* Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003,
* 2005, 2006, 2007, 2008 by Larry Wall and others
* You may distribute under the terms of either the GNU General Public
* License or the Artistic License, as specified in the README file.
* Definitions etc. for regexp(3) routines.
* Caveat: this is V8 regexp(3) [actually, a reimplementation thereof],
#ifndef PLUGGABLE_RE_EXTENSION
/* we don't want to include this stuff if we are inside of
an external regex engine based on the core one - like re 'debug'*/
typedef struct regnode regnode;
struct reg_substr_datum {
SSize_t min_offset; /* min pos (in chars) that substr must appear */
SSize_t max_offset /* max pos (in chars) that substr must appear */;
SV *substr; /* non-utf8 variant */
SV *utf8_substr; /* utf8 variant */
SSize_t end_shift; /* how many fixed chars must end the string */
U8 check_ix; /* index into data[] of check substr */
struct reg_substr_datum data[3]; /* Actual array */
#define SV_SAVED_COPY SV *saved_copy; /* If non-NULL, SV which is COW from original */
/* offsets within a string of a particular /(.)/ capture */
typedef struct regexp_paren_pair {
/* 'start_tmp' records a new opening position before the matching end
* has been found, so that the old start and end values are still
* "abc" =~ /(.(?{print "[$1]"}))+/
* This field is not part of the API. */
#if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_UTF8_C)
#define _invlist_union(a, b, output) _invlist_union_maybe_complement_2nd(a, b, FALSE, output)
#define _invlist_intersection(a, b, output) _invlist_intersection_maybe_complement_2nd(a, b, FALSE, output)
/* Subtracting b from a leaves in a everything that was there that isn't in b,
* that is the intersection of a with b's complement */
#define _invlist_subtract(a, b, output) _invlist_intersection_maybe_complement_2nd(a, b, TRUE, output)
/* record the position of a (?{...}) within a pattern */
/* array of reg_code_block's plus header info */
int refcnt; /* we may be pointed to from a regex and from the savestack */
int count; /* how many code blocks */
struct reg_code_block *cb; /* array of reg_code_block's */
The regexp/REGEXP struct, see L<perlreapi> for further documentation
on the individual fields. The struct is ordered so that the most
commonly used fields are placed at the start.
Any patch that adds items to this struct will need to include
changes to F<sv.c> (C<Perl_re_dup()>) and F<regcomp.c>
(C<pregfree()>). This involves freeing or cloning items in the
regexp's data array based on the data item's type.
/* what engine created this regexp? */ \
const struct regexp_engine* engine; \
REGEXP *mother_re; /* what re is this a lightweight copy of? */ \
HV *paren_names; /* Optional hash of paren names */ \
/*--------------------------------------------------------*/ \
/* Information about the match that the perl core uses to */ \
U32 extflags; /* Flags used both externally and internally */ \
SSize_t minlen; /* mininum possible number of chars in string to match */\
SSize_t minlenret; /* mininum possible number of chars in $& */ \
STRLEN gofs; /* chars left of pos that we search from */ \
/* substring data about strings that must appear in the */ \
/* final match, used for optimisations */ \
struct reg_substr_data *substrs; \
U32 nparens; /* number of capture buffers */ \
/* private engine specific data */ \
U32 intflags; /* Engine Specific Internal flags */ \
void *pprivate; /* Data private to the regex engine which */ \
/* created this object. */ \
/*--------------------------------------------------------*/ \
/* Data about the last/current match. These are modified */ \
U32 lastparen; /* last open paren matched */ \
U32 lastcloseparen; /* last close paren matched */ \
/* Array of offsets for (@-) and (@+) */ \
regexp_paren_pair *offs; \
char **recurse_locinput; /* used to detect infinite recursion, XXX: move to internal */ \
/*--------------------------------------------------------*/ \
/* saved or original string so \digit works forever. */ \
SV_SAVED_COPY /* If non-NULL, SV which is COW from original */\
SSize_t sublen; /* Length of string pointed by subbeg */ \
SSize_t suboffset; /* byte offset of subbeg from logical start of str */ \
SSize_t subcoffset; /* suboffset equiv, but in chars (for @-/@+) */ \
/* Information about the match that isn't often used */ \
SSize_t maxlen; /* mininum possible number of chars in string to match */\
/*--------------------------------------------------------*/ \
/* offset from wrapped to the start of precomp */ \
PERL_BITFIELD32 pre_prefix:4; \
/* original flags used to compile the pattern, may differ */ \
/* from extflags in various ways */ \
PERL_BITFIELD32 compflags:9; \
/*--------------------------------------------------------*/ \
CV *qr_anoncv /* the anon sub wrapped round qr/(?{..})/ */
#define RXp_PAREN_NAMES(rx) ((rx)->paren_names)
/* used for high speed searches */
typedef struct re_scream_pos_data_s
char **scream_olds; /* match pos */
SSize_t *scream_pos; /* Internal iterator of scream. */
/* regexp_engine structure. This is the dispatch table for regexes.
* Any regex engine implementation must be able to build one of these.
typedef struct regexp_engine {
REGEXP* (*comp) (pTHX_ SV * const pattern, U32 flags);
I32 (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend,
char* strbeg, SSize_t minend, SV* sv,
const char * const strbeg,
re_scream_pos_data *data);
SV* (*checkstr) (pTHX_ REGEXP * const rx);
void (*rxfree) (pTHX_ REGEXP * const rx);
void (*numbered_buff_FETCH) (pTHX_ REGEXP * const rx, const I32 paren,
void (*numbered_buff_STORE) (pTHX_ REGEXP * const rx, const I32 paren,
I32 (*numbered_buff_LENGTH) (pTHX_ REGEXP * const rx, const SV * const sv,
SV* (*named_buff) (pTHX_ REGEXP * const rx, SV * const key,
SV * const value, const U32 flags);
SV* (*named_buff_iter) (pTHX_ REGEXP * const rx, const SV * const lastkey,
SV* (*qr_package)(pTHX_ REGEXP * const rx);
void* (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
REGEXP* (*op_comp) (pTHX_ SV ** const patternp, int pat_count,
OP *expr, const struct regexp_engine* eng,
bool *is_bare_re, U32 orig_rx_flags, U32 pm_flags);
These are passed to the numbered capture variable callbacks as the
paren name. >= 1 is reserved for actual numbered captures, i.e. $1,
#define RX_BUFF_IDX_CARET_PREMATCH -5 /* ${^PREMATCH} */
#define RX_BUFF_IDX_CARET_POSTMATCH -4 /* ${^POSTMATCH} */
#define RX_BUFF_IDX_CARET_FULLMATCH -3 /* ${^MATCH} */
#define RX_BUFF_IDX_PREMATCH -2 /* $` */
#define RX_BUFF_IDX_POSTMATCH -1 /* $' */
#define RX_BUFF_IDX_FULLMATCH 0 /* $& */
Flags that are passed to the named_buff and named_buff_iter
callbacks above. Those routines are called from universal.c via the
Tie::Hash::NamedCapture interface for %+ and %- and the re::
functions in the same file.
/* The Tie::Hash::NamedCapture operation this is part of, if any */
#define RXapif_FETCH 0x0001
#define RXapif_STORE 0x0002
#define RXapif_DELETE 0x0004
#define RXapif_CLEAR 0x0008
#define RXapif_EXISTS 0x0010
#define RXapif_SCALAR 0x0020
#define RXapif_FIRSTKEY 0x0040
#define RXapif_NEXTKEY 0x0080
/* Whether %+ or %- is being operated on */
#define RXapif_ONE 0x0100 /* %+ */
#define RXapif_ALL 0x0200 /* %- */
/* Whether this is being called from a re:: function */
#define RXapif_REGNAME 0x0400
#define RXapif_REGNAMES 0x0800
#define RXapif_REGNAMES_COUNT 0x1000
=for apidoc Am|REGEXP *|SvRX|SV *sv
Convenience macro to get the REGEXP from a SV. This is approximately
equivalent to the following snippet:
sv = MUTABLE_SV(SvRV(sv));
if (SvTYPE(sv) == SVt_REGEXP)
C<NULL> will be returned if a REGEXP* is not found.
=for apidoc Am|bool|SvRXOK|SV* sv
Returns a boolean indicating whether the SV (or the one it references)
If you want to do something with the REGEXP* later use SvRX instead
#define SvRX(sv) (Perl_get_re_arg(aTHX_ sv))
#define SvRXOK(sv) cBOOL(Perl_get_re_arg(aTHX_ sv))
/* Flags stored in regexp->extflags
* These are used by code external to the regexp engine
* Note that the flags whose names start with RXf_PMf_ are defined in
* op_reg_common.h, being copied from the parallel flags of op_pmflags
* NOTE: if you modify any RXf flags you should run regen.pl or
* regen/regcomp.pl so that regnodes.h is updated with the changes.
#include "op_reg_common.h"
#define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED|RXf_PMf_EXTENDED_MORE|RXf_PMf_NOCAPTURE)
#define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl, x_count) \
case IGNORE_PAT_MOD: *(pmfl) |= RXf_PMf_FOLD; break; \
case MULTILINE_PAT_MOD: *(pmfl) |= RXf_PMf_MULTILINE; break; \
case SINGLE_PAT_MOD: *(pmfl) |= RXf_PMf_SINGLELINE; break; \
case XTENDED_PAT_MOD: if (x_count == 0) { \
*(pmfl) |= RXf_PMf_EXTENDED; \
*(pmfl) &= ~RXf_PMf_EXTENDED_MORE; \
*(pmfl) |= RXf_PMf_EXTENDED \
|RXf_PMf_EXTENDED_MORE; \
case NOCAPTURE_PAT_MOD: *(pmfl) |= RXf_PMf_NOCAPTURE; break;
/* Note, includes charset ones, assumes 0 is the default for them */
#define STD_PMMOD_FLAGS_CLEAR(pmfl) \
*(pmfl) &= ~(RXf_PMf_FOLD|RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_EXTENDED|RXf_PMf_EXTENDED_MORE|RXf_PMf_CHARSET|RXf_PMf_NOCAPTURE)
/* chars and strings used as regex pattern modifiers
* Singular is a 'c'har, plural is a "string"
* NOTE, KEEPCOPY was originally 'k', but was changed to 'p' for preserve
* for compatibility reasons with Regexp::Common which highjacked (?k:...)
* for its own uses. So 'k' is out as well.
#define DEFAULT_PAT_MOD '^' /* Short for all the default modifiers */
#define KEEPCOPY_PAT_MOD 'p'
#define NOCAPTURE_PAT_MOD 'n'
#define GLOBAL_PAT_MOD 'g'
#define CONTINUE_PAT_MOD 'c'
#define MULTILINE_PAT_MOD 'm'
#define SINGLE_PAT_MOD 's'
#define IGNORE_PAT_MOD 'i'
#define XTENDED_PAT_MOD 'x'
#define NONDESTRUCT_PAT_MOD 'r'
#define LOCALE_PAT_MOD 'l'
#define UNICODE_PAT_MOD 'u'
#define DEPENDS_PAT_MOD 'd'
#define ASCII_RESTRICT_PAT_MOD 'a'
#define ONCE_PAT_MODS "o"
#define KEEPCOPY_PAT_MODS "p"
#define NOCAPTURE_PAT_MODS "n"
#define EXEC_PAT_MODS "e"
#define LOOP_PAT_MODS "gc"
#define NONDESTRUCT_PAT_MODS "r"
#define LOCALE_PAT_MODS "l"
#define UNICODE_PAT_MODS "u"
#define DEPENDS_PAT_MODS "d"
#define ASCII_RESTRICT_PAT_MODS "a"
#define ASCII_MORE_RESTRICT_PAT_MODS "aa"
/* This string is expected by regcomp.c to be ordered so that the first
* character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of extflags; the next
* character is bit +1, etc. */
#define STD_PAT_MODS "msixxn"
#define CHARSET_PAT_MODS ASCII_RESTRICT_PAT_MODS DEPENDS_PAT_MODS LOCALE_PAT_MODS UNICODE_PAT_MODS
/* This string is expected by XS_re_regexp_pattern() in universal.c to be ordered
* so that the first character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of
* extflags; the next character is in bit +1, etc. */
#define INT_PAT_MODS STD_PAT_MODS KEEPCOPY_PAT_MODS
#define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS NOCAPTURE_PAT_MODS
#define QR_PAT_MODS STD_PAT_MODS EXT_PAT_MODS CHARSET_PAT_MODS
#define M_PAT_MODS QR_PAT_MODS LOOP_PAT_MODS
#define S_PAT_MODS M_PAT_MODS EXEC_PAT_MODS NONDESTRUCT_PAT_MODS
* NOTE: if you modify any RXf flags you should run regen.pl or
* regen/regcomp.pl so that regnodes.h is updated with the changes.
Set in Perl_pmruntime for a split. Will be used by regex engines to
check whether they should set RXf_SKIPWHITE
#define RXf_SPLIT RXf_PMf_SPLIT
/* Currently the regex flags occupy a single 32-bit word. Not all bits are
* currently used. The lower bits are shared with their corresponding PMf flag
* bits, up to but not including _RXf_PMf_SHIFT_NEXT. The unused bits
* immediately follow; finally the used RXf-only (unshared) bits, so that the
* highest bit in the word is used. This gathers all the unused bits as a pool
* in the middle, like so: 11111111111111110000001111111111
* where the '1's represent used bits, and the '0's unused. This design allows
* us to allocate off one end of the pool if we need to add a shared bit, and
* off the other end if we need a non-shared bit, without disturbing the other
* bits. This maximizes the likelihood of being able to change things without
* breaking binary compatibility.
* To add shared bits, do so in op_reg_common.h. This should change
* _RXf_PMf_SHIFT_NEXT so that things won't compile. Then come to regexp.h and
* op.h and adjust the constant adders in the definitions of RXf_BASE_SHIFT and
* Pmf_BASE_SHIFT down by the number of shared bits you added. That's it.
* Things should be binary compatible. But if either of these gets to having
* to subtract rather than add, leave at 0 and instead adjust all the entries
* that are in terms of it. But if the first one of those is already
* RXf_BASE_SHIFT+0, there are no bits left, and a redesign is in order.
* To remove unshared bits, just delete its entry. If you're where breaking
* binary compatibility is ok to do, you might want to adjust things to move
* the newly opened space so that it gets absorbed into the common pool.
* To add unshared bits, first use up any gaps in the middle. Otherwise,
* allocate off the low end until you get to RXf_BASE_SHIFT+0. If that isn't
* enough, move RXf_BASE_SHIFT down (if possible) and add the new bit at the
* other end instead; this preserves binary compatibility.
* For the regexp bits, PL_reg_extflags_name[] in regnodes.h has a comment
* giving which bits are used/unused */
#define RXf_BASE_SHIFT (_RXf_PMf_SHIFT_NEXT + 2)
#define RXf_NO_INPLACE_SUBST (1U<<(RXf_BASE_SHIFT+2))
#define RXf_EVAL_SEEN (1U<<(RXf_BASE_SHIFT+3))
#define RXf_UNBOUNDED_QUANTIFIER_SEEN (1U<<(RXf_BASE_SHIFT+4))
#define RXf_CHECK_ALL (1U<<(RXf_BASE_SHIFT+5))
#define RXf_MATCH_UTF8 (1U<<(RXf_BASE_SHIFT+6)) /* $1 etc are utf8 */
#define RXf_USE_INTUIT_NOML (1U<<(RXf_BASE_SHIFT+7))
#define RXf_USE_INTUIT_ML (1U<<(RXf_BASE_SHIFT+8))
#define RXf_INTUIT_TAIL (1U<<(RXf_BASE_SHIFT+9))
#define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML)
/* Do we have some sort of anchor? */
#define RXf_IS_ANCHORED (1U<<(RXf_BASE_SHIFT+10))
/* Copy and tainted info */
#define RXf_COPY_DONE (1U<<(RXf_BASE_SHIFT+11))
/* post-execution: $1 et al are tainted */
#define RXf_TAINTED_SEEN (1U<<(RXf_BASE_SHIFT+12))
/* this pattern was tainted during compilation */
#define RXf_TAINTED (1U<<(RXf_BASE_SHIFT+13))
/* Flags indicating special patterns */
#define RXf_START_ONLY (1U<<(RXf_BASE_SHIFT+14)) /* Pattern is /^/ */
#define RXf_SKIPWHITE (1U<<(RXf_BASE_SHIFT+15)) /* Pattern is for a */
#define RXf_WHITE (1U<<(RXf_BASE_SHIFT+16)) /* Pattern is /\s+/ */
#define RXf_NULL (1U<<(RXf_BASE_SHIFT+17)) /* Pattern is // */
/* See comments at the beginning of these defines about adding bits. The
* highest bit position should be used, so that if RXf_BASE_SHIFT gets
* increased, the #error below will be triggered so that you will be reminded
* to adjust things at the other end to keep the bit positions unchanged */
#if RXf_BASE_SHIFT+17 > 31
# error Too many RXf_PMf bits used. See comments at beginning of these for what to do
* NOTE: if you modify any RXf flags you should run regen.pl or
* regen/regcomp.pl so that regnodes.h is updated with the changes.
# define RX_ISTAINTED(prog) 0
# define RX_TAINT_on(prog) NOOP
# define RXp_MATCH_TAINTED(prog) 0
# define RX_MATCH_TAINTED(prog) 0
# define RXp_MATCH_TAINTED_on(prog) NOOP
# define RX_MATCH_TAINTED_on(prog) NOOP
# define RX_MATCH_TAINTED_off(prog) NOOP
# define RX_ISTAINTED(prog) (RX_EXTFLAGS(prog) & RXf_TAINTED)
# define RX_TAINT_on(prog) (RX_EXTFLAGS(prog) |= RXf_TAINTED)
# define RXp_MATCH_TAINTED(prog) (RXp_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
# define RX_MATCH_TAINTED(prog) (RX_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
# define RXp_MATCH_TAINTED_on(prog) (RXp_EXTFLAGS(prog) |= RXf_TAINTED_SEEN)
# define RX_MATCH_TAINTED_on(prog) (RX_EXTFLAGS(prog) |= RXf_TAINTED_SEEN)
# define RX_MATCH_TAINTED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_TAINTED_SEEN)
#define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN)
#define RX_MATCH_TAINTED_set(prog, t) ((t) \
? RX_MATCH_TAINTED_on(prog) \
: RX_MATCH_TAINTED_off(prog))
#define RXp_MATCH_COPIED(prog) (RXp_EXTFLAGS(prog) & RXf_COPY_DONE)
#define RX_MATCH_COPIED(prog) (RX_EXTFLAGS(prog) & RXf_COPY_DONE)
#define RXp_MATCH_COPIED_on(prog) (RXp_EXTFLAGS(prog) |= RXf_COPY_DONE)
#define RX_MATCH_COPIED_on(prog) (RX_EXTFLAGS(prog) |= RXf_COPY_DONE)
#define RXp_MATCH_COPIED_off(prog) (RXp_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
#define RX_MATCH_COPIED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
#define RX_MATCH_COPIED_set(prog,t) ((t) \
? RX_MATCH_COPIED_on(prog) \
: RX_MATCH_COPIED_off(prog))
#define RXp_EXTFLAGS(rx) ((rx)->extflags)
#define RXp_COMPFLAGS(rx) ((rx)->compflags)
/* For source compatibility. We used to store these explicitly. */
#define RX_PRECOMP(prog) (RX_WRAPPED(prog) + ReANY(prog)->pre_prefix)
#define RX_PRECOMP_const(prog) (RX_WRAPPED_const(prog) + ReANY(prog)->pre_prefix)
/* FIXME? Are we hardcoding too much here and constraining plugin extension
writers? Specifically, the value 1 assumes that the wrapped version always
has exactly one character at the end, a ')'. Will that always be true? */