Commit | Line | Data |
b3eb6a9b |
1 | package re; |
2 | |
de8c5301 |
3 | # pragma for controlling the regex engine |
4 | use strict; |
5 | use warnings; |
6 | |
192c1e27 |
7 | our $VERSION = "0.09"; |
de8c5301 |
8 | our @ISA = qw(Exporter); |
192c1e27 |
9 | my @XS_FUNCTIONS = qw(regmust); |
10 | my %XS_FUNCTIONS = map { $_ => 1 } @XS_FUNCTIONS; |
11 | our @EXPORT_OK = (@XS_FUNCTIONS, |
12 | qw(is_regexp regexp_pattern |
13 | regname regnames regnames_count)); |
de8c5301 |
14 | our %EXPORT_OK = map { $_ => 1 } @EXPORT_OK; |
15 | |
16 | # *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING *** |
17 | # |
18 | # If you modify these values see comment below! |
19 | |
20 | my %bitmask = ( |
21 | taint => 0x00100000, # HINT_RE_TAINT |
22 | eval => 0x00200000, # HINT_RE_EVAL |
23 | ); |
24 | |
25 | # - File::Basename contains a literal for 'taint' as a fallback. If |
26 | # taint is changed here, File::Basename must be updated as well. |
27 | # |
28 | # - ExtUtils::ParseXS uses a hardcoded |
29 | # BEGIN { $^H |= 0x00200000 } |
30 | # in it to allow re.xs to be built. So if 'eval' is changed here then |
31 | # ExtUtils::ParseXS must be changed as well. |
32 | # |
33 | # *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING *** |
34 | |
35 | sub setcolor { |
36 | eval { # Ignore errors |
37 | require Term::Cap; |
38 | |
39 | my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning. |
40 | my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue'; |
41 | my @props = split /,/, $props; |
42 | my $colors = join "\t", map {$terminal->Tputs($_,1)} @props; |
43 | |
44 | $colors =~ s/\0//g; |
45 | $ENV{PERL_RE_COLORS} = $colors; |
46 | }; |
47 | if ($@) { |
48 | $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t'; |
49 | } |
50 | |
51 | } |
52 | |
53 | my %flags = ( |
54 | COMPILE => 0x0000FF, |
55 | PARSE => 0x000001, |
56 | OPTIMISE => 0x000002, |
57 | TRIEC => 0x000004, |
58 | DUMP => 0x000008, |
f7819f85 |
59 | FLAGS => 0x000010, |
de8c5301 |
60 | |
61 | EXECUTE => 0x00FF00, |
62 | INTUIT => 0x000100, |
63 | MATCH => 0x000200, |
64 | TRIEE => 0x000400, |
65 | |
66 | EXTRA => 0xFF0000, |
67 | TRIEM => 0x010000, |
68 | OFFSETS => 0x020000, |
69 | OFFSETSDBG => 0x040000, |
70 | STATE => 0x080000, |
71 | OPTIMISEM => 0x100000, |
72 | STACK => 0x280000, |
e7707071 |
73 | BUFFERS => 0x400000, |
de8c5301 |
74 | ); |
e7707071 |
75 | $flags{ALL} = -1 & ~($flags{OFFSETS}|$flags{OFFSETSDBG}|$flags{BUFFERS}); |
de8c5301 |
76 | $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE}; |
77 | $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE}; |
78 | $flags{More} = $flags{MORE} = $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE}; |
79 | $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE}; |
80 | $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC}; |
81 | |
82 | my $installed; |
83 | my $installed_error; |
84 | |
85 | sub _do_install { |
86 | if ( ! defined($installed) ) { |
87 | require XSLoader; |
88 | $installed = eval { XSLoader::load('re', $VERSION) } || 0; |
89 | $installed_error = $@; |
90 | } |
91 | } |
92 | |
93 | sub _load_unload { |
94 | my ($on)= @_; |
95 | if ($on) { |
96 | _do_install(); |
97 | if ( ! $installed ) { |
98 | die "'re' not installed!? ($installed_error)"; |
99 | } else { |
100 | # We call install() every time, as if we didn't, we wouldn't |
101 | # "see" any changes to the color environment var since |
102 | # the last time it was called. |
103 | |
104 | # install() returns an integer, which if casted properly |
105 | # in C resolves to a structure containing the regex |
106 | # hooks. Setting it to a random integer will guarantee |
107 | # segfaults. |
108 | $^H{regcomp} = install(); |
109 | } |
110 | } else { |
111 | delete $^H{regcomp}; |
112 | } |
113 | } |
114 | |
115 | sub bits { |
116 | my $on = shift; |
117 | my $bits = 0; |
118 | unless (@_) { |
119 | require Carp; |
120 | Carp::carp("Useless use of \"re\" pragma"); |
121 | } |
122 | foreach my $idx (0..$#_){ |
123 | my $s=$_[$idx]; |
124 | if ($s eq 'Debug' or $s eq 'Debugcolor') { |
125 | setcolor() if $s =~/color/i; |
126 | ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS}; |
127 | for my $idx ($idx+1..$#_) { |
128 | if ($flags{$_[$idx]}) { |
129 | if ($on) { |
130 | ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]}; |
131 | } else { |
132 | ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]}; |
133 | } |
134 | } else { |
135 | require Carp; |
136 | Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ", |
137 | join(", ",sort keys %flags ) ); |
138 | } |
139 | } |
140 | _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS}); |
141 | last; |
142 | } elsif ($s eq 'debug' or $s eq 'debugcolor') { |
143 | setcolor() if $s =~/color/i; |
144 | _load_unload($on); |
66e6b4c5 |
145 | last; |
de8c5301 |
146 | } elsif (exists $bitmask{$s}) { |
147 | $bits |= $bitmask{$s}; |
192c1e27 |
148 | } elsif ($XS_FUNCTIONS{$s}) { |
149 | _do_install(); |
150 | if (! $installed) { |
151 | require Carp; |
152 | Carp::croak("\"re\" function '$s' not available"); |
153 | } |
154 | require Exporter; |
155 | re->export_to_level(2, 're', $s); |
de8c5301 |
156 | } elsif ($EXPORT_OK{$s}) { |
de8c5301 |
157 | require Exporter; |
158 | re->export_to_level(2, 're', $s); |
159 | } else { |
160 | require Carp; |
161 | Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ", |
162 | join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask), |
163 | ")"); |
164 | } |
165 | } |
166 | $bits; |
167 | } |
168 | |
169 | sub import { |
170 | shift; |
171 | $^H |= bits(1, @_); |
172 | } |
173 | |
174 | sub unimport { |
175 | shift; |
176 | $^H &= ~ bits(0, @_); |
177 | } |
178 | |
179 | 1; |
180 | |
181 | __END__ |
56953603 |
182 | |
b3eb6a9b |
183 | =head1 NAME |
184 | |
185 | re - Perl pragma to alter regular expression behaviour |
186 | |
187 | =head1 SYNOPSIS |
188 | |
e4d48cc9 |
189 | use re 'taint'; |
190 | ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here |
b3eb6a9b |
191 | |
2cd61cdb |
192 | $pat = '(?{ $foo = 1 })'; |
e4d48cc9 |
193 | use re 'eval'; |
2cd61cdb |
194 | /foo${pat}bar/; # won't fail (when not under -T switch) |
e4d48cc9 |
195 | |
196 | { |
197 | no re 'taint'; # the default |
198 | ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here |
199 | |
200 | no re 'eval'; # the default |
2cd61cdb |
201 | /foo${pat}bar/; # disallowed (with or without -T switch) |
e4d48cc9 |
202 | } |
b3eb6a9b |
203 | |
1e2e3d02 |
204 | use re 'debug'; # output debugging info during |
205 | /^(.*)$/s; # compile and run time |
206 | |
2cd61cdb |
207 | |
02ea72ae |
208 | use re 'debugcolor'; # same as 'debug', but with colored output |
209 | ... |
210 | |
a3621e74 |
211 | use re qw(Debug All); # Finer tuned debugging options. |
4ee9a43f |
212 | use re qw(Debug More); |
fe759410 |
213 | no re qw(Debug ALL); # Turn of all re debugging in this scope |
4ee9a43f |
214 | |
de8c5301 |
215 | use re qw(is_regexp regexp_pattern); # import utility functions |
216 | my ($pat,$mods)=regexp_pattern(qr/foo/i); |
217 | if (is_regexp($obj)) { |
218 | print "Got regexp: ", |
219 | scalar regexp_pattern($obj); # just as perl would stringify it |
220 | } # but no hassle with blessed re's. |
a3621e74 |
221 | |
3ffabb8c |
222 | (We use $^X in these examples because it's tainted by default.) |
223 | |
b3eb6a9b |
224 | =head1 DESCRIPTION |
225 | |
de8c5301 |
226 | =head2 'taint' mode |
227 | |
b3eb6a9b |
228 | When C<use re 'taint'> is in effect, and a tainted string is the target |
229 | of a regex, the regex memories (or values returned by the m// operator |
e4d48cc9 |
230 | in list context) are tainted. This feature is useful when regex operations |
231 | on tainted data aren't meant to extract safe substrings, but to perform |
232 | other transformations. |
b3eb6a9b |
233 | |
de8c5301 |
234 | =head2 'eval' mode |
235 | |
e4d48cc9 |
236 | When C<use re 'eval'> is in effect, a regex is allowed to contain |
2cd61cdb |
237 | C<(?{ ... })> zero-width assertions even if regular expression contains |
ffbc6a93 |
238 | variable interpolation. That is normally disallowed, since it is a |
2cd61cdb |
239 | potential security risk. Note that this pragma is ignored when the regular |
240 | expression is obtained from tainted data, i.e. evaluation is always |
3c4b39be |
241 | disallowed with tainted regular expressions. See L<perlre/(?{ code })>. |
2cd61cdb |
242 | |
ffbc6a93 |
243 | For the purpose of this pragma, interpolation of precompiled regular |
0a92e3a8 |
244 | expressions (i.e., the result of C<qr//>) is I<not> considered variable |
245 | interpolation. Thus: |
2cd61cdb |
246 | |
247 | /foo${pat}bar/ |
248 | |
ffbc6a93 |
249 | I<is> allowed if $pat is a precompiled regular expression, even |
2cd61cdb |
250 | if $pat contains C<(?{ ... })> assertions. |
251 | |
de8c5301 |
252 | =head2 'debug' mode |
253 | |
ffbc6a93 |
254 | When C<use re 'debug'> is in effect, perl emits debugging messages when |
2cd61cdb |
255 | compiling and using regular expressions. The output is the same as that |
256 | obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the |
257 | B<-Dr> switch. It may be quite voluminous depending on the complexity |
02ea72ae |
258 | of the match. Using C<debugcolor> instead of C<debug> enables a |
259 | form of output that can be used to get a colorful display on terminals |
260 | that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a |
261 | comma-separated list of C<termcap> properties to use for highlighting |
ffbc6a93 |
262 | strings on/off, pre-point part on/off. |
2cd61cdb |
263 | See L<perldebug/"Debugging regular expressions"> for additional info. |
264 | |
de8c5301 |
265 | As of 5.9.5 the directive C<use re 'debug'> and its equivalents are |
266 | lexically scoped, as the other directives are. However they have both |
267 | compile-time and run-time effects. |
268 | |
269 | See L<perlmodlib/Pragmatic Modules>. |
270 | |
271 | =head2 'Debug' mode |
272 | |
a3621e74 |
273 | Similarly C<use re 'Debug'> produces debugging output, the difference |
274 | being that it allows the fine tuning of what debugging output will be |
be8e71aa |
275 | emitted. Options are divided into three groups, those related to |
276 | compilation, those related to execution and those related to special |
277 | purposes. The options are as follows: |
278 | |
279 | =over 4 |
280 | |
281 | =item Compile related options |
282 | |
283 | =over 4 |
284 | |
285 | =item COMPILE |
286 | |
287 | Turns on all compile related debug options. |
288 | |
289 | =item PARSE |
290 | |
291 | Turns on debug output related to the process of parsing the pattern. |
292 | |
293 | =item OPTIMISE |
294 | |
295 | Enables output related to the optimisation phase of compilation. |
296 | |
24b23f37 |
297 | =item TRIEC |
be8e71aa |
298 | |
299 | Detailed info about trie compilation. |
300 | |
301 | =item DUMP |
302 | |
303 | Dump the final program out after it is compiled and optimised. |
304 | |
be8e71aa |
305 | =back |
306 | |
307 | =item Execute related options |
308 | |
309 | =over 4 |
310 | |
311 | =item EXECUTE |
312 | |
313 | Turns on all execute related debug options. |
314 | |
315 | =item MATCH |
316 | |
317 | Turns on debugging of the main matching loop. |
318 | |
24b23f37 |
319 | =item TRIEE |
be8e71aa |
320 | |
321 | Extra debugging of how tries execute. |
322 | |
323 | =item INTUIT |
324 | |
325 | Enable debugging of start point optimisations. |
326 | |
327 | =back |
328 | |
329 | =item Extra debugging options |
330 | |
331 | =over 4 |
332 | |
333 | =item EXTRA |
334 | |
335 | Turns on all "extra" debugging options. |
336 | |
e7707071 |
337 | =item BUFFERS |
338 | |
339 | Enable debugging the capture buffer storage during match. Warning, |
340 | this can potentially produce extremely large output. |
341 | |
24b23f37 |
342 | =item TRIEM |
343 | |
344 | Enable enhanced TRIE debugging. Enhances both TRIEE |
345 | and TRIEC. |
346 | |
347 | =item STATE |
348 | |
4ee9a43f |
349 | Enable debugging of states in the engine. |
24b23f37 |
350 | |
351 | =item STACK |
be8e71aa |
352 | |
24b23f37 |
353 | Enable debugging of the recursion stack in the engine. Enabling |
354 | or disabling this option automatically does the same for debugging |
355 | states as well. This output from this can be quite large. |
356 | |
357 | =item OPTIMISEM |
358 | |
359 | Enable enhanced optimisation debugging and start point optimisations. |
360 | Probably not useful except when debugging the regex engine itself. |
361 | |
362 | =item OFFSETS |
363 | |
364 | Dump offset information. This can be used to see how regops correlate |
365 | to the pattern. Output format is |
366 | |
367 | NODENUM:POSITION[LENGTH] |
368 | |
369 | Where 1 is the position of the first char in the string. Note that position |
370 | can be 0, or larger than the actual length of the pattern, likewise length |
371 | can be zero. |
be8e71aa |
372 | |
24b23f37 |
373 | =item OFFSETSDBG |
be8e71aa |
374 | |
375 | Enable debugging of offsets information. This emits copious |
fe759410 |
376 | amounts of trace information and doesn't mesh well with other |
be8e71aa |
377 | debug options. |
378 | |
fe759410 |
379 | Almost definitely only useful to people hacking |
be8e71aa |
380 | on the offsets part of the debug engine. |
381 | |
382 | =back |
383 | |
384 | =item Other useful flags |
385 | |
386 | These are useful shortcuts to save on the typing. |
387 | |
388 | =over 4 |
389 | |
390 | =item ALL |
391 | |
e7707071 |
392 | Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS |
be8e71aa |
393 | |
394 | =item All |
395 | |
fe759410 |
396 | Enable DUMP and all execute options. Equivalent to: |
be8e71aa |
397 | |
398 | use re 'debug'; |
399 | |
400 | =item MORE |
401 | |
402 | =item More |
403 | |
24b23f37 |
404 | Enable TRIEM and all execute compile and execute options. |
be8e71aa |
405 | |
dba3f186 |
406 | =back |
be8e71aa |
407 | |
dba3f186 |
408 | =back |
a3621e74 |
409 | |
1e2e3d02 |
410 | As of 5.9.5 the directive C<use re 'debug'> and its equivalents are |
4ee9a43f |
411 | lexically scoped, as the other directives are. However they have both |
1e2e3d02 |
412 | compile-time and run-time effects. |
b3eb6a9b |
413 | |
de8c5301 |
414 | =head2 Exportable Functions |
b3eb6a9b |
415 | |
de8c5301 |
416 | As of perl 5.9.5 're' debug contains a number of utility functions that |
4ee9a43f |
417 | may be optionally exported into the caller's namespace. They are listed |
de8c5301 |
418 | below. |
b3eb6a9b |
419 | |
de8c5301 |
420 | =over 4 |
b3eb6a9b |
421 | |
de8c5301 |
422 | =item is_regexp($ref) |
02ea72ae |
423 | |
de8c5301 |
424 | Returns true if the argument is a compiled regular expression as returned |
4ee9a43f |
425 | by C<qr//>, false if it is not. |
02ea72ae |
426 | |
4ee9a43f |
427 | This function will not be confused by overloading or blessing. In |
428 | internals terms, this extracts the regexp pointer out of the |
de8c5301 |
429 | PERL_MAGIC_qr structure so it it cannot be fooled. |
894be9b7 |
430 | |
de8c5301 |
431 | =item regexp_pattern($ref) |
02ea72ae |
432 | |
4ee9a43f |
433 | If the argument is a compiled regular expression as returned by C<qr//>, |
434 | then this function returns the pattern. |
be8e71aa |
435 | |
4ee9a43f |
436 | In list context it returns a two element list, the first element |
437 | containing the pattern and the second containing the modifiers used when |
438 | the pattern was compiled. |
be8e71aa |
439 | |
4ee9a43f |
440 | my ($pat, $mods) = regexp_pattern($ref); |
a3621e74 |
441 | |
4ee9a43f |
442 | In scalar context it returns the same as perl would when strigifying a raw |
443 | C<qr//> with the same pattern inside. If the argument is not a compiled |
444 | reference then this routine returns false but defined in scalar context, |
445 | and the empty list in list context. Thus the following |
f9f4320a |
446 | |
de8c5301 |
447 | if (regexp_pattern($ref) eq '(?i-xsm:foo)') |
dba3f186 |
448 | |
de8c5301 |
449 | will be warning free regardless of what $ref actually is. |
380e0b81 |
450 | |
4ee9a43f |
451 | Like C<is_regexp> this function will not be confused by overloading |
452 | or blessing of the object. |
b3eb6a9b |
453 | |
256ddcd0 |
454 | =item regmust($ref) |
455 | |
432acd5f |
456 | If the argument is a compiled regular expression as returned by C<qr//>, |
457 | then this function returns what the optimiser consiers to be the longest |
458 | anchored fixed string and longest floating fixed string in the pattern. |
459 | |
460 | A I<fixed string> is defined as being a substring that must appear for the |
461 | pattern to match. An I<anchored fixed string> is a fixed string that must |
462 | appear at a particular offset from the beginning of the match. A I<floating |
463 | fixed string> is defined as a fixed string that can appear at any point in |
464 | a range of positions relative to the start of the match. For example, |
465 | |
466 | my $qr = qr/here .* there/x; |
467 | my ($anchored, $floating) = regmust($qr); |
256ddcd0 |
468 | print "anchored:'$anchored'\nfloating:'$floating'\n"; |
432acd5f |
469 | |
256ddcd0 |
470 | results in |
471 | |
472 | anchored:'here' |
473 | floating:'there' |
474 | |
432acd5f |
475 | Because the C<here> is before the C<.*> in the pattern, its position |
476 | can be determined exactly. That's not true, however, for the C<there>; |
477 | it could appear at any point after where the anchored string appeared. |
256ddcd0 |
478 | Perl uses both for its optimisations, prefering the longer, or, if they are |
479 | equal, the floating. |
480 | |
481 | B<NOTE:> This may not necessarily be the definitive longest anchored and |
432acd5f |
482 | floating string. This will be what the optimiser of the Perl that you |
256ddcd0 |
483 | are using thinks is the longest. If you believe that the result is wrong |
484 | please report it via the L<perlbug> utility. |
485 | |
28d8d7f4 |
486 | =item regname($name,$all) |
44a2ac75 |
487 | |
28d8d7f4 |
488 | Returns the contents of a named buffer of the last successful match. If |
489 | $all is true, then returns an array ref containing one entry per buffer, |
44a2ac75 |
490 | otherwise returns the first defined buffer. |
491 | |
28d8d7f4 |
492 | =item regnames($all) |
44a2ac75 |
493 | |
28d8d7f4 |
494 | Returns a list of all of the named buffers defined in the last successful |
495 | match. If $all is true, then it returns all names defined, if not it returns |
496 | only names which were involved in the match. |
44a2ac75 |
497 | |
28d8d7f4 |
498 | =item regnames_count() |
44a2ac75 |
499 | |
28d8d7f4 |
500 | Returns the number of distinct names defined in the pattern used |
501 | for the last successful match. |
44a2ac75 |
502 | |
28d8d7f4 |
503 | B<Note:> this result is always the actual number of distinct |
504 | named buffers defined, it may not actually match that which is |
505 | returned by C<regnames()> and related routines when those routines |
506 | have not been called with the $all parameter set. |
44a2ac75 |
507 | |
de8c5301 |
508 | =back |
b3eb6a9b |
509 | |
de8c5301 |
510 | =head1 SEE ALSO |
b3eb6a9b |
511 | |
de8c5301 |
512 | L<perlmodlib/Pragmatic Modules>. |
513 | |
514 | =cut |