Commit | Line | Data |
b3eb6a9b |
1 | package re; |
2 | |
99cc5cc6 |
3 | # pragma for controlling the regexp engine |
de8c5301 |
4 | use strict; |
5 | use warnings; |
6 | |
0a0b870a |
7 | our $VERSION = "0.11"; |
de8c5301 |
8 | our @ISA = qw(Exporter); |
ec781434 |
9 | our @EXPORT_OK = ('regmust', |
192c1e27 |
10 | qw(is_regexp regexp_pattern |
11 | regname regnames regnames_count)); |
de8c5301 |
12 | our %EXPORT_OK = map { $_ => 1 } @EXPORT_OK; |
13 | |
de8c5301 |
14 | my %bitmask = ( |
15 | taint => 0x00100000, # HINT_RE_TAINT |
16 | eval => 0x00200000, # HINT_RE_EVAL |
17 | ); |
18 | |
de8c5301 |
19 | sub setcolor { |
20 | eval { # Ignore errors |
21 | require Term::Cap; |
22 | |
23 | my $terminal = Tgetent Term::Cap ({OSPEED => 9600}); # Avoid warning. |
24 | my $props = $ENV{PERL_RE_TC} || 'md,me,so,se,us,ue'; |
25 | my @props = split /,/, $props; |
26 | my $colors = join "\t", map {$terminal->Tputs($_,1)} @props; |
27 | |
28 | $colors =~ s/\0//g; |
29 | $ENV{PERL_RE_COLORS} = $colors; |
30 | }; |
31 | if ($@) { |
32 | $ENV{PERL_RE_COLORS} ||= qq'\t\t> <\t> <\t\t'; |
33 | } |
34 | |
35 | } |
36 | |
37 | my %flags = ( |
38 | COMPILE => 0x0000FF, |
39 | PARSE => 0x000001, |
40 | OPTIMISE => 0x000002, |
41 | TRIEC => 0x000004, |
42 | DUMP => 0x000008, |
f7819f85 |
43 | FLAGS => 0x000010, |
de8c5301 |
44 | |
45 | EXECUTE => 0x00FF00, |
46 | INTUIT => 0x000100, |
47 | MATCH => 0x000200, |
48 | TRIEE => 0x000400, |
49 | |
50 | EXTRA => 0xFF0000, |
51 | TRIEM => 0x010000, |
52 | OFFSETS => 0x020000, |
53 | OFFSETSDBG => 0x040000, |
54 | STATE => 0x080000, |
55 | OPTIMISEM => 0x100000, |
56 | STACK => 0x280000, |
e7707071 |
57 | BUFFERS => 0x400000, |
2c296965 |
58 | GPOS => 0x800000, |
de8c5301 |
59 | ); |
e7707071 |
60 | $flags{ALL} = -1 & ~($flags{OFFSETS}|$flags{OFFSETSDBG}|$flags{BUFFERS}); |
de8c5301 |
61 | $flags{All} = $flags{all} = $flags{DUMP} | $flags{EXECUTE}; |
2c296965 |
62 | $flags{Extra} = $flags{EXECUTE} | $flags{COMPILE} | $flags{GPOS}; |
de8c5301 |
63 | $flags{More} = $flags{MORE} = $flags{All} | $flags{TRIEC} | $flags{TRIEM} | $flags{STATE}; |
64 | $flags{State} = $flags{DUMP} | $flags{EXECUTE} | $flags{STATE}; |
65 | $flags{TRIE} = $flags{DUMP} | $flags{EXECUTE} | $flags{TRIEC}; |
66 | |
ec781434 |
67 | if (defined &DynaLoader::boot_DynaLoader) { |
68 | require XSLoader; |
69 | XSLoader::load( __PACKAGE__, $VERSION); |
de8c5301 |
70 | } |
ec781434 |
71 | # else we're miniperl |
72 | # We need to work for miniperl, because the XS toolchain uses Text::Wrap, which |
73 | # uses re 'taint'. |
de8c5301 |
74 | |
75 | sub _load_unload { |
76 | my ($on)= @_; |
77 | if ($on) { |
ec781434 |
78 | # We call install() every time, as if we didn't, we wouldn't |
79 | # "see" any changes to the color environment var since |
80 | # the last time it was called. |
81 | |
82 | # install() returns an integer, which if casted properly |
99cc5cc6 |
83 | # in C resolves to a structure containing the regexp |
ec781434 |
84 | # hooks. Setting it to a random integer will guarantee |
85 | # segfaults. |
86 | $^H{regcomp} = install(); |
de8c5301 |
87 | } else { |
88 | delete $^H{regcomp}; |
89 | } |
90 | } |
91 | |
92 | sub bits { |
93 | my $on = shift; |
94 | my $bits = 0; |
95 | unless (@_) { |
96 | require Carp; |
97 | Carp::carp("Useless use of \"re\" pragma"); |
98 | } |
99 | foreach my $idx (0..$#_){ |
100 | my $s=$_[$idx]; |
101 | if ($s eq 'Debug' or $s eq 'Debugcolor') { |
102 | setcolor() if $s =~/color/i; |
103 | ${^RE_DEBUG_FLAGS} = 0 unless defined ${^RE_DEBUG_FLAGS}; |
104 | for my $idx ($idx+1..$#_) { |
105 | if ($flags{$_[$idx]}) { |
106 | if ($on) { |
107 | ${^RE_DEBUG_FLAGS} |= $flags{$_[$idx]}; |
108 | } else { |
109 | ${^RE_DEBUG_FLAGS} &= ~ $flags{$_[$idx]}; |
110 | } |
111 | } else { |
112 | require Carp; |
113 | Carp::carp("Unknown \"re\" Debug flag '$_[$idx]', possible flags: ", |
114 | join(", ",sort keys %flags ) ); |
115 | } |
116 | } |
117 | _load_unload($on ? 1 : ${^RE_DEBUG_FLAGS}); |
118 | last; |
119 | } elsif ($s eq 'debug' or $s eq 'debugcolor') { |
120 | setcolor() if $s =~/color/i; |
121 | _load_unload($on); |
66e6b4c5 |
122 | last; |
de8c5301 |
123 | } elsif (exists $bitmask{$s}) { |
124 | $bits |= $bitmask{$s}; |
125 | } elsif ($EXPORT_OK{$s}) { |
de8c5301 |
126 | require Exporter; |
127 | re->export_to_level(2, 're', $s); |
128 | } else { |
129 | require Carp; |
130 | Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ", |
131 | join(', ', map {qq('$_')} 'debug', 'debugcolor', sort keys %bitmask), |
132 | ")"); |
133 | } |
134 | } |
135 | $bits; |
136 | } |
137 | |
138 | sub import { |
139 | shift; |
140 | $^H |= bits(1, @_); |
141 | } |
142 | |
143 | sub unimport { |
144 | shift; |
145 | $^H &= ~ bits(0, @_); |
146 | } |
147 | |
148 | 1; |
149 | |
150 | __END__ |
56953603 |
151 | |
b3eb6a9b |
152 | =head1 NAME |
153 | |
154 | re - Perl pragma to alter regular expression behaviour |
155 | |
156 | =head1 SYNOPSIS |
157 | |
e4d48cc9 |
158 | use re 'taint'; |
159 | ($x) = ($^X =~ /^(.*)$/s); # $x is tainted here |
b3eb6a9b |
160 | |
2cd61cdb |
161 | $pat = '(?{ $foo = 1 })'; |
e4d48cc9 |
162 | use re 'eval'; |
2cd61cdb |
163 | /foo${pat}bar/; # won't fail (when not under -T switch) |
e4d48cc9 |
164 | |
165 | { |
166 | no re 'taint'; # the default |
167 | ($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here |
168 | |
169 | no re 'eval'; # the default |
2cd61cdb |
170 | /foo${pat}bar/; # disallowed (with or without -T switch) |
e4d48cc9 |
171 | } |
b3eb6a9b |
172 | |
1e2e3d02 |
173 | use re 'debug'; # output debugging info during |
174 | /^(.*)$/s; # compile and run time |
175 | |
2cd61cdb |
176 | |
02ea72ae |
177 | use re 'debugcolor'; # same as 'debug', but with colored output |
178 | ... |
179 | |
a3621e74 |
180 | use re qw(Debug All); # Finer tuned debugging options. |
4ee9a43f |
181 | use re qw(Debug More); |
fe759410 |
182 | no re qw(Debug ALL); # Turn of all re debugging in this scope |
4ee9a43f |
183 | |
de8c5301 |
184 | use re qw(is_regexp regexp_pattern); # import utility functions |
185 | my ($pat,$mods)=regexp_pattern(qr/foo/i); |
186 | if (is_regexp($obj)) { |
187 | print "Got regexp: ", |
188 | scalar regexp_pattern($obj); # just as perl would stringify it |
189 | } # but no hassle with blessed re's. |
a3621e74 |
190 | |
3ffabb8c |
191 | (We use $^X in these examples because it's tainted by default.) |
192 | |
b3eb6a9b |
193 | =head1 DESCRIPTION |
194 | |
de8c5301 |
195 | =head2 'taint' mode |
196 | |
b3eb6a9b |
197 | When C<use re 'taint'> is in effect, and a tainted string is the target |
99cc5cc6 |
198 | of a regexp, the regexp memories (or values returned by the m// operator |
199 | in list context) are tainted. This feature is useful when regexp operations |
e4d48cc9 |
200 | on tainted data aren't meant to extract safe substrings, but to perform |
201 | other transformations. |
b3eb6a9b |
202 | |
de8c5301 |
203 | =head2 'eval' mode |
204 | |
99cc5cc6 |
205 | When C<use re 'eval'> is in effect, a regexp is allowed to contain |
0b370c0a |
206 | C<(?{ ... })> zero-width assertions and C<(??{ ... })> postponed |
207 | subexpressions, even if the regular expression contains |
ffbc6a93 |
208 | variable interpolation. That is normally disallowed, since it is a |
2cd61cdb |
209 | potential security risk. Note that this pragma is ignored when the regular |
210 | expression is obtained from tainted data, i.e. evaluation is always |
0b370c0a |
211 | disallowed with tainted regular expressions. See L<perlre/(?{ code })> |
bb1773de |
212 | and L<perlre/(??{ code })>. |
2cd61cdb |
213 | |
ffbc6a93 |
214 | For the purpose of this pragma, interpolation of precompiled regular |
0a92e3a8 |
215 | expressions (i.e., the result of C<qr//>) is I<not> considered variable |
216 | interpolation. Thus: |
2cd61cdb |
217 | |
218 | /foo${pat}bar/ |
219 | |
ffbc6a93 |
220 | I<is> allowed if $pat is a precompiled regular expression, even |
0b370c0a |
221 | if $pat contains C<(?{ ... })> assertions or C<(??{ ... })> subexpressions. |
2cd61cdb |
222 | |
de8c5301 |
223 | =head2 'debug' mode |
224 | |
ffbc6a93 |
225 | When C<use re 'debug'> is in effect, perl emits debugging messages when |
2cd61cdb |
226 | compiling and using regular expressions. The output is the same as that |
227 | obtained by running a C<-DDEBUGGING>-enabled perl interpreter with the |
228 | B<-Dr> switch. It may be quite voluminous depending on the complexity |
02ea72ae |
229 | of the match. Using C<debugcolor> instead of C<debug> enables a |
230 | form of output that can be used to get a colorful display on terminals |
231 | that understand termcap color sequences. Set C<$ENV{PERL_RE_TC}> to a |
232 | comma-separated list of C<termcap> properties to use for highlighting |
ffbc6a93 |
233 | strings on/off, pre-point part on/off. |
2cd61cdb |
234 | See L<perldebug/"Debugging regular expressions"> for additional info. |
235 | |
de8c5301 |
236 | As of 5.9.5 the directive C<use re 'debug'> and its equivalents are |
237 | lexically scoped, as the other directives are. However they have both |
238 | compile-time and run-time effects. |
239 | |
240 | See L<perlmodlib/Pragmatic Modules>. |
241 | |
242 | =head2 'Debug' mode |
243 | |
a3621e74 |
244 | Similarly C<use re 'Debug'> produces debugging output, the difference |
245 | being that it allows the fine tuning of what debugging output will be |
be8e71aa |
246 | emitted. Options are divided into three groups, those related to |
247 | compilation, those related to execution and those related to special |
248 | purposes. The options are as follows: |
249 | |
250 | =over 4 |
251 | |
252 | =item Compile related options |
253 | |
254 | =over 4 |
255 | |
256 | =item COMPILE |
257 | |
258 | Turns on all compile related debug options. |
259 | |
260 | =item PARSE |
261 | |
262 | Turns on debug output related to the process of parsing the pattern. |
263 | |
264 | =item OPTIMISE |
265 | |
266 | Enables output related to the optimisation phase of compilation. |
267 | |
24b23f37 |
268 | =item TRIEC |
be8e71aa |
269 | |
270 | Detailed info about trie compilation. |
271 | |
272 | =item DUMP |
273 | |
274 | Dump the final program out after it is compiled and optimised. |
275 | |
be8e71aa |
276 | =back |
277 | |
278 | =item Execute related options |
279 | |
280 | =over 4 |
281 | |
282 | =item EXECUTE |
283 | |
284 | Turns on all execute related debug options. |
285 | |
286 | =item MATCH |
287 | |
288 | Turns on debugging of the main matching loop. |
289 | |
24b23f37 |
290 | =item TRIEE |
be8e71aa |
291 | |
292 | Extra debugging of how tries execute. |
293 | |
294 | =item INTUIT |
295 | |
296 | Enable debugging of start point optimisations. |
297 | |
298 | =back |
299 | |
300 | =item Extra debugging options |
301 | |
302 | =over 4 |
303 | |
304 | =item EXTRA |
305 | |
306 | Turns on all "extra" debugging options. |
307 | |
e7707071 |
308 | =item BUFFERS |
309 | |
310 | Enable debugging the capture buffer storage during match. Warning, |
311 | this can potentially produce extremely large output. |
312 | |
24b23f37 |
313 | =item TRIEM |
314 | |
315 | Enable enhanced TRIE debugging. Enhances both TRIEE |
316 | and TRIEC. |
317 | |
318 | =item STATE |
319 | |
4ee9a43f |
320 | Enable debugging of states in the engine. |
24b23f37 |
321 | |
322 | =item STACK |
be8e71aa |
323 | |
24b23f37 |
324 | Enable debugging of the recursion stack in the engine. Enabling |
325 | or disabling this option automatically does the same for debugging |
326 | states as well. This output from this can be quite large. |
327 | |
328 | =item OPTIMISEM |
329 | |
330 | Enable enhanced optimisation debugging and start point optimisations. |
99cc5cc6 |
331 | Probably not useful except when debugging the regexp engine itself. |
24b23f37 |
332 | |
333 | =item OFFSETS |
334 | |
335 | Dump offset information. This can be used to see how regops correlate |
336 | to the pattern. Output format is |
337 | |
338 | NODENUM:POSITION[LENGTH] |
339 | |
340 | Where 1 is the position of the first char in the string. Note that position |
341 | can be 0, or larger than the actual length of the pattern, likewise length |
342 | can be zero. |
be8e71aa |
343 | |
24b23f37 |
344 | =item OFFSETSDBG |
be8e71aa |
345 | |
346 | Enable debugging of offsets information. This emits copious |
fe759410 |
347 | amounts of trace information and doesn't mesh well with other |
be8e71aa |
348 | debug options. |
349 | |
fe759410 |
350 | Almost definitely only useful to people hacking |
be8e71aa |
351 | on the offsets part of the debug engine. |
352 | |
353 | =back |
354 | |
355 | =item Other useful flags |
356 | |
357 | These are useful shortcuts to save on the typing. |
358 | |
359 | =over 4 |
360 | |
361 | =item ALL |
362 | |
e7707071 |
363 | Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS |
be8e71aa |
364 | |
365 | =item All |
366 | |
fe759410 |
367 | Enable DUMP and all execute options. Equivalent to: |
be8e71aa |
368 | |
369 | use re 'debug'; |
370 | |
371 | =item MORE |
372 | |
373 | =item More |
374 | |
24b23f37 |
375 | Enable TRIEM and all execute compile and execute options. |
be8e71aa |
376 | |
dba3f186 |
377 | =back |
be8e71aa |
378 | |
dba3f186 |
379 | =back |
a3621e74 |
380 | |
1e2e3d02 |
381 | As of 5.9.5 the directive C<use re 'debug'> and its equivalents are |
4ee9a43f |
382 | lexically scoped, as the other directives are. However they have both |
1e2e3d02 |
383 | compile-time and run-time effects. |
b3eb6a9b |
384 | |
de8c5301 |
385 | =head2 Exportable Functions |
b3eb6a9b |
386 | |
de8c5301 |
387 | As of perl 5.9.5 're' debug contains a number of utility functions that |
4ee9a43f |
388 | may be optionally exported into the caller's namespace. They are listed |
de8c5301 |
389 | below. |
b3eb6a9b |
390 | |
de8c5301 |
391 | =over 4 |
b3eb6a9b |
392 | |
de8c5301 |
393 | =item is_regexp($ref) |
02ea72ae |
394 | |
de8c5301 |
395 | Returns true if the argument is a compiled regular expression as returned |
4ee9a43f |
396 | by C<qr//>, false if it is not. |
02ea72ae |
397 | |
4ee9a43f |
398 | This function will not be confused by overloading or blessing. In |
399 | internals terms, this extracts the regexp pointer out of the |
de8c5301 |
400 | PERL_MAGIC_qr structure so it it cannot be fooled. |
894be9b7 |
401 | |
de8c5301 |
402 | =item regexp_pattern($ref) |
02ea72ae |
403 | |
4ee9a43f |
404 | If the argument is a compiled regular expression as returned by C<qr//>, |
405 | then this function returns the pattern. |
be8e71aa |
406 | |
4ee9a43f |
407 | In list context it returns a two element list, the first element |
408 | containing the pattern and the second containing the modifiers used when |
409 | the pattern was compiled. |
be8e71aa |
410 | |
4ee9a43f |
411 | my ($pat, $mods) = regexp_pattern($ref); |
a3621e74 |
412 | |
99cc5cc6 |
413 | In scalar context it returns the same as perl would when stringifying a raw |
4ee9a43f |
414 | C<qr//> with the same pattern inside. If the argument is not a compiled |
415 | reference then this routine returns false but defined in scalar context, |
416 | and the empty list in list context. Thus the following |
f9f4320a |
417 | |
de8c5301 |
418 | if (regexp_pattern($ref) eq '(?i-xsm:foo)') |
dba3f186 |
419 | |
de8c5301 |
420 | will be warning free regardless of what $ref actually is. |
380e0b81 |
421 | |
4ee9a43f |
422 | Like C<is_regexp> this function will not be confused by overloading |
423 | or blessing of the object. |
b3eb6a9b |
424 | |
256ddcd0 |
425 | =item regmust($ref) |
426 | |
432acd5f |
427 | If the argument is a compiled regular expression as returned by C<qr//>, |
99cc5cc6 |
428 | then this function returns what the optimiser considers to be the longest |
432acd5f |
429 | anchored fixed string and longest floating fixed string in the pattern. |
430 | |
431 | A I<fixed string> is defined as being a substring that must appear for the |
432 | pattern to match. An I<anchored fixed string> is a fixed string that must |
433 | appear at a particular offset from the beginning of the match. A I<floating |
434 | fixed string> is defined as a fixed string that can appear at any point in |
435 | a range of positions relative to the start of the match. For example, |
436 | |
437 | my $qr = qr/here .* there/x; |
438 | my ($anchored, $floating) = regmust($qr); |
256ddcd0 |
439 | print "anchored:'$anchored'\nfloating:'$floating'\n"; |
432acd5f |
440 | |
256ddcd0 |
441 | results in |
442 | |
443 | anchored:'here' |
444 | floating:'there' |
445 | |
432acd5f |
446 | Because the C<here> is before the C<.*> in the pattern, its position |
447 | can be determined exactly. That's not true, however, for the C<there>; |
448 | it could appear at any point after where the anchored string appeared. |
256ddcd0 |
449 | Perl uses both for its optimisations, prefering the longer, or, if they are |
450 | equal, the floating. |
451 | |
452 | B<NOTE:> This may not necessarily be the definitive longest anchored and |
432acd5f |
453 | floating string. This will be what the optimiser of the Perl that you |
256ddcd0 |
454 | are using thinks is the longest. If you believe that the result is wrong |
455 | please report it via the L<perlbug> utility. |
456 | |
28d8d7f4 |
457 | =item regname($name,$all) |
44a2ac75 |
458 | |
28d8d7f4 |
459 | Returns the contents of a named buffer of the last successful match. If |
460 | $all is true, then returns an array ref containing one entry per buffer, |
44a2ac75 |
461 | otherwise returns the first defined buffer. |
462 | |
28d8d7f4 |
463 | =item regnames($all) |
44a2ac75 |
464 | |
28d8d7f4 |
465 | Returns a list of all of the named buffers defined in the last successful |
466 | match. If $all is true, then it returns all names defined, if not it returns |
467 | only names which were involved in the match. |
44a2ac75 |
468 | |
28d8d7f4 |
469 | =item regnames_count() |
44a2ac75 |
470 | |
28d8d7f4 |
471 | Returns the number of distinct names defined in the pattern used |
472 | for the last successful match. |
44a2ac75 |
473 | |
28d8d7f4 |
474 | B<Note:> this result is always the actual number of distinct |
475 | named buffers defined, it may not actually match that which is |
476 | returned by C<regnames()> and related routines when those routines |
477 | have not been called with the $all parameter set. |
44a2ac75 |
478 | |
de8c5301 |
479 | =back |
b3eb6a9b |
480 | |
de8c5301 |
481 | =head1 SEE ALSO |
b3eb6a9b |
482 | |
de8c5301 |
483 | L<perlmodlib/Pragmatic Modules>. |
484 | |
485 | =cut |