X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexp.h;h=a2e0fc1a31cf76d191846b659c84adb3c0f2a0aa;hb=798b63bc924a07589315b3229311582adce06136;hp=4ea7c694e97e74225520efd9bcee88759e463ef6;hpb=be8e71aa96b3d344129617fadcaca7d958882caa;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regexp.h b/regexp.h index 4ea7c69..a2e0fc1 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 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. @@ -15,8 +15,8 @@ * not the System V one. */ #ifndef PLUGGABLE_RE_EXTENSION -/* we don't want to include this stuff if we are inside Nicholas' - * pluggable regex engine code */ +/* 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; @@ -30,90 +30,176 @@ struct reg_substr_data; struct reg_data; +struct regexp_engine; + +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 +#define SV_SAVED_COPY SV *saved_copy; /* If non-NULL, SV which is COW from original */ +#else +#define SV_SAVED_COPY +#endif +/* this is ordered such that the most commonly used + fields are at the start of the struct */ 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. */ + /* what engine created this regexp? */ + const struct regexp_engine* engine; + + /* 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 */ + I32 *startp; /* Array of offsets from start of string (@-) */ + I32 *endp; /* Array of offsets from start of string (@+) */ char *subbeg; /* saved or original string so \digit works forever. */ -#ifdef PERL_OLD_COPY_ON_WRITE - SV *saved_copy; /* If non-NULL, SV which is COW from original */ -#endif - U32 *offsets; /* offset annotations 20001228 MJD */ I32 sublen; /* Length of string pointed by subbeg */ - I32 refcnt; - I32 minlen; /* mininum possible length of $& */ + SV_SAVED_COPY /* If non-NULL, SV which is COW from original */ + + + /* Information about the match that isn't often used */ + char *precomp; /* pre-compilation regular expression */ 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. */ + I32 seen_evals; /* 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 */ } 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 0x00000001 -#define ROPT_ANCH_MBOL 0x00000002 -#define ROPT_ANCH_SBOL 0x00000004 -#define ROPT_ANCH_GPOS 0x00000008 -#define ROPT_SKIP 0x00000010 -#define ROPT_IMPLICIT 0x00000020 /* Converted .* to ^.* */ -#define ROPT_NOSCAN 0x00000040 /* Check-string always at start. */ -#define ROPT_GPOS_SEEN 0x00000080 -#define ROPT_CHECK_ALL 0x00000100 -#define ROPT_LOOKBEHIND_SEEN 0x00000200 -#define ROPT_EVAL_SEEN 0x00000400 -#define ROPT_CANY_SEEN 0x00000800 -#define ROPT_SANY_SEEN ROPT_CANY_SEEN /* src bckwrd cmpt */ - -/* 0xf800 of reganch is used by PMf_COMPILETIME */ - -#define ROPT_UTF8 0x00010000 -#define ROPT_NAUGHTY 0x00020000 /* how exponential is this pattern? */ -#define ROPT_COPY_DONE 0x00040000 /* subbeg is a copy of the string */ -#define ROPT_TAINTED_SEEN 0x00080000 -#define ROPT_MATCH_UTF8 0x10000000 /* subbeg is utf-8 */ - -#define RE_USE_INTUIT_NOML 0x00100000 /* Best to intuit before matching */ -#define RE_USE_INTUIT_ML 0x00200000 -#define REINT_AUTORITATIVE_NOML 0x00400000 /* Can trust a positive answer */ -#define REINT_AUTORITATIVE_ML 0x00800000 -#define REINT_ONCE_NOML 0x01000000 /* Intuit can succed once only. */ -#define REINT_ONCE_ML 0x02000000 -#define RE_INTUIT_ONECHAR 0x04000000 -#define RE_INTUIT_TAIL 0x08000000 - - -#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) +/* 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_ char* exp, char* xend, PMOP* pm); + I32 (*exec) (pTHX_ regexp* prog, char* stringarg, char* strend, + char* strbeg, I32 minend, SV* screamer, + void* data, U32 flags); + char* (*intuit) (pTHX_ regexp *prog, SV *sv, char *strpos, + char *strend, U32 flags, + struct re_scream_pos_data_s *data); + SV* (*checkstr) (pTHX_ regexp *prog); + void (*free) (pTHX_ struct regexp* r); +#ifdef USE_ITHREADS + void* (*dupe) (pTHX_ const regexp *r, CLONE_PARAMS *param); +#endif +} regexp_engine; + +/* 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. + */ + +/* Anchor and GPOS related stuff */ +#define RXf_ANCH_BOL 0x00000001 +#define RXf_ANCH_MBOL 0x00000002 +#define RXf_ANCH_SBOL 0x00000004 +#define RXf_ANCH_GPOS 0x00000008 +#define RXf_GPOS_SEEN 0x00000010 +#define RXf_GPOS_FLOAT 0x00000020 +/* 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) + +/* Flags indicating special patterns */ +#define RXf_START_ONLY 0x00000200 /* Pattern is /^/ */ +#define RXf_WHITE 0x00000400 /* Pattern is /\s+/ */ + +/* 0xF800 of extflags is used by (RXf_)PMf_COMPILETIME */ +#define RXf_PMf_LOCALE 0x00000800 /* use locale */ +#define RXf_PMf_MULTILINE 0x00001000 /* /m */ +#define RXf_PMf_SINGLELINE 0x00002000 /* /s */ +#define RXf_PMf_FOLD 0x00004000 /* /i */ +#define RXf_PMf_EXTENDED 0x00008000 /* /x */ +#define RXf_PMf_KEEPCOPY 0x00010000 /* /k */ +/* these flags are transfered from the PMOP->op_pmflags member during compilation */ +#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 + +/* What we have seen */ +#define RXf_LOOKBEHIND_SEEN 0x00020000 +#define RXf_EVAL_SEEN 0x00040000 +#define RXf_CANY_SEEN 0x00080000 + +/* Special */ +#define RXf_NOSCAN 0x00100000 +#define RXf_CHECK_ALL 0x00200000 + +/* UTF8 related */ +#define RXf_UTF8 0x00400000 +#define RXf_MATCH_UTF8 0x00800000 + +/* Intuit related */ +#define RXf_USE_INTUIT_NOML 0x01000000 +#define RXf_USE_INTUIT_ML 0x02000000 +#define RXf_INTUIT_TAIL 0x04000000 +/* one bit here */ +#define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML) + +/* Copy and tainted info */ +#define RXf_COPY_DONE 0x10000000 +#define RXf_TAINTED_SEEN 0x20000000 +/* two bits here */ + + +#define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN) +#define RX_MATCH_TAINTED(prog) ((prog)->extflags & RXf_TAINTED_SEEN) +#define RX_MATCH_TAINTED_on(prog) ((prog)->extflags |= RXf_TAINTED_SEEN) +#define RX_MATCH_TAINTED_off(prog) ((prog)->extflags &= ~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 RX_MATCH_COPIED(prog) ((prog)->extflags & RXf_COPY_DONE) +#define RX_MATCH_COPIED_on(prog) ((prog)->extflags |= RXf_COPY_DONE) +#define RX_MATCH_COPIED_off(prog) ((prog)->extflags &= ~RXf_COPY_DONE) #define RX_MATCH_COPIED_set(prog,t) ((t) \ ? RX_MATCH_COPIED_on(prog) \ : RX_MATCH_COPIED_off(prog)) + #endif /* PLUGGABLE_RE_EXTENSION */ /* Stuff that needs to be included in the plugable extension goes below here */ -#define RE_DEBUG_BIT 0x20000000 -#define RX_DEBUG(prog) ((prog)->reganch & RE_DEBUG_BIT) -#define RX_DEBUG_on(prog) ((prog)->reganch |= RE_DEBUG_BIT) - #ifdef PERL_OLD_COPY_ON_WRITE #define RX_MATCH_COPY_FREE(rx) \ STMT_START {if (rx->saved_copy) { \ @@ -131,9 +217,9 @@ typedef struct regexp { }} 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 RX_MATCH_UTF8(prog) ((prog)->extflags & RXf_MATCH_UTF8) +#define RX_MATCH_UTF8_on(prog) ((prog)->extflags |= RXf_MATCH_UTF8) +#define RX_MATCH_UTF8_off(prog) ((prog)->extflags &= ~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))) @@ -145,7 +231,7 @@ typedef struct regexp { #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) +#define ReREFCNT_dec(re) CALLREGFREE(re) #define FBMcf_TAIL_DOLLAR 1 #define FBMcf_TAIL_DOLLARM 2 @@ -155,8 +241,6 @@ typedef struct regexp { #define FBMrf_MULTILINE 1 -struct re_scream_pos_data_s; - /* an accepting state/position*/ struct _reg_trie_accepted { U8 *endpos; @@ -173,54 +257,19 @@ typedef struct { 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; +#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 50 +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 { @@ -232,68 +281,107 @@ typedef struct regmatch_state { } yes; struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; reg_trie_accepted *accept_buff; - U32 accepted; /* how many 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; struct { /* this first element must match u.yes */ struct regmatch_state *prev_yes_state; + struct regmatch_state *prev_eval; + struct regmatch_state *prev_curlyx; regexp *prev_rex; - int toggleutf; + 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; + regnode *B; /* the node following us */ + U32 close_paren; /* which close bracket is our end */ } eval; struct { - CHECKPOINT cp; /* remember current savestack indexes */ - struct regmatch_state *outercc; /* outer CURLYX state if any */ - - /* these contain the current curly state, and are accessed - * by subsequent WHILEMs */ + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + struct regmatch_state *prev_curlyx; /* previous cur_curlyx */ + 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 */ + regnode *A, *B; /* the nodes corresponding to /A*B/ */ + + /* 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; + U32 lastparen; + regnode *next_branch; /* next branch node */ + CHECKPOINT cp; + } branch; + + struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; 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; + 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; } u; } regmatch_state; @@ -320,7 +408,6 @@ typedef struct regmatch_slab { #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 @@ -348,7 +435,6 @@ struct re_save_state { 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 */ @@ -360,7 +446,7 @@ struct re_save_state { I32 re_state_reg_leftiter; /* wait until caching pos */ 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 */ + U32 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 */ @@ -369,6 +455,7 @@ 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