X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/0cadcf80572b885a4a6b7680439582b19fd09887..19720b39a307f3ff449747dea71012038b7a8714:/regexp.h diff --git a/regexp.h b/regexp.h index 0493267..90e3406 100644 --- a/regexp.h +++ b/regexp.h @@ -1,7 +1,7 @@ /* regexp.h * * Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003, - * by Larry Wall and others + * 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. @@ -14,7 +14,9 @@ * Caveat: this is V8 regexp(3) [actually, a reimplementation thereof], * not the System V one. */ - +#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'*/ struct regnode { U8 flags; @@ -28,114 +30,457 @@ struct reg_substr_data; struct reg_data; -typedef struct regexp { - I32 *startp; - I32 *endp; - regnode *regstclass; - struct reg_substr_data *substrs; - char *precomp; /* pre-compilation regular expression */ - struct reg_data *data; /* Additional data. */ - char *subbeg; /* saved or original string - so \digit works forever. */ +struct regexp_engine; +struct regexp; + +struct reg_substr_datum { + I32 min_offset; + I32 max_offset; + SV *substr; /* non-utf8 variant */ + SV *utf8_substr; /* utf8 variant */ + I32 end_shift; +}; +struct reg_substr_data { + struct reg_substr_datum data[3]; /* Actual array */ +}; + #ifdef PERL_OLD_COPY_ON_WRITE - SV *saved_copy; /* If non-NULL, SV which is COW from original */ +#define SV_SAVED_COPY SV *saved_copy; /* If non-NULL, SV which is COW from original */ +#else +#define SV_SAVED_COPY #endif - U32 *offsets; /* offset annotations 20001228 MJD */ - I32 sublen; /* Length of string pointed by subbeg */ - I32 refcnt; - I32 minlen; /* mininum possible length of $& */ - I32 prelen; /* length of precomp */ - U32 nparens; /* number of parentheses */ - U32 lastparen; /* last paren matched */ - U32 lastcloseparen; /* last paren matched */ - U32 reganch; /* Internal use only + - Tainted information used by regexec? */ - regnode program[1]; /* Unwarranted chumminess with compiler. */ + +typedef struct regexp_paren_pair { + I32 start; + I32 end; +} regexp_paren_pair; + +/* + The regexp/REGEXP struct, see L 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 (C) and F + (C). This involves freeing or cloning items in the + regexp's data array based on the data item's type. +*/ + +#define _REGEXP_COMMON \ + /* what engine created this regexp? */ \ + const struct regexp_engine* engine; \ + REGEXP *mother_re; /* what re is this a lightweight copy of? */ \ + /* Information about the match that the perl core uses to */ \ + /* manage things */ \ + U32 extflags; /* Flags used both externally and internally */ \ + I32 minlen; /* mininum possible length of string to match */\ + I32 minlenret; /* mininum possible length of $& */ \ + U32 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 */ \ + /* during matching */ \ + U32 lastparen; /* last open paren matched */ \ + U32 lastcloseparen; /* last close paren matched */ \ + regexp_paren_pair *swap; /* Unused: 5.10.1 and later */ \ + /* Array of offsets for (@-) and (@+) */ \ + regexp_paren_pair *offs; \ + /* saved or original string so \digit works forever. */ \ + char *subbeg; \ + SV_SAVED_COPY /* If non-NULL, SV which is COW from original */\ + I32 sublen; /* Length of string pointed by subbeg */ \ + /* Information about the match that isn't often used */ \ + /* offset from wrapped to the start of precomp */ \ + PERL_BITFIELD32 pre_prefix:4; \ + /* number of eval groups in the pattern - for security checks */\ + PERL_BITFIELD32 seen_evals:28 + +typedef struct regexp { + _XPV_HEAD; + _XPVMG_HEAD; + _REGEXP_COMMON; } regexp; -#define ROPT_ANCH (ROPT_ANCH_BOL|ROPT_ANCH_MBOL|ROPT_ANCH_GPOS|ROPT_ANCH_SBOL) -#define ROPT_ANCH_SINGLE (ROPT_ANCH_SBOL|ROPT_ANCH_GPOS) -#define ROPT_ANCH_BOL 0x00001 -#define ROPT_ANCH_MBOL 0x00002 -#define ROPT_ANCH_SBOL 0x00004 -#define ROPT_ANCH_GPOS 0x00008 -#define ROPT_SKIP 0x00010 -#define ROPT_IMPLICIT 0x00020 /* Converted .* to ^.* */ -#define ROPT_NOSCAN 0x00040 /* Check-string always at start. */ -#define ROPT_GPOS_SEEN 0x00080 -#define ROPT_CHECK_ALL 0x00100 -#define ROPT_LOOKBEHIND_SEEN 0x00200 -#define ROPT_EVAL_SEEN 0x00400 -#define ROPT_CANY_SEEN 0x00800 -#define ROPT_SANY_SEEN ROPT_CANY_SEEN /* src bckwrd cmpt */ - -/* 0xf800 of reganch is used by PMf_COMPILETIME */ - -#define ROPT_UTF8 0x10000 -#define ROPT_NAUGHTY 0x20000 /* how exponential is this pattern? */ -#define ROPT_COPY_DONE 0x40000 /* subbeg is a copy of the string */ -#define ROPT_TAINTED_SEEN 0x80000 -#define ROPT_MATCH_UTF8 0x10000000 /* subbeg is utf-8 */ - -#define RE_USE_INTUIT_NOML 0x0100000 /* Best to intuit before matching */ -#define RE_USE_INTUIT_ML 0x0200000 -#define REINT_AUTORITATIVE_NOML 0x0400000 /* Can trust a positive answer */ -#define REINT_AUTORITATIVE_ML 0x0800000 -#define REINT_ONCE_NOML 0x1000000 /* Intuit can succed once only. */ -#define REINT_ONCE_ML 0x2000000 -#define RE_INTUIT_ONECHAR 0x4000000 -#define RE_INTUIT_TAIL 0x8000000 - -#define RE_USE_INTUIT (RE_USE_INTUIT_NOML|RE_USE_INTUIT_ML) -#define REINT_AUTORITATIVE (REINT_AUTORITATIVE_NOML|REINT_AUTORITATIVE_ML) -#define REINT_ONCE (REINT_ONCE_NOML|REINT_ONCE_ML) - -#define RX_MATCH_TAINTED(prog) ((prog)->reganch & ROPT_TAINTED_SEEN) -#define RX_MATCH_TAINTED_on(prog) ((prog)->reganch |= ROPT_TAINTED_SEEN) -#define RX_MATCH_TAINTED_off(prog) ((prog)->reganch &= ~ROPT_TAINTED_SEEN) +/* HV *paren_names; Optional hash of paren names + now stored in the IV union */ + +#define RXp_PAREN_NAMES(rx) ((rx)->xiv_u.xivu_hv) + +/* used for high speed searches */ +typedef struct re_scream_pos_data_s +{ + char **scream_olds; /* match pos */ + I32 *scream_pos; /* Internal iterator of scream. */ +} re_scream_pos_data; + +/* 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, I32 minend, SV* screamer, + void* data, U32 flags); + char* (*intuit) (pTHX_ REGEXP * const rx, SV *sv, char *strpos, + char *strend, const U32 flags, + re_scream_pos_data *data); + SV* (*checkstr) (pTHX_ REGEXP * const rx); + void (*free) (pTHX_ REGEXP * const rx); + void (*numbered_buff_FETCH) (pTHX_ REGEXP * const rx, const I32 paren, + SV * const sv); + void (*numbered_buff_STORE) (pTHX_ REGEXP * const rx, const I32 paren, + SV const * const value); + I32 (*numbered_buff_LENGTH) (pTHX_ REGEXP * const rx, const SV * const sv, + const I32 paren); + 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, + const U32 flags); + SV* (*qr_package)(pTHX_ REGEXP * const rx); +#ifdef USE_ITHREADS + void* (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param); +#endif +} regexp_engine; + +/* + These are passed to the numbered capture variable callbacks as the + paren name. >= 1 is reserved for actual numbered captures, i.e. $1, + $2 etc. +*/ +#define RX_BUFF_IDX_PREMATCH -2 /* $` / ${^PREMATCH} */ +#define RX_BUFF_IDX_POSTMATCH -1 /* $' / ${^POSTMATCH} */ +#define RX_BUFF_IDX_FULLMATCH 0 /* $& / ${^MATCH} */ + +/* + 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 + +/* +=head1 REGEXP Functions + +=for apidoc Am|REGEXP *|SvRX|SV *sv + +Convenience macro to get the REGEXP from a SV. This is approximately +equivalent to the following snippet: + + if (SvMAGICAL(sv)) + mg_get(sv); + if (SvROK(sv) && + (tmpsv = (SV*)SvRV(sv)) && + SvTYPE(tmpsv) == SVt_PVMG && + (tmpmg = mg_find(tmpsv, PERL_MAGIC_qr))) + { + return (REGEXP *)tmpmg->mg_obj; + } + +NULL will be returned if a REGEXP* is not found. + +=for apidoc Am|bool|SvRXOK|SV* sv + +Returns a boolean indicating whether the SV contains qr magic +(PERL_MAGIC_qr). + +If you want to do something with the REGEXP* later use SvRX instead +and check for NULL. + +=cut +*/ + +#define SvRX(sv) (Perl_get_re_arg(aTHX_ sv)) +#define SvRXOK(sv) (Perl_get_re_arg(aTHX_ sv) ? TRUE : FALSE) + + +/* Flags stored in regexp->extflags + * These are used by code external to the regexp engine + * + * Note that flags starting with RXf_PMf_ have exact equivalents + * stored in op_pmflags and which are defined in op.h, they are defined + * numerically here only for clarity. + * + * NOTE: if you modify any RXf flags you should run regen.pl or regcomp.pl + * so that regnodes.h is updated with the changes. + * + */ + +/* 0x3F of extflags is used by (RXf_)PMf_COMPILETIME + * If you change these you need to change the equivalent flags in op.h, and + * vice versa. */ +#define RXf_PMf_MULTILINE 0x00000001 /* /m */ +#define RXf_PMf_SINGLELINE 0x00000002 /* /s */ +#define RXf_PMf_FOLD 0x00000004 /* /i */ +#define RXf_PMf_EXTENDED 0x00000008 /* /x */ +#define RXf_PMf_KEEPCOPY 0x00000010 /* /p */ +#define RXf_PMf_LOCALE 0x00000020 /* use locale */ +/* these flags are transfered from the PMOP->op_pmflags member during compilation */ +#define RXf_PMf_STD_PMMOD_SHIFT 0 +#define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED) +#define RXf_PMf_COMPILETIME (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_LOCALE|RXf_PMf_FOLD|RXf_PMf_EXTENDED|RXf_PMf_KEEPCOPY) + +#define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl) \ + 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: *(pmfl) |= RXf_PMf_EXTENDED; break + +/* chars and strings used as regex pattern modifiers + * Singlular 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 EXEC_PAT_MOD 'e' +#define KEEPCOPY_PAT_MOD 'p' +#define ONCE_PAT_MOD 'o' +#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 ONCE_PAT_MODS "o" +#define KEEPCOPY_PAT_MODS "p" +#define EXEC_PAT_MODS "e" +#define LOOP_PAT_MODS "gc" + +#define STD_PAT_MODS "msix" + +#define INT_PAT_MODS STD_PAT_MODS KEEPCOPY_PAT_MODS + +#define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS +#define QR_PAT_MODS STD_PAT_MODS EXT_PAT_MODS +#define M_PAT_MODS QR_PAT_MODS LOOP_PAT_MODS +#define S_PAT_MODS M_PAT_MODS EXEC_PAT_MODS + +/* + * NOTE: if you modify any RXf flags you should run regen.pl or regcomp.pl + * so that regnodes.h is updated with the changes. + * + */ + +/* Anchor and GPOS related stuff */ +#define RXf_ANCH_BOL 0x00000100 +#define RXf_ANCH_MBOL 0x00000200 +#define RXf_ANCH_SBOL 0x00000400 +#define RXf_ANCH_GPOS 0x00000800 +#define RXf_GPOS_SEEN 0x00001000 +#define RXf_GPOS_FLOAT 0x00002000 +/* two bits here */ +#define RXf_ANCH (RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL) +#define RXf_GPOS_CHECK (RXf_GPOS_SEEN|RXf_ANCH_GPOS) +#define RXf_ANCH_SINGLE (RXf_ANCH_SBOL|RXf_ANCH_GPOS) + +/* What we have seen */ +#define RXf_LOOKBEHIND_SEEN 0x00004000 +#define RXf_EVAL_SEEN 0x00008000 +#define RXf_CANY_SEEN 0x00010000 + +/* Special */ +#define RXf_NOSCAN 0x00020000 +#define RXf_CHECK_ALL 0x00040000 + +/* UTF8 related */ +#define RXf_MATCH_UTF8 0x00100000 + +/* Intuit related */ +#define RXf_USE_INTUIT_NOML 0x00200000 +#define RXf_USE_INTUIT_ML 0x00400000 +#define RXf_INTUIT_TAIL 0x00800000 + +/* + Set in Perl_pmruntime if op_flags & OPf_SPECIAL, i.e. split. Will + be used by regex engines to check whether they should set + RXf_SKIPWHITE +*/ +#define RXf_SPLIT 0x01000000 + +#define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML) + +/* Copy and tainted info */ +#define RXf_COPY_DONE 0x02000000 +#define RXf_TAINTED_SEEN 0x04000000 +#define RXf_TAINTED 0x08000000 /* this pattern is tainted */ + +/* Flags indicating special patterns */ +#define RXf_START_ONLY 0x10000000 /* Pattern is /^/ */ +#define RXf_SKIPWHITE 0x20000000 /* Pattern is for a split / / */ +#define RXf_WHITE 0x40000000 /* Pattern is /\s+/ */ +#define RXf_NULL 0x80000000 /* Pattern is // */ + +/* + * NOTE: if you modify any RXf flags you should run regen.pl or regcomp.pl + * so that regnodes.h is updated with the changes. + * + */ + +#define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN) +#define RXp_MATCH_TAINTED(prog) (RXp_EXTFLAGS(prog) & RXf_TAINTED_SEEN) +#define RX_MATCH_TAINTED(prog) (RX_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_MATCH_TAINTED_set(prog, t) ((t) \ ? RX_MATCH_TAINTED_on(prog) \ : RX_MATCH_TAINTED_off(prog)) -#define RX_MATCH_COPIED(prog) ((prog)->reganch & ROPT_COPY_DONE) -#define RX_MATCH_COPIED_on(prog) ((prog)->reganch |= ROPT_COPY_DONE) -#define RX_MATCH_COPIED_off(prog) ((prog)->reganch &= ~ROPT_COPY_DONE) +#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) + +/* For source compatibility. We used to store these explicitly. */ +#define RX_PRECOMP(prog) (RX_WRAPPED(prog) + ((struct regexp *)SvANY(prog))->pre_prefix) +#define RX_PRECOMP_const(prog) (RX_WRAPPED_const(prog) + ((struct regexp *)SvANY(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? */ +#define RX_PRELEN(prog) (RX_WRAPLEN(prog) - ((struct regexp *)SvANY(prog))->pre_prefix - 1) +#define RX_WRAPPED(prog) SvPVX(prog) +#define RX_WRAPPED_const(prog) SvPVX_const(prog) +#define RX_WRAPLEN(prog) SvCUR(prog) +#define RX_CHECK_SUBSTR(prog) (((struct regexp *)SvANY(prog))->check_substr) +#define RX_REFCNT(prog) SvREFCNT(prog) +#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) +# define RX_EXTFLAGS(prog) \ + (*({ \ + const REGEXP *const _rx_extflags = (prog); \ + assert(SvTYPE(_rx_extflags) == SVt_REGEXP); \ + &RXp_EXTFLAGS(SvANY(_rx_extflags)); \ + })) +# define RX_ENGINE(prog) \ + (*({ \ + const REGEXP *const _rx_engine = (prog); \ + assert(SvTYPE(_rx_engine) == SVt_REGEXP); \ + &SvANY(_rx_engine)->engine; \ + })) +# define RX_SUBBEG(prog) \ + (*({ \ + const REGEXP *const _rx_subbeg = (prog); \ + assert(SvTYPE(_rx_subbeg) == SVt_REGEXP); \ + &SvANY(_rx_subbeg)->subbeg; \ + })) +# define RX_OFFS(prog) \ + (*({ \ + const REGEXP *const _rx_offs = (prog); \ + assert(SvTYPE(_rx_offs) == SVt_REGEXP); \ + &SvANY(_rx_offs)->offs; \ + })) +# define RX_NPARENS(prog) \ + (*({ \ + const REGEXP *const _rx_nparens = (prog); \ + assert(SvTYPE(_rx_nparens) == SVt_REGEXP); \ + &SvANY(_rx_nparens)->nparens; \ + })) +#else +# define RX_EXTFLAGS(prog) RXp_EXTFLAGS((struct regexp *)SvANY(prog)) +# define RX_ENGINE(prog) (((struct regexp *)SvANY(prog))->engine) +# define RX_SUBBEG(prog) (((struct regexp *)SvANY(prog))->subbeg) +# define RX_OFFS(prog) (((struct regexp *)SvANY(prog))->offs) +# define RX_NPARENS(prog) (((struct regexp *)SvANY(prog))->nparens) +#endif +#define RX_SUBLEN(prog) (((struct regexp *)SvANY(prog))->sublen) +#define RX_MINLEN(prog) (((struct regexp *)SvANY(prog))->minlen) +#define RX_MINLENRET(prog) (((struct regexp *)SvANY(prog))->minlenret) +#define RX_GOFS(prog) (((struct regexp *)SvANY(prog))->gofs) +#define RX_LASTPAREN(prog) (((struct regexp *)SvANY(prog))->lastparen) +#define RX_LASTCLOSEPAREN(prog) (((struct regexp *)SvANY(prog))->lastcloseparen) +#define RX_SEEN_EVALS(prog) (((struct regexp *)SvANY(prog))->seen_evals) +#define RX_SAVED_COPY(prog) (((struct regexp *)SvANY(prog))->saved_copy) + +#endif /* PLUGGABLE_RE_EXTENSION */ + +/* Stuff that needs to be included in the plugable extension goes below here */ + #ifdef PERL_OLD_COPY_ON_WRITE #define RX_MATCH_COPY_FREE(rx) \ - STMT_START {if (rx->saved_copy) { \ - SV_CHECK_THINKFIRST_COW_DROP(rx->saved_copy); \ + STMT_START {if (RX_SAVED_COPY(rx)) { \ + SV_CHECK_THINKFIRST_COW_DROP(RX_SAVED_COPY(rx)); \ } \ if (RX_MATCH_COPIED(rx)) { \ - Safefree(rx->subbeg); \ + Safefree(RX_SUBBEG(rx)); \ RX_MATCH_COPIED_off(rx); \ }} STMT_END #else #define RX_MATCH_COPY_FREE(rx) \ STMT_START {if (RX_MATCH_COPIED(rx)) { \ - Safefree(rx->subbeg); \ + Safefree(RX_SUBBEG(rx)); \ RX_MATCH_COPIED_off(rx); \ }} STMT_END #endif -#define RX_MATCH_UTF8(prog) ((prog)->reganch & ROPT_MATCH_UTF8) -#define RX_MATCH_UTF8_on(prog) ((prog)->reganch |= ROPT_MATCH_UTF8) -#define RX_MATCH_UTF8_off(prog) ((prog)->reganch &= ~ROPT_MATCH_UTF8) +#define RXp_MATCH_UTF8(prog) (RXp_EXTFLAGS(prog) & RXf_MATCH_UTF8) +#define RX_MATCH_UTF8(prog) (RX_EXTFLAGS(prog) & RXf_MATCH_UTF8) +#define RX_MATCH_UTF8_on(prog) (RX_EXTFLAGS(prog) |= RXf_MATCH_UTF8) +#define RX_MATCH_UTF8_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_MATCH_UTF8) #define RX_MATCH_UTF8_set(prog, t) ((t) \ ? (RX_MATCH_UTF8_on(prog), (PL_reg_match_utf8 = 1)) \ : (RX_MATCH_UTF8_off(prog), (PL_reg_match_utf8 = 0))) - + +/* Whether the pattern stored at RX_WRAPPED is in UTF-8 */ +#define RX_UTF8(prog) SvUTF8(prog) + #define REXEC_COPY_STR 0x01 /* Need to copy the string. */ #define REXEC_CHECKED 0x02 /* check_substr already checked. */ #define REXEC_SCREAM 0x04 /* use scream table. */ #define REXEC_IGNOREPOS 0x08 /* \G matches at start. */ #define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */ -#define ReREFCNT_inc(re) ((void)(re && re->refcnt++), re) -#define ReREFCNT_dec(re) CALLREGFREE(aTHX_ re) +#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) +# define ReREFCNT_inc(re) \ + ({ \ + /* This is here to generate a casting warning if incorrect. */ \ + REGEXP *const _rerefcnt_inc = (re); \ + assert(SvTYPE(_rerefcnt_inc) == SVt_REGEXP); \ + SvREFCNT_inc(_rerefcnt_inc); \ + _rerefcnt_inc; \ + }) +# define ReREFCNT_dec(re) \ + ({ \ + /* This is here to generate a casting warning if incorrect. */ \ + REGEXP *const _rerefcnt_dec = (re); \ + SvREFCNT_dec(_rerefcnt_dec); \ + }) +#else +# define ReREFCNT_dec(re) SvREFCNT_dec(re) +# define ReREFCNT_inc(re) ((REGEXP *) SvREFCNT_inc(re)) +#endif + +/* FIXME for plugins. */ #define FBMcf_TAIL_DOLLAR 1 #define FBMcf_TAIL_DOLLARM 2 @@ -145,8 +490,6 @@ typedef struct regexp { #define FBMrf_MULTILINE 1 -struct re_scream_pos_data_s; - /* an accepting state/position*/ struct _reg_trie_accepted { U8 *endpos; @@ -158,59 +501,26 @@ typedef struct _reg_trie_accepted reg_trie_accepted; * Perl_regexec_flags and then passed to regtry(), regmatch() etc */ typedef struct { - regexp *prog; + REGEXP *prog; char *bol; char *till; SV *sv; char *ganch; + char *cutpoint; } regmatch_info; /* structures for holding and saving the state maintained by regmatch() */ -typedef I32 CHECKPOINT; - -typedef enum { - resume_TRIE1, - resume_TRIE2, - resume_EVAL, - resume_CURLYX, - resume_WHILEM1, - resume_WHILEM2, - resume_WHILEM3, - resume_WHILEM4, - resume_WHILEM5, - resume_WHILEM6, - resume_CURLYM1, - resume_CURLYM2, - resume_CURLYM3, - resume_IFMATCH, - resume_PLUS1, - resume_PLUS2, - resume_PLUS3, - resume_PLUS4 -} regmatch_resume_states; +#ifndef MAX_RECURSE_EVAL_NOCHANGE_DEPTH +#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 1000 +#endif +typedef I32 CHECKPOINT; typedef struct regmatch_state { - - /* these vars contain state that needs to be maintained - * across the main while loop ... */ - - regmatch_resume_states resume_state; /* where to jump to on return */ - regnode *scan; /* Current node. */ - regnode *next; /* Next node. */ - bool minmod; /* the next "{n,m}" is a "{n,m}?" */ - bool sw; /* the condition value in (?(cond)a|b) */ - int logical; - I32 unwind; /* savestack index of current unwind block */ - struct regmatch_state *cc; /* current innermost curly state */ - char *locinput; - - /* ... while the rest of these are local to an individual branch */ - - I32 n; /* no or next */ - I32 ln; /* len or last */ + int resume_state; /* where to jump to on return */ + char *locinput; /* where to backtrack in string on failure */ union { @@ -221,69 +531,128 @@ typedef struct regmatch_state { struct regmatch_state *prev_yes_state; } yes; + /* branchlike members */ + /* this is a fake union member that matches the first elements + * of each member that needs to behave like a branch */ + struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + U32 lastparen; + CHECKPOINT cp; + + } branchlike; + + struct { + /* the first elements must match u.branchlike */ + struct regmatch_state *prev_yes_state; + U32 lastparen; + CHECKPOINT cp; + + regnode *next_branch; /* next branch node */ + } branch; + struct { - reg_trie_accepted *accept_buff; - U32 accepted; /* how many accepting states we have seen */ + /* the first elements must match u.branchlike */ + struct regmatch_state *prev_yes_state; + U32 lastparen; + CHECKPOINT cp; + + reg_trie_accepted *accept_buff; /* accepting states we have seen */ + U32 accepted; /* how many accepting states we have seen */ + U16 *jump; /* positive offsets from me */ + regnode *B; /* node following the trie */ + regnode *me; /* Which node am I - needed for jump tries*/ } trie; + /* special types - these members are used to store state for special + regops like eval, if/then, lookaround and the markpoint state */ struct { /* this first element must match u.yes */ struct regmatch_state *prev_yes_state; - regexp *prev_rex; - int toggleutf; + struct regmatch_state *prev_eval; + struct regmatch_state *prev_curlyx; + REGEXP *prev_rex; + U32 toggle_reg_flags; /* what bits in PL_reg_flags to + flip when transitioning between + inner and outer rexen */ CHECKPOINT cp; /* remember current savestack indexes */ CHECKPOINT lastcp; + U32 close_paren; /* which close bracket is our end */ + regnode *B; /* the node following us */ } eval; struct { - CHECKPOINT cp; /* remember current savestack indexes */ - struct regmatch_state *outercc; /* outer CURLYX state if any */ + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + I32 wanted; + I32 logical; /* saved copy of 'logical' var */ + regnode *me; /* the IFMATCH/SUSPEND/UNLESSM node */ + } ifmatch; /* and SUSPEND/UNLESSM */ + + struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + struct regmatch_state *prev_mark; + SV* mark_name; + char *mark_loc; + } mark; + + struct { + int val; + } keeper; - /* these contain the current curly state, and are accessed - * by subsequent WHILEMs */ + /* quantifiers - these members are used for storing state for + for the regops used to implement quantifiers */ + struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + struct regmatch_state *prev_curlyx; /* previous cur_curlyx */ + regnode *A, *B; /* the nodes corresponding to /A*B/ */ + CHECKPOINT cp; /* remember current savestack index */ + bool minmod; int parenfloor;/* how far back to strip paren data */ - int cur; /* how many instances of scan we've matched */ - int min; /* the minimal number of scans to match */ - int max; /* the maximal number of scans to match */ - regnode * scan; /* the thing to match */ - char * lastloc;/* where we started matching this scan */ + int min; /* the minimal number of A's to match */ + int max; /* the maximal number of A's to match */ + + /* these two are modified by WHILEM */ + int count; /* how many instances of A we've matched */ + char *lastloc;/* where previous A matched (0-len detect) */ } curlyx; struct { - CHECKPOINT cp; /* remember current savestack indexes */ - CHECKPOINT lastcp; - struct regmatch_state *savecc; - char *lastloc; /* Detection of 0-len. */ - I32 cache_offset; - I32 cache_bit; + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + struct regmatch_state *save_curlyx; + CHECKPOINT cp; /* remember current savestack indexes */ + CHECKPOINT lastcp; + char *save_lastloc; /* previous curlyx.lastloc */ + I32 cache_offset; + I32 cache_mask; } whilem; struct { /* this first element must match u.yes */ struct regmatch_state *prev_yes_state; - I32 paren; I32 c1, c2; /* case fold search */ - CHECKPOINT lastcp; - I32 l; - I32 matches; - I32 maxwanted; + CHECKPOINT cp; + I32 alen; /* length of first-matched A string */ + I32 count; bool minmod; + regnode *A, *B; /* the nodes corresponding to /A*B/ */ + regnode *me; /* the curlym node */ } curlym; struct { - I32 paren; - CHECKPOINT lastcp; + U32 paren; + CHECKPOINT cp; I32 c1, c2; /* case fold search */ - char *e; - char *old; + char *maxpos; /* highest possible point in string to match */ + char *oldloc; /* the previous locinput */ int count; - } plus; /* and CURLYN/CURLY/STAR */ + int min, max; /* {m,n} */ + regnode *A, *B; /* the nodes corresponding to /A*B/ */ + } curly; /* and CURLYN/PLUS/STAR */ - struct { - /* this first element must match u.yes */ - struct regmatch_state *prev_yes_state; - I32 wanted; - } ifmatch; /* and SUSPEND/UNLESSM */ } u; } regmatch_state; @@ -303,14 +672,12 @@ typedef struct regmatch_slab { #define PL_bostr PL_reg_state.re_state_bostr #define PL_reginput PL_reg_state.re_state_reginput #define PL_regeol PL_reg_state.re_state_regeol -#define PL_regstartp PL_reg_state.re_state_regstartp -#define PL_regendp PL_reg_state.re_state_regendp +#define PL_regoffs PL_reg_state.re_state_regoffs #define PL_reglastparen PL_reg_state.re_state_reglastparen #define PL_reglastcloseparen PL_reg_state.re_state_reglastcloseparen #define PL_reg_start_tmp PL_reg_state.re_state_reg_start_tmp #define PL_reg_start_tmpl PL_reg_state.re_state_reg_start_tmpl #define PL_reg_eval_set PL_reg_state.re_state_reg_eval_set -#define PL_regindent PL_reg_state.re_state_regindent #define PL_reg_match_utf8 PL_reg_state.re_state_reg_match_utf8 #define PL_reg_magic PL_reg_state.re_state_reg_magic #define PL_reg_oldpos PL_reg_state.re_state_reg_oldpos @@ -328,29 +695,27 @@ typedef struct regmatch_slab { struct re_save_state { U32 re_state_reg_flags; /* from regexec.c */ + U32 re_state_reg_start_tmpl; /* from regexec.c */ + I32 re_state_reg_eval_set; /* from regexec.c */ + bool re_state_reg_match_utf8; /* from regexec.c */ char *re_state_bostr; char *re_state_reginput; /* String-input pointer. */ char *re_state_regeol; /* End of input, for $ check. */ - I32 *re_state_regstartp; /* Pointer to startp array. */ - I32 *re_state_regendp; /* Ditto for endp. */ + regexp_paren_pair *re_state_regoffs; /* Pointer to start/end pairs */ U32 *re_state_reglastparen; /* Similarly for lastparen. */ U32 *re_state_reglastcloseparen; /* Similarly for lastcloseparen. */ char **re_state_reg_start_tmp; /* from regexec.c */ - U32 re_state_reg_start_tmpl; /* from regexec.c */ - I32 re_state_reg_eval_set; /* from regexec.c */ - int re_state_regindent; /* from regexec.c */ - bool re_state_reg_match_utf8; /* from regexec.c */ MAGIC *re_state_reg_magic; /* from regexec.c */ - I32 re_state_reg_oldpos; /* from regexec.c */ PMOP *re_state_reg_oldcurpm; /* from regexec.c */ PMOP *re_state_reg_curpm; /* from regexec.c */ char *re_state_reg_oldsaved; /* old saved substr during match */ STRLEN re_state_reg_oldsavedlen; /* old length of saved substr during match */ + STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */ + I32 re_state_reg_oldpos; /* from regexec.c */ I32 re_state_reg_maxiter; /* max wait until caching pos */ I32 re_state_reg_leftiter; /* wait until caching pos */ + U32 re_state_regsize; /* from regexec.c */ char *re_state_reg_poscache; /* cache of pos of WHILEM */ - STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */ - I32 re_state_regsize; /* from regexec.c */ char *re_state_reg_starttry; /* from regexec.c */ #ifdef PERL_OLD_COPY_ON_WRITE SV *re_state_nrs; /* was placeholder: unused since 5.8.0 (5.7.2 patch #12027 for bug ID 20010815.012). Used to save rx->saved_copy */ @@ -359,6 +724,7 @@ struct re_save_state { #define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \ (1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack))) + /* * Local variables: * c-indentation-style: bsd