X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/1497789321a3812581136bb80750207bffc72afc..77cb431f8c908a31b286bd0ba26649cb6041bd72:/regexp.h diff --git a/regexp.h b/regexp.h index 6217aaa..455f048 100644 --- a/regexp.h +++ b/regexp.h @@ -1,4 +1,11 @@ /* regexp.h + * + * 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. + * */ /* @@ -30,6 +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; @@ -65,6 +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 @@ -93,12 +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) @@ -112,3 +146,217 @@ 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; + +/* 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; + char *bol; + char *till; + SV *sv; + char *ganch; +} 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_CURLYM4, + resume_IFMATCH, + resume_PLUS1, + resume_PLUS2, + resume_PLUS3, + resume_PLUS4 +} 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 { + + /* this is a fake union member that matches the first element + * of each member that needs to store positive backtrack + * information */ + struct { + struct regmatch_state *prev_yes_state; + } yes; + + struct { + reg_trie_accepted *accept_buff; + U32 accepted; /* how many accepting states we have seen */ + } trie; + + struct { + /* this first element must match u.yes */ + struct regmatch_state *prev_yes_state; + regexp *prev_rex; + int toggleutf; + 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 */ + } u; +} 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; + +#define PL_reg_flags PL_reg_state.re_state_reg_flags +#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_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 +#define PL_reg_oldcurpm PL_reg_state.re_state_reg_oldcurpm +#define PL_reg_curpm PL_reg_state.re_state_reg_curpm +#define PL_reg_oldsaved PL_reg_state.re_state_reg_oldsaved +#define PL_reg_oldsavedlen PL_reg_state.re_state_reg_oldsavedlen +#define PL_reg_maxiter PL_reg_state.re_state_reg_maxiter +#define PL_reg_leftiter PL_reg_state.re_state_reg_leftiter +#define PL_reg_poscache PL_reg_state.re_state_reg_poscache +#define PL_reg_poscache_size PL_reg_state.re_state_reg_poscache_size +#define PL_regsize PL_reg_state.re_state_regsize +#define PL_reg_starttry PL_reg_state.re_state_reg_starttry +#define PL_nrs PL_reg_state.re_state_nrs + +struct re_save_state { + U32 re_state_reg_flags; /* 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. */ + 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 */ + I32 re_state_reg_maxiter; /* max wait until caching pos */ + 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 */ + 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 */ +#endif +}; + +#define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \ + (1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack))) +/* + * Local variables: + * c-indentation-style: bsd + * c-basic-offset: 4 + * indent-tabs-mode: t + * End: + * + * ex: set ts=8 sts=4 sw=4 noet: + */