X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexp.h;h=a9c2130db60ea0f2d9939b560f674d55455a1b4b;hb=d8319b27ddc4c4d74cdae9e6de28ec2be390c5a1;hp=81552baf6cec490c4a2730fc3cdabc483b2db232;hpb=d6376244c2261025775e8b11a6a64d7c9e0d1612;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regexp.h b/regexp.h index 81552ba..a9c2130 100644 --- a/regexp.h +++ b/regexp.h @@ -1,6 +1,7 @@ /* regexp.h * - * Copyright (c) 1997-2002, Larry Wall + * Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003, + * 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. @@ -36,6 +37,9 @@ typedef struct regexp { struct reg_data *data; /* Additional data. */ 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; @@ -71,6 +75,7 @@ typedef struct regexp { #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 @@ -99,12 +104,35 @@ typedef struct regexp { ? RX_MATCH_COPIED_on(prog) \ : RX_MATCH_COPIED_off(prog)) +#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); \ + } \ + if (RX_MATCH_COPIED(rx)) { \ + Safefree(rx->subbeg); \ + RX_MATCH_COPIED_off(rx); \ + }} STMT_END +#else +#define RX_MATCH_COPY_FREE(rx) \ + STMT_START {if (RX_MATCH_COPIED(rx)) { \ + Safefree(rx->subbeg); \ + 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 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))) + #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 REXEC_ML 0x20 /* $* was set. */ #define ReREFCNT_inc(re) ((void)(re && re->refcnt++), re) #define ReREFCNT_dec(re) CALLREGFREE(aTHX_ re) @@ -118,3 +146,144 @@ typedef struct regexp { #define FBMrf_MULTILINE 1 struct re_scream_pos_data_s; + +/* an accepting state/position*/ +struct _reg_trie_accepted { + U8 *endpos; + U16 wordnum; +}; +typedef struct _reg_trie_accepted reg_trie_accepted; + + +/* structures for holding and saving the state maintained by regmatch() */ + +typedef I32 CHECKPOINT; + +typedef struct re_cc_state +{ + I32 ss; + regnode *node; + struct re_cc_state *prev; + struct regmatch_state *cc; /* state corresponding to the current curly */ + regexp *re; +} re_cc_state; + + +typedef enum { + resume_TRIE1, + resume_TRIE2, + resume_CURLYX, + resume_WHILEM1, + resume_WHILEM2, + resume_WHILEM3, + resume_WHILEM4, + resume_WHILEM5, + resume_WHILEM6, + resume_CURLYM1, + resume_CURLYM2, + resume_CURLYM3, + resume_CURLYM4, + resume_IFMATCH, + resume_PLUS1, + resume_PLUS2, + resume_PLUS3, + resume_PLUS4, + resume_END +} regmatch_resume_states; + + +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 */ + + union { + struct { + reg_trie_accepted *accept_buff; + U32 accepted; /* how many accepting states we have seen */ + } trie; + + struct { + CHECKPOINT cp; /* remember current savestack indexes */ + CHECKPOINT lastcp; + } 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 */ + 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 */ + } 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; + } whilem; + + struct { + I32 paren; + I32 c1, c2; /* case fold search */ + CHECKPOINT lastcp; + I32 l; + I32 matches; + I32 maxwanted; + } curlym; + + struct { + I32 paren; + CHECKPOINT lastcp; + I32 c1, c2; /* case fold search */ + char *e; + char *old; + int count; + } plus; /* and CURLYN/CURLY/STAR */ + + struct { + CHECKPOINT cp; /* remember current savestack indexes */ + CHECKPOINT lastcp; + struct regmatch_state *savecc; + re_cc_state *cur_call_cc; + regexp *end_re; + } end; + } u; + + re_cc_state *reg_call_cc; /* saved value of PL_reg_call_cc */ +} regmatch_state; + +/* how many regmatch_state structs to allocate as a single slab. + * We do it in 4K blocks for efficiency. The "3" is 2 for the next/prev + * pointers, plus 1 for any mythical malloc overhead. */ + +#define PERL_REGMATCH_SLAB_SLOTS \ + ((4096 - 3 * sizeof (void*)) / sizeof(regmatch_state)) + +typedef struct regmatch_slab { + regmatch_state states[PERL_REGMATCH_SLAB_SLOTS]; + struct regmatch_slab *prev, *next; +} regmatch_slab;