3 perlreapi - perl regular expression plugin interface
7 As of Perl 5.9.5 there is a new interface for using other regexp engines than
8 the default one. Each engine is supposed to provide access to a constant
9 structure of the following format:
11 typedef struct regexp_engine {
12 REGEXP* (*comp) (pTHX_ const SV * const pattern, const U32 flags);
13 I32 (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend,
14 char* strbeg, I32 minend, SV* screamer,
15 void* data, U32 flags);
16 char* (*intuit) (pTHX_ REGEXP * const rx, SV *sv, char *strpos,
17 char *strend, U32 flags,
18 struct re_scream_pos_data_s *data);
19 SV* (*checkstr) (pTHX_ REGEXP * const rx);
20 void (*free) (pTHX_ REGEXP * const rx);
21 void (*numbered_buff_get) (pTHX_ REGEXP * const rx,
22 const I32 paren, SV * const usesv);
23 SV* (*named_buff_get)(pTHX_ REGEXP * const rx, SV * const namesv,
25 SV* (*qr_package)(pTHX_ REGEXP * const rx);
27 void* (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
31 When a regexp is compiled, its C<engine> field is then set to point at
32 the appropriate structure so that when it needs to be used Perl can find
33 the right routines to do so.
35 In order to install a new regexp handler, C<$^H{regcomp}> is set
36 to an integer which (when casted appropriately) resolves to one of these
37 structures. When compiling, the C<comp> method is executed, and the
38 resulting regexp structure's engine field is expected to point back at
41 The pTHX_ symbol in the definition is a macro used by perl under threading
42 to provide an extra argument to the routine holding a pointer back to
43 the interpreter that is executing the regexp. So under threading all
44 routines get an extra argument.
46 The routines are as follows:
50 REGEXP* comp(pTHX_ const SV * const pattern, const U32 flags);
52 Compile the pattern stored in C<pattern> using the given C<flags> and
53 return a pointer to a prepared C<REGEXP> structure that can perform
54 the match. See L</The REGEXP structure> below for an explanation of
55 the individual fields in the REGEXP struct.
57 The C<pattern> parameter is the scalar that was used as the
58 pattern. previous versions of perl would pass two C<char*> indicating
59 the start and end of the stringifed pattern, the following snippet can
60 be used to get the old parameters:
63 char* exp = SvPV(pattern, plen);
64 char* xend = exp + plen;
66 Since any scalar can be passed as a pattern it's possible to implement
67 an engine that does something with an array (C<< "ook" =~ [ qw/ eek
68 hlagh / ] >>) or with the non-stringified form of a compiled regular
69 expression (C<< "ook" =~ qr/eek/ >>). perl's own engine will always
70 stringify everything using the snippet above but that doesn't mean
71 other engines have to.
73 The C<flags> paramater is a bitfield which indicates which of the
74 C<msixk> flags the regex was compiled with. In addition it contains
75 info about whether C<use locale> is in effect and optimization info
76 for C<split>. A regex engine might want to use the same split
77 optimizations with a different syntax, for instance a Perl6 engine
78 would treat C<split /^^/> equivalently to perl's C<split /^/>, see
79 L<split documentation|perlfunc> and the relevant code in C<pp_split>
80 in F<pp.c> to find out whether your engine should be setting these.
82 The C<eogc> flags are stripped out before being passed to the comp
83 routine. The regex engine does not need to know whether any of these
84 are set as those flags should only affect what perl does with the
85 pattern and its match variables, not how it gets compiled & executed.
91 C<split ' '> or C<split> with no arguments (which really means
92 C<split(' ', $_> see L<split|perlfunc>).
96 Set if the pattern is C</^/> (C<<r->prelen == 1 && r->precomp[0] ==
97 '^'>>). Will be used by the C<split> operator to split the given
98 string on C<\n> (even under C</^/s>, see L<split|perlfunc>).
102 Set if the pattern is exactly C</\s+/> and used by C<split>, the
103 definition of whitespace varies depending on whether RXf_UTF8 or
104 RXf_PMf_LOCALE is set.
108 Makes C<split> use the locale dependant definition of whitespace under C<use
109 locale> when RXf_SKIPWHITE or RXf_WHITE is in effect. Under ASCII whitespace is
110 defined as per L<isSPACE|perlapi/ISSPACE>, and by the internal macros
111 C<is_utf8_space> under UTF-8 and C<isSPACE_LC> under C<use locale>.
113 =item RXf_PMf_MULTILINE
115 The C</m> flag, this ends up being passed to C<Perl_fbm_instr> by
116 C<pp_split> regardless of the engine.
118 =item RXf_PMf_SINGLELINE
120 The C</s> flag. Guaranteed not to be used outside the regex engine.
124 The C</i> flag. Guaranteed not to be used outside the regex engine.
126 =item RXf_PMf_EXTENDED
128 The C</x> flag. Guaranteed not to be used outside the regex
129 engine. However if present on a regex C<#> comments will be stripped
130 by the tokenizer regardless of the engine currently in use.
132 =item RXf_PMf_KEEPCOPY
138 Set if the pattern is L<SvUTF8()|perlapi/SvUTF8>, set by Perl_pmruntime.
142 In general these flags should be preserved in regex->extflags after
143 compilation, although it is possible the regex includes constructs
144 that changes them. The perl engine for instance may upgrade non-utf8
145 strings to utf8 if the pattern includes constructs such as C<\x{...}>
146 that can only match unicode values. RXf_SKIPWHITE should always be
147 preserved verbatim in regex->extflags.
151 I32 exec(pTHX_ REGEXP * const rx,
152 char *stringarg, char* strend, char* strbeg,
153 I32 minend, SV* screamer,
154 void* data, U32 flags);
160 char* intuit(pTHX_ REGEXP * const rx,
161 SV *sv, char *strpos, char *strend,
162 const U32 flags, struct re_scream_pos_data_s *data);
164 Find the start position where a regex match should be attempted,
165 or possibly whether the regex engine should not be run because the
166 pattern can't match. This is called as appropriate by the core
167 depending on the values of the extflags member of the regexp
172 SV* checkstr(pTHX_ REGEXP * const rx);
174 Return a SV containing a string that must appear in the pattern. Used
175 by C<split> for optimising matches.
179 void free(pTHX_ REGEXP * const rx);
181 Called by perl when it is freeing a regexp pattern so that the engine
182 can release any resources pointed to by the C<pprivate> member of the
183 regexp structure. This is only responsible for freeing private data;
184 perl will handle releasing anything else contained in the regexp structure.
186 =head2 numbered_buff_get
188 void numbered_buff_get(pTHX_ REGEXP * const rx, const I32 paren,
191 Called to get the value of C<$`>, C<$'>, C<$&> (and their named
192 equivalents, see L<perlvar>) and the numbered capture buffers (C<$1>,
195 The C<paren> paramater will be C<-2> for C<$`>, C<-1> for C<$'>, C<0>
196 for C<$&>, C<1> for C<$1> and so forth.
198 C<usesv> should be set to the scalar to return, the scalar is passed
199 as an argument rather than being returned from the function because
200 when it's called perl already has a scalar to store the value,
201 creating another one would be redundant. The scalar can be set with
202 C<sv_setsv>, C<sv_setpvn> and friends, see L<perlapi>.
204 This callback is where perl untaints its own capture variables under
205 taint mode (see L<perlsec>). See the C<Perl_reg_numbered_buff_get>
206 function in F<regcomp.c> for how to untaint capture variables if
207 that's something you'd like your engine to do as well.
209 =head2 named_buff_get
211 SV* named_buff_get(pTHX_ REGEXP * const rx, SV * const namesv,
214 Called to get the value of key in the C<%+> and C<%-> hashes,
215 C<namesv> is the hash key being requested and if C<flags & 1> is true
216 C<%-> is being requested (and C<%+> if it's not).
220 SV* qr_package(pTHX_ REGEXP * const rx);
222 The package the qr// magic object is blessed into (as seen by C<ref
223 qr//>). It is recommended that engines change this to their package
224 name for identification regardless of whether they implement methods
227 A callback implementation might be:
230 Example_reg_qr_package(pTHX_ REGEXP * const rx)
233 return newSVpvs("re::engine::Example");
236 Any method calls on an object created with C<qr//> will be dispatched to the
237 package as a normal object.
239 use re::engine::Example;
241 $re->meth; # dispatched to re::engine::Example::meth()
243 To retrieve the C<REGEXP> object from the scalar in an XS function use the
254 (sv = (SV*)SvRV(sv)) && /* assignment deliberate */
255 SvTYPE(sv) == SVt_PVMG &&
256 (mg = mg_find(sv, PERL_MAGIC_qr))) /* assignment deliberate */
258 re = (REGEXP *)mg->mg_obj;
261 Or use the (CURRENTLY UNDOCUMENETED!) C<Perl_get_re_arg> function:
265 const REGEXP * const re = (REGEXP *)Perl_get_re_arg( aTHX_ rv, 0, NULL );
269 void* dupe(pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
271 On threaded builds a regexp may need to be duplicated so that the pattern
272 can be used by mutiple threads. This routine is expected to handle the
273 duplication of any private data pointed to by the C<pprivate> member of
274 the regexp structure. It will be called with the preconstructed new
275 regexp structure as an argument, the C<pprivate> member will point at
276 the B<old> private structue, and it is this routine's responsibility to
277 construct a copy and return a pointer to it (which perl will then use to
278 overwrite the field as passed to this routine.)
280 This allows the engine to dupe its private data but also if necessary
281 modify the final structure if it really must.
283 On unthreaded builds this field doesn't exist.
285 =head1 The REGEXP structure
287 The REGEXP struct is defined in F<regexp.h>. All regex engines must be able to
288 correctly build such a structure in their L</comp> routine.
290 The REGEXP structure contains all the data that perl needs to be aware of
291 to properly work with the regular expression. It includes data about
292 optimisations that perl can use to determine if the regex engine should
293 really be used, and various other control info that is needed to properly
294 execute patterns in various contexts such as is the pattern anchored in
295 some way, or what flags were used during the compile, or whether the
296 program contains special constructs that perl needs to be aware of.
298 In addition it contains two fields that are intended for the private use
299 of the regex engine that compiled the pattern. These are the C<intflags>
300 and pprivate members. The C<pprivate> is a void pointer to an arbitrary
301 structure whose use and management is the responsibility of the compiling
302 engine. perl will never modify either of these values.
304 typedef struct regexp {
305 /* what engine created this regexp? */
306 const struct regexp_engine* engine;
308 /* what re is this a lightweight copy of? */
309 struct regexp* mother_re;
311 /* Information about the match that the perl core uses to manage things */
312 U32 extflags; /* Flags used both externally and internally */
313 I32 minlen; /* mininum possible length of string to match */
314 I32 minlenret; /* mininum possible length of $& */
315 U32 gofs; /* chars left of pos that we search from */
317 /* substring data about strings that must appear
318 in the final match, used for optimisations */
319 struct reg_substr_data *substrs;
321 U32 nparens; /* number of capture buffers */
323 /* private engine specific data */
324 U32 intflags; /* Engine Specific Internal flags */
325 void *pprivate; /* Data private to the regex engine which
326 created this object. */
328 /* Data about the last/current match. These are modified during matching*/
329 U32 lastparen; /* last open paren matched */
330 U32 lastcloseparen; /* last close paren matched */
331 regexp_paren_pair *swap; /* Swap copy of *offs */
332 regexp_paren_pair *offs; /* Array of offsets for (@-) and (@+) */
334 char *subbeg; /* saved or original string so \digit works forever. */
335 SV_SAVED_COPY /* If non-NULL, SV which is COW from original */
336 I32 sublen; /* Length of string pointed by subbeg */
338 /* Information about the match that isn't often used */
339 I32 prelen; /* length of precomp */
340 const char *precomp; /* pre-compilation regular expression */
342 /* wrapped can't be const char*, as it is returned by sv_2pv_flags */
343 char *wrapped; /* wrapped version of the pattern */
344 I32 wraplen; /* length of wrapped */
346 I32 seen_evals; /* number of eval groups in the pattern - for security checks */
347 HV *paren_names; /* Optional hash of paren names */
349 /* Refcount of this regexp */
350 I32 refcnt; /* Refcount of this regexp */
353 The fields are discussed in more detail below:
359 This field points at a regexp_engine structure which contains pointers
360 to the subroutines that are to be used for performing a match. It
361 is the compiling routine's responsibility to populate this field before
362 returning the regexp object.
364 Internally this is set to C<NULL> unless a custom engine is specified in
365 C<$^H{regcomp}>, perl's own set of callbacks can be accessed in the struct
366 pointed to by C<RE_ENGINE_PTR>.
370 TODO, see L<http://www.mail-archive.com/perl5-changes@perl.org/msg17328.html>
374 This will be used by perl to see what flags the regexp was compiled with, this
375 will normally be set to the value of the flags parameter on L</comp>.
377 =item C<minlen> C<minlenret>
379 The minimum string length required for the pattern to match. This is used to
380 prune the search space by not bothering to match any closer to the end of a
381 string than would allow a match. For instance there is no point in even
382 starting the regex engine if the minlen is 10 but the string is only 5
383 characters long. There is no way that the pattern can match.
385 C<minlenret> is the minimum length of the string that would be found
388 The difference between C<minlen> and C<minlenret> can be seen in the
393 where the C<minlen> would be 3 but C<minlenret> would only be 2 as the \d is
394 required to match but is not actually included in the matched content. This
395 distinction is particularly important as the substitution logic uses the
396 C<minlenret> to tell whether it can do in-place substition which can result in
397 considerable speedup.
401 Left offset from pos() to start match at.
407 =item C<nparens>, C<lasparen>, and C<lastcloseparen>
409 These fields are used to keep track of how many paren groups could be matched
410 in the pattern, which was the last open paren to be entered, and which was
411 the last close paren to be entered.
415 The engine's private copy of the flags the pattern was compiled with. Usually
416 this is the same as C<extflags> unless the engine chose to modify one of them
420 A void* pointing to an engine-defined data structure. The perl engine uses the
421 C<regexp_internal> structure (see L<perlreguts/Base Structures>) but a custom
422 engine should use something else.
430 A C<regexp_paren_pair> structure which defines offsets into the string being
431 matched which correspond to the C<$&> and C<$1>, C<$2> etc. captures, the
432 C<regexp_paren_pair> struct is defined as follows:
434 typedef struct regexp_paren_pair {
439 If C<< ->offs[num].start >> or C<< ->offs[num].end >> is C<-1> then that
440 capture buffer did not match. C<< ->offs[0].start/end >> represents C<$&> (or
441 C<${^MATCH> under C<//p>) and C<< ->offs[paren].end >> matches C<$$paren> where
444 =item C<precomp> C<prelen>
446 Used for debugging purposes. C<precomp> holds a copy of the pattern
447 that was compiled and C<prelen> its length.
451 This is a hash used internally to track named capture buffers and their
452 offsets. The keys are the names of the buffers the values are dualvars,
453 with the IV slot holding the number of buffers with the given name and the
454 pv being an embedded array of I32. The values may also be contained
455 independently in the data array in cases where named backreferences are
458 =item C<reg_substr_data>
460 Holds information on the longest string that must occur at a fixed
461 offset from the start of the pattern, and the longest string that must
462 occur at a floating offset from the start of the pattern. Used to do
463 Fast-Boyer-Moore searches on the string to find out if its worth using
464 the regex engine at all, and if so where in the string to search.
466 =item C<subbeg> C<sublen> C<saved_copy>
468 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
469 if (RX_MATCH_COPIED(ret))
470 ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
474 C<PL_sawampersand || rx->extflags & RXf_PMf_KEEPCOPY>
476 These are used during execution phase for managing search and replace
479 =item C<wrapped> C<wraplen>
481 Stores the string C<qr//> stringifies to, for example C<(?-xism:eek)>
482 in the case of C<qr/eek/>.
484 When using a custom engine that doesn't support the C<(?:)> construct for
485 inline modifiers it's best to have C<qr//> stringify to the supplied pattern,
486 note that this will create invalid patterns in cases such as:
488 my $x = qr/a|b/; # "a|b"
490 my $z = qr/$x$y/; # "a|bc"
492 There's no solution for such problems other than making the custom engine
493 understand some for of inline modifiers.
495 The C<Perl_reg_stringify> in F<regcomp.c> does the stringification work.
499 This stores the number of eval groups in the pattern. This is used for security
500 purposes when embedding compiled regexes into larger patterns with C<qr//>.
504 The number of times the structure is referenced. When this falls to 0 the
505 regexp is automatically freed by a call to pregfree. This should be set to 1 in
506 each engine's L</comp> routine.
510 =head2 De-allocation and Cloning
512 Any patch that adds data items to the REGEXP struct will need to include
513 changes to F<sv.c> (C<Perl_re_dup()>) and F<regcomp.c> (C<pregfree()>). This
514 involves freeing or cloning items in the regexp's data array based on the data
519 Originally part of L<perlreguts>.
523 Originally written by Yves Orton, expanded by E<AElig>var ArnfjE<ouml>rE<eth>
528 Copyright 2006 Yves Orton and 2007 E<AElig>var ArnfjE<ouml>rE<eth> Bjarmason.
530 This program is free software; you can redistribute it and/or modify it under
531 the same terms as Perl itself.