X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/07bc277f32c1d7aff237dd3f55d558b5d4b93314..406871851914a339b834e2e26f758de1e24104bd:/regexp.h diff --git a/regexp.h b/regexp.h index 17dfbb6..f87031c 100644 --- a/regexp.h +++ b/regexp.h @@ -1,7 +1,7 @@ /* regexp.h * * Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003, - * 2005, 2006, 2007, 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. @@ -66,49 +66,49 @@ typedef struct regexp_paren_pair { 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? */ \ + HV *paren_names; /* Optional hash of paren names */ \ + /* 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 { - /* what engine created this regexp? */ - const struct regexp_engine* engine; - struct 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 */ - struct reg_substr_data *substrs; /* substring data about strings that must appear - in the final match, used for optimisations */ - 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; /* Swap copy of *offs */ - regexp_paren_pair *offs; /* Array of offsets for (@-) and (@+) */ - - char *subbeg; /* saved or original string - so \digit works forever. */ - 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 */ - /* wrapped can't be const char*, as it is returned by sv_2pv_flags */ - char *wrapped; /* wrapped version of the pattern */ - I32 wraplen; /* length of wrapped */ - unsigned pre_prefix:4; /* offset from wrapped to the start of precomp */ - unsigned seen_evals:28; /* number of eval groups in the pattern - for security checks */ - HV *paren_names; /* Optional hash of paren names */ - - /* Refcount of this regexp */ - I32 refcnt; /* Refcount of this regexp */ + _XPV_HEAD; + _REGEXP_COMMON; } regexp; +#define RXp_PAREN_NAMES(rx) ((rx)->paren_names) + /* used for high speed searches */ typedef struct re_scream_pos_data_s { @@ -120,7 +120,7 @@ typedef struct re_scream_pos_data_s * Any regex engine implementation must be able to build one of these. */ typedef struct regexp_engine { - REGEXP* (*comp) (pTHX_ const SV * const pattern, const U32 flags); + 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); @@ -178,7 +178,7 @@ typedef struct regexp_engine { /* Whether this is being called from a re:: function */ #define RXapif_REGNAME 0x0400 #define RXapif_REGNAMES 0x0800 -#define RXapif_REGNAMES_COUNT 0x1000 +#define RXapif_REGNAMES_COUNT 0x1000 /* =head1 REGEXP Functions @@ -215,7 +215,7 @@ and check for NULL. #define SvRXOK(sv) (Perl_get_re_arg(aTHX_ sv) ? TRUE : FALSE) -/* Flags stored in regexp->extflags +/* 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 @@ -223,7 +223,7 @@ and check for NULL. * 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. + * so that regnodes.h is updated with the changes. * */ @@ -263,11 +263,13 @@ and check for NULL. #define SINGLE_PAT_MOD 's' #define IGNORE_PAT_MOD 'i' #define XTENDED_PAT_MOD 'x' +#define NONDESTRUCT_PAT_MOD 'r' #define ONCE_PAT_MODS "o" #define KEEPCOPY_PAT_MODS "p" #define EXEC_PAT_MODS "e" #define LOOP_PAT_MODS "gc" +#define NONDESTRUCT_PAT_MODS "r" #define STD_PAT_MODS "msix" @@ -276,11 +278,11 @@ and check for NULL. #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 +#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 regcomp.pl - * so that regnodes.h is updated with the changes. + * so that regnodes.h is updated with the changes. * */ @@ -306,7 +308,6 @@ and check for NULL. #define RXf_CHECK_ALL 0x00040000 /* UTF8 related */ -#define RXf_UTF8 0x00080000 #define RXf_MATCH_UTF8 0x00100000 /* Intuit related */ @@ -336,7 +337,7 @@ and check for NULL. /* * NOTE: if you modify any RXf flags you should run regen.pl or regcomp.pl - * so that regnodes.h is updated with the changes. + * so that regnodes.h is updated with the changes. * */ @@ -359,35 +360,66 @@ and check for NULL. ? RX_MATCH_COPIED_on(prog) \ : RX_MATCH_COPIED_off(prog)) -#define RXp_PRECOMP(rx) ((rx)->wrapped + (rx)->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 RXp_PRELEN(rx) ((rx)->wraplen - (rx)->pre_prefix - 1) -#define RXp_WRAPPED(rx) ((rx)->wrapped) -#define RXp_WRAPLEN(rx) ((rx)->wraplen) #define RXp_EXTFLAGS(rx) ((rx)->extflags) /* For source compatibility. We used to store these explicitly. */ -#define RX_PRECOMP(prog) ((prog)->wrapped + (prog)->pre_prefix) -#define RX_PRELEN(prog) ((prog)->wraplen - (prog)->pre_prefix - 1) -#define RX_WRAPPED(prog) ((prog)->wrapped) -#define RX_WRAPLEN(prog) ((prog)->wraplen) -#define RX_CHECK_SUBSTR(prog) ((prog)->check_substr) -#define RX_EXTFLAGS(prog) ((prog)->extflags) -#define RX_REFCNT(prog) ((prog)->refcnt) -#define RX_ENGINE(prog) ((prog)->engine) -#define RX_SUBBEG(prog) ((prog)->subbeg) -#define RX_OFFS(prog) ((prog)->offs) -#define RX_NPARENS(prog) ((prog)->nparens) -#define RX_SUBLEN(prog) ((prog)->sublen) -#define RX_SUBBEG(prog) ((prog)->subbeg) -#define RX_MINLEN(prog) ((prog)->minlen) -#define RX_MINLENRET(prog) ((prog)->minlenret) -#define RX_GOFS(prog) ((prog)->gofs) -#define RX_LASTPAREN(prog) ((prog)->lastparen) -#define RX_LASTCLOSEPAREN(prog) ((prog)->lastcloseparen) -#define RX_SEEN_EVALS(prog) ((prog)->seen_evals) +#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 */ @@ -395,8 +427,8 @@ and check for NULL. #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(rx)); \ @@ -417,15 +449,37 @@ and check for NULL. #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(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 @@ -435,18 +489,11 @@ and check for NULL. #define FBMrf_MULTILINE 1 -/* an accepting state/position*/ -struct _reg_trie_accepted { - U8 *endpos; - U16 wordnum; -}; -typedef struct _reg_trie_accepted reg_trie_accepted; - /* some basic information about the current match that is created by * Perl_regexec_flags and then passed to regtry(), regmatch() etc */ typedef struct { - regexp *prog; + REGEXP *prog; char *bol; char *till; SV *sv; @@ -502,11 +549,15 @@ typedef struct regmatch_state { U32 lastparen; CHECKPOINT cp; - reg_trie_accepted *accept_buff; /* accepting states we have seen */ - U32 accepted; /* how many accepting states we have seen */ + U32 accepted; /* how many accepting states left */ U16 *jump; /* positive offsets from me */ regnode *B; /* node following the trie */ regnode *me; /* Which node am I - needed for jump tries*/ + U8 *firstpos;/* pos in string of first trie match */ + U32 firstchars;/* len in chars of firstpos from start */ + U16 nextword;/* next word to try */ + U16 topword; /* longest accepted word */ + bool longfold;/* saw a fold with a 1->n char mapping */ } trie; /* special types - these members are used to store state for special @@ -516,7 +567,7 @@ typedef struct regmatch_state { struct regmatch_state *prev_yes_state; struct regmatch_state *prev_eval; struct regmatch_state *prev_curlyx; - regexp *prev_rex; + REGEXP *prev_rex; U32 toggle_reg_flags; /* what bits in PL_reg_flags to flip when transitioning between inner and outer rexen */ @@ -552,12 +603,11 @@ typedef struct regmatch_state { /* 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/ */ + regnode *me; /* the CURLYX node */ + regnode *B; /* the B node in /A*B/ */ CHECKPOINT cp; /* remember current savestack index */ bool minmod; int parenfloor;/* how far back to strip paren data */ - 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 */