of the here document), |
1757 | |
1758 | =item [6] |
1759 | |
1760 | "EOMSG" (i.e. the right delimiter of the here document), |
1761 | |
1762 | =item [7..10] |
1763 | |
1764 | "" (a here document has no second left delimiter, second text, second right |
1765 | delimiter, or trailing modifiers). |
1766 | |
1767 | =back |
1768 | |
1769 | However, the matching position of the input variable would be set to |
1770 | "exit;" (i.e. I<after> the closing delimiter of the here document), |
1771 | which would cause the earlier " || die;\nexit;" to be skipped in any |
1772 | sequence of code fragment extractions. |
1773 | |
d1be9408 |
1774 | To avoid this problem, when it encounters a here document while |
55a1c97c |
1775 | extracting from a modifiable string, C<extract_quotelike> silently |
1776 | rearranges the string to an equivalent piece of Perl: |
1777 | |
1778 | <<'EOMSG' |
1779 | This is the message. |
1780 | EOMSG |
1781 | || die; |
1782 | exit; |
1783 | |
1784 | in which the here document I<is> contiguous. It still leaves the |
1785 | matching position after the here document, but now the rest of the line |
1786 | on which the here document starts is not skipped. |
1787 | |
1788 | To prevent <extract_quotelike> from mucking about with the input in this way |
1789 | (this is the only case where a list-context C<extract_quotelike> does so), |
1790 | you can pass the input variable as an interpolated literal: |
1791 | |
1792 | $quotelike = extract_quotelike("$var"); |
1793 | |
1794 | |
1795 | =head2 C<extract_codeblock> |
1796 | |
1797 | C<extract_codeblock> attempts to recognize and extract a balanced |
1798 | bracket delimited substring that may contain unbalanced brackets |
1799 | inside Perl quotes or quotelike operations. That is, C<extract_codeblock> |
1800 | is like a combination of C<"extract_bracketed"> and |
1801 | C<"extract_quotelike">. |
1802 | |
1803 | C<extract_codeblock> takes the same initial three parameters as C<extract_bracketed>: |
1804 | a text to process, a set of delimiter brackets to look for, and a prefix to |
1805 | match first. It also takes an optional fourth parameter, which allows the |
1806 | outermost delimiter brackets to be specified separately (see below). |
1807 | |
1808 | Omitting the first argument (input text) means process C<$_> instead. |
1809 | Omitting the second argument (delimiter brackets) indicates that only C<'{'> is to be used. |
1810 | Omitting the third argument (prefix argument) implies optional whitespace at the start. |
1811 | Omitting the fourth argument (outermost delimiter brackets) indicates that the |
1812 | value of the second argument is to be used for the outermost delimiters. |
1813 | |
d1be9408 |
1814 | Once the prefix an the outermost opening delimiter bracket have been |
55a1c97c |
1815 | recognized, code blocks are extracted by stepping through the input text and |
1816 | trying the following alternatives in sequence: |
1817 | |
1818 | =over 4 |
1819 | |
1820 | =item 1. |
1821 | |
1822 | Try and match a closing delimiter bracket. If the bracket was the same |
1823 | species as the last opening bracket, return the substring to that |
1824 | point. If the bracket was mismatched, return an error. |
1825 | |
1826 | =item 2. |
1827 | |
1828 | Try to match a quote or quotelike operator. If found, call |
1829 | C<extract_quotelike> to eat it. If C<extract_quotelike> fails, return |
1830 | the error it returned. Otherwise go back to step 1. |
1831 | |
1832 | =item 3. |
1833 | |
1834 | Try to match an opening delimiter bracket. If found, call |
1835 | C<extract_codeblock> recursively to eat the embedded block. If the |
1836 | recursive call fails, return an error. Otherwise, go back to step 1. |
1837 | |
1838 | =item 4. |
1839 | |
1840 | Unconditionally match a bareword or any other single character, and |
1841 | then go back to step 1. |
1842 | |
1843 | =back |
1844 | |
1845 | |
1846 | Examples: |
1847 | |
1848 | # Find a while loop in the text |
1849 | |
1850 | if ($text =~ s/.*?while\s*\{/{/) |
1851 | { |
1852 | $loop = "while " . extract_codeblock($text); |
1853 | } |
1854 | |
1855 | # Remove the first round-bracketed list (which may include |
1856 | # round- or curly-bracketed code blocks or quotelike operators) |
1857 | |
1858 | extract_codeblock $text, "(){}", '[^(]*'; |
1859 | |
1860 | |
1861 | The ability to specify a different outermost delimiter bracket is useful |
1862 | in some circumstances. For example, in the Parse::RecDescent module, |
1863 | parser actions which are to be performed only on a successful parse |
1864 | are specified using a C<E<lt>defer:...E<gt>> directive. For example: |
1865 | |
1866 | sentence: subject verb object |
1867 | <defer: {$::theVerb = $item{verb}} > |
1868 | |
1869 | Parse::RecDescent uses C<extract_codeblock($text, '{}E<lt>E<gt>')> to extract the code |
1870 | within the C<E<lt>defer:...E<gt>> directive, but there's a problem. |
1871 | |
1872 | A deferred action like this: |
1873 | |
1874 | <defer: {if ($count>10) {$count--}} > |
1875 | |
1876 | will be incorrectly parsed as: |
1877 | |
1878 | <defer: {if ($count> |
1879 | |
1880 | because the "less than" operator is interpreted as a closing delimiter. |
1881 | |
1882 | But, by extracting the directive using |
1883 | S<C<extract_codeblock($text, '{}', undef, 'E<lt>E<gt>')>> |
1884 | the '>' character is only treated as a delimited at the outermost |
1885 | level of the code block, so the directive is parsed correctly. |
1886 | |
1887 | =head2 C<extract_multiple> |
1888 | |
1889 | The C<extract_multiple> subroutine takes a string to be processed and a |
1890 | list of extractors (subroutines or regular expressions) to apply to that string. |
1891 | |
1892 | In an array context C<extract_multiple> returns an array of substrings |
1893 | of the original string, as extracted by the specified extractors. |
1894 | In a scalar context, C<extract_multiple> returns the first |
1895 | substring successfully extracted from the original string. In both |
1896 | scalar and void contexts the original string has the first successfully |
1897 | extracted substring removed from it. In all contexts |
1898 | C<extract_multiple> starts at the current C<pos> of the string, and |
1899 | sets that C<pos> appropriately after it matches. |
1900 | |
d1be9408 |
1901 | Hence, the aim of a call to C<extract_multiple> in a list context |
55a1c97c |
1902 | is to split the processed string into as many non-overlapping fields as |
1903 | possible, by repeatedly applying each of the specified extractors |
1904 | to the remainder of the string. Thus C<extract_multiple> is |
1905 | a generalized form of Perl's C<split> subroutine. |
1906 | |
1907 | The subroutine takes up to four optional arguments: |
1908 | |
1909 | =over 4 |
1910 | |
1911 | =item 1. |
1912 | |
1913 | A string to be processed (C<$_> if the string is omitted or C<undef>) |
1914 | |
1915 | =item 2. |
1916 | |
1917 | A reference to a list of subroutine references and/or qr// objects and/or |
1918 | literal strings and/or hash references, specifying the extractors |
1919 | to be used to split the string. If this argument is omitted (or |
1920 | C<undef>) the list: |
1921 | |
1922 | [ |
1923 | sub { extract_variable($_[0], '') }, |
1924 | sub { extract_quotelike($_[0],'') }, |
1925 | sub { extract_codeblock($_[0],'{}','') }, |
1926 | ] |
1927 | |
1928 | is used. |
1929 | |
1930 | |
1931 | =item 3. |
1932 | |
d1be9408 |
1933 | A number specifying the maximum number of fields to return. If this |
55a1c97c |
1934 | argument is omitted (or C<undef>), split continues as long as possible. |
1935 | |
1936 | If the third argument is I<N>, then extraction continues until I<N> fields |
1937 | have been successfully extracted, or until the string has been completely |
1938 | processed. |
1939 | |
1940 | Note that in scalar and void contexts the value of this argument is |
1941 | automatically reset to 1 (under C<-w>, a warning is issued if the argument |
1942 | has to be reset). |
1943 | |
1944 | =item 4. |
1945 | |
1946 | A value indicating whether unmatched substrings (see below) within the |
1947 | text should be skipped or returned as fields. If the value is true, |
1948 | such substrings are skipped. Otherwise, they are returned. |
1949 | |
1950 | =back |
1951 | |
1952 | The extraction process works by applying each extractor in |
a7602084 |
1953 | sequence to the text string. |
1954 | |
1955 | If the extractor is a subroutine it is called in a list context and is |
1956 | expected to return a list of a single element, namely the extracted |
1957 | text. It may optionally also return two further arguments: a string |
1958 | representing the text left after extraction (like $' for a pattern |
1959 | match), and a string representing any prefix skipped before the |
1960 | extraction (like $` in a pattern match). Note that this is designed |
1961 | to facilitate the use of other Text::Balanced subroutines with |
1962 | C<extract_multiple>. Note too that the value returned by an extractor |
1963 | subroutine need not bear any relationship to the corresponding substring |
1964 | of the original text (see examples below). |
55a1c97c |
1965 | |
1966 | If the extractor is a precompiled regular expression or a string, |
1967 | it is matched against the text in a scalar context with a leading |
1968 | '\G' and the gc modifiers enabled. The extracted value is either |
1969 | $1 if that variable is defined after the match, or else the |
1970 | complete match (i.e. $&). |
1971 | |
1972 | If the extractor is a hash reference, it must contain exactly one element. |
1973 | The value of that element is one of the |
1974 | above extractor types (subroutine reference, regular expression, or string). |
1975 | The key of that element is the name of a class into which the successful |
1976 | return value of the extractor will be blessed. |
1977 | |
1978 | If an extractor returns a defined value, that value is immediately |
1979 | treated as the next extracted field and pushed onto the list of fields. |
1980 | If the extractor was specified in a hash reference, the field is also |
1981 | blessed into the appropriate class, |
1982 | |
1983 | If the extractor fails to match (in the case of a regex extractor), or returns an empty list or an undefined value (in the case of a subroutine extractor), it is |
1984 | assumed to have failed to extract. |
1985 | If none of the extractor subroutines succeeds, then one |
1986 | character is extracted from the start of the text and the extraction |
1987 | subroutines reapplied. Characters which are thus removed are accumulated and |
1988 | eventually become the next field (unless the fourth argument is true, in which |
d1be9408 |
1989 | case they are discarded). |
55a1c97c |
1990 | |
1991 | For example, the following extracts substrings that are valid Perl variables: |
1992 | |
1993 | @fields = extract_multiple($text, |
1994 | [ sub { extract_variable($_[0]) } ], |
1995 | undef, 1); |
1996 | |
1997 | This example separates a text into fields which are quote delimited, |
1998 | curly bracketed, and anything else. The delimited and bracketed |
1999 | parts are also blessed to identify them (the "anything else" is unblessed): |
2000 | |
2001 | @fields = extract_multiple($text, |
2002 | [ |
2003 | { Delim => sub { extract_delimited($_[0],q{'"}) } }, |
2004 | { Brack => sub { extract_bracketed($_[0],'{}') } }, |
2005 | ]); |
2006 | |
2007 | This call extracts the next single substring that is a valid Perl quotelike |
2008 | operator (and removes it from $text): |
2009 | |
2010 | $quotelike = extract_multiple($text, |
2011 | [ |
2012 | sub { extract_quotelike($_[0]) }, |
2013 | ], undef, 1); |
2014 | |
2015 | Finally, here is yet another way to do comma-separated value parsing: |
2016 | |
2017 | @fields = extract_multiple($csv_text, |
2018 | [ |
2019 | sub { extract_delimited($_[0],q{'"}) }, |
2020 | qr/([^,]+)(.*)/, |
2021 | ], |
2022 | undef,1); |
2023 | |
2024 | The list in the second argument means: |
2025 | I<"Try and extract a ' or " delimited string, otherwise extract anything up to a comma...">. |
2026 | The undef third argument means: |
2027 | I<"...as many times as possible...">, |
2028 | and the true value in the fourth argument means |
2029 | I<"...discarding anything else that appears (i.e. the commas)">. |
2030 | |
2031 | If you wanted the commas preserved as separate fields (i.e. like split |
2032 | does if your split pattern has capturing parentheses), you would |
2033 | just make the last parameter undefined (or remove it). |
2034 | |
2035 | |
2036 | =head2 C<gen_delimited_pat> |
2037 | |
2038 | The C<gen_delimited_pat> subroutine takes a single (string) argument and |
2039 | > builds a Friedl-style optimized regex that matches a string delimited |
2040 | by any one of the characters in the single argument. For example: |
2041 | |
2042 | gen_delimited_pat(q{'"}) |
2043 | |
2044 | returns the regex: |
2045 | |
2046 | (?:\"(?:\\\"|(?!\").)*\"|\'(?:\\\'|(?!\').)*\') |
2047 | |
2048 | Note that the specified delimiters are automatically quotemeta'd. |
2049 | |
2050 | A typical use of C<gen_delimited_pat> would be to build special purpose tags |
2051 | for C<extract_tagged>. For example, to properly ignore "empty" XML elements |
2052 | (which might contain quoted strings): |
2053 | |
2054 | my $empty_tag = '<(' . gen_delimited_pat(q{'"}) . '|.)+/>'; |
2055 | |
2056 | extract_tagged($text, undef, undef, undef, {ignore => [$empty_tag]} ); |
2057 | |
2058 | |
2059 | C<gen_delimited_pat> may also be called with an optional second argument, |
2060 | which specifies the "escape" character(s) to be used for each delimiter. |
2061 | For example to match a Pascal-style string (where ' is the delimiter |
2062 | and '' is a literal ' within the string): |
2063 | |
2064 | gen_delimited_pat(q{'},q{'}); |
2065 | |
2066 | Different escape characters can be specified for different delimiters. |
2067 | For example, to specify that '/' is the escape for single quotes |
2068 | and '%' is the escape for double quotes: |
2069 | |
2070 | gen_delimited_pat(q{'"},q{/%}); |
2071 | |
2072 | If more delimiters than escape chars are specified, the last escape char |
2073 | is used for the remaining delimiters. |
2074 | If no escape char is specified for a given specified delimiter, '\' is used. |
2075 | |
2076 | Note that |
2077 | C<gen_delimited_pat> was previously called |
2078 | C<delimited_pat>. That name may still be used, but is now deprecated. |
2079 | |
2080 | |
2081 | =head1 DIAGNOSTICS |
2082 | |
2083 | In a list context, all the functions return C<(undef,$original_text)> |
2084 | on failure. In a scalar context, failure is indicated by returning C<undef> |
2085 | (in this case the input text is not modified in any way). |
2086 | |
2087 | In addition, on failure in I<any> context, the C<$@> variable is set. |
2088 | Accessing C<$@-E<gt>{error}> returns one of the error diagnostics listed |
2089 | below. |
2090 | Accessing C<$@-E<gt>{pos}> returns the offset into the original string at |
2091 | which the error was detected (although not necessarily where it occurred!) |
2092 | Printing C<$@> directly produces the error message, with the offset appended. |
2093 | On success, the C<$@> variable is guaranteed to be C<undef>. |
2094 | |
2095 | The available diagnostics are: |
2096 | |
2097 | =over 4 |
2098 | |
2099 | =item C<Did not find a suitable bracket: "%s"> |
2100 | |
2101 | The delimiter provided to C<extract_bracketed> was not one of |
2102 | C<'()[]E<lt>E<gt>{}'>. |
2103 | |
2104 | =item C<Did not find prefix: /%s/> |
2105 | |
2106 | A non-optional prefix was specified but wasn't found at the start of the text. |
2107 | |
2108 | =item C<Did not find opening bracket after prefix: "%s"> |
2109 | |
2110 | C<extract_bracketed> or C<extract_codeblock> was expecting a |
2111 | particular kind of bracket at the start of the text, and didn't find it. |
2112 | |
2113 | =item C<No quotelike operator found after prefix: "%s"> |
2114 | |
2115 | C<extract_quotelike> didn't find one of the quotelike operators C<q>, |
2116 | C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> at the start of the substring |
2117 | it was extracting. |
2118 | |
2119 | =item C<Unmatched closing bracket: "%c"> |
2120 | |
2121 | C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> encountered |
2122 | a closing bracket where none was expected. |
2123 | |
2124 | =item C<Unmatched opening bracket(s): "%s"> |
2125 | |
2126 | C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> ran |
2127 | out of characters in the text before closing one or more levels of nested |
2128 | brackets. |
2129 | |
2130 | =item C<Unmatched embedded quote (%s)> |
2131 | |
2132 | C<extract_bracketed> attempted to match an embedded quoted substring, but |
2133 | failed to find a closing quote to match it. |
2134 | |
2135 | =item C<Did not find closing delimiter to match '%s'> |
2136 | |
2137 | C<extract_quotelike> was unable to find a closing delimiter to match the |
2138 | one that opened the quote-like operation. |
2139 | |
2140 | =item C<Mismatched closing bracket: expected "%c" but found "%s"> |
2141 | |
2142 | C<extract_bracketed>, C<extract_quotelike> or C<extract_codeblock> found |
2143 | a valid bracket delimiter, but it was the wrong species. This usually |
2144 | indicates a nesting error, but may indicate incorrect quoting or escaping. |
2145 | |
2146 | =item C<No block delimiter found after quotelike "%s"> |
2147 | |
2148 | C<extract_quotelike> or C<extract_codeblock> found one of the |
2149 | quotelike operators C<q>, C<qq>, C<qw>, C<qx>, C<s>, C<tr> or C<y> |
2150 | without a suitable block after it. |
2151 | |
2152 | =item C<Did not find leading dereferencer> |
2153 | |
2154 | C<extract_variable> was expecting one of '$', '@', or '%' at the start of |
2155 | a variable, but didn't find any of them. |
2156 | |
2157 | =item C<Bad identifier after dereferencer> |
2158 | |
2159 | C<extract_variable> found a '$', '@', or '%' indicating a variable, but that |
2160 | character was not followed by a legal Perl identifier. |
2161 | |
2162 | =item C<Did not find expected opening bracket at %s> |
2163 | |
2164 | C<extract_codeblock> failed to find any of the outermost opening brackets |
2165 | that were specified. |
2166 | |
2167 | =item C<Improperly nested codeblock at %s> |
2168 | |
2169 | A nested code block was found that started with a delimiter that was specified |
2170 | as being only to be used as an outermost bracket. |
2171 | |
2172 | =item C<Missing second block for quotelike "%s"> |
2173 | |
2174 | C<extract_codeblock> or C<extract_quotelike> found one of the |
2175 | quotelike operators C<s>, C<tr> or C<y> followed by only one block. |
2176 | |
2177 | =item C<No match found for opening bracket> |
2178 | |
2179 | C<extract_codeblock> failed to find a closing bracket to match the outermost |
2180 | opening bracket. |
2181 | |
2182 | =item C<Did not find opening tag: /%s/> |
2183 | |
2184 | C<extract_tagged> did not find a suitable opening tag (after any specified |
2185 | prefix was removed). |
2186 | |
2187 | =item C<Unable to construct closing tag to match: /%s/> |
2188 | |
2189 | C<extract_tagged> matched the specified opening tag and tried to |
2190 | modify the matched text to produce a matching closing tag (because |
2191 | none was specified). It failed to generate the closing tag, almost |
2192 | certainly because the opening tag did not start with a |
2193 | bracket of some kind. |
2194 | |
2195 | =item C<Found invalid nested tag: %s> |
2196 | |
2197 | C<extract_tagged> found a nested tag that appeared in the "reject" list |
2198 | (and the failure mode was not "MAX" or "PARA"). |
2199 | |
2200 | =item C<Found unbalanced nested tag: %s> |
2201 | |
2202 | C<extract_tagged> found a nested opening tag that was not matched by a |
2203 | corresponding nested closing tag (and the failure mode was not "MAX" or "PARA"). |
2204 | |
2205 | =item C<Did not find closing tag> |
2206 | |
2207 | C<extract_tagged> reached the end of the text without finding a closing tag |
2208 | to match the original opening tag (and the failure mode was not |
2209 | "MAX" or "PARA"). |
2210 | |
2211 | |
2212 | |
2213 | |
2214 | =back |
2215 | |
2216 | |
2217 | =head1 AUTHOR |
2218 | |
2219 | Damian Conway (damian@conway.org) |
2220 | |
2221 | |
2222 | =head1 BUGS AND IRRITATIONS |
2223 | |
2224 | There are undoubtedly serious bugs lurking somewhere in this code, if |
2225 | only because parts of it give the impression of understanding a great deal |
2226 | more about Perl than they really do. |
2227 | |
2228 | Bug reports and other feedback are most welcome. |
2229 | |
2230 | |
2231 | =head1 COPYRIGHT |
2232 | |
2233 | Copyright (c) 1997-2001, Damian Conway. All Rights Reserved. |
2234 | This module is free software. It may be used, redistributed |
2235 | and/or modified under the same terms as Perl itself. |