* 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 Nicholas'
+ * pluggable regex engine code */
struct regnode {
U8 flags;
struct reg_data;
+struct regexp_engine;
+
typedef struct regexp {
I32 *startp;
I32 *endp;
U32 lastcloseparen; /* last paren matched */
U32 reganch; /* Internal use only +
Tainted information used by regexec? */
+ const struct regexp_engine* engine;
regnode program[1]; /* Unwarranted chumminess with compiler. */
} regexp;
+
+typedef struct re_scream_pos_data_s
+{
+ char **scream_olds; /* match pos */
+ I32 *scream_pos; /* Internal iterator of scream. */
+} re_scream_pos_data;
+
+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
+ regexp* (*dupe) (pTHX_ const regexp *r, CLONE_PARAMS *param);
+#endif
+} regexp_engine;
+
#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_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 */
+#define ROPT_GPOS_CHECK (ROPT_GPOS_SEEN|ROPT_ANCH_GPOS)
+#define ROPT_RECURSE_SEEN 0x00001000
/* 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_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 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_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)
? 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) { \
#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
};
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 */
-/* structures for holding and saving the state maintained by regmatch() */
-
-typedef I32 CHECKPOINT;
+typedef struct {
+ regexp *prog;
+ char *bol;
+ char *till;
+ SV *sv;
+ char *ganch;
+} regmatch_info;
+
-/* Current curly descriptor */
-typedef struct curcur CURCUR;
-struct curcur {
- 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 */
- int minmod; /* whether to work our way up or down */
- regnode * scan; /* the thing to match */
- regnode * next; /* what has to match after it */
- char * lastloc; /* where we started matching this scan */
- CURCUR * oldcc; /* current curly before we started this one */
-};
+/* structures for holding and saving the state maintained by regmatch() */
-typedef struct re_cc_state
-{
- I32 ss;
- regnode *node;
- struct re_cc_state *prev;
- CURCUR *cc;
- 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;
+#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 50
+typedef I32 CHECKPOINT;
-typedef struct {
+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 */
+ int 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 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 */
- CURCUR *cc; /* current innermost curly struct */
char *locinput;
/* ... while the rest of these are local to an individual branch */
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 */
+ U32 accepted; /* how many accepting states we have seen */
+ U16 *jump; /* negative offsets from B */
+ regnode *B; /* node following the trie */
+ regnode *me; /* only needed for debugging */
} trie;
struct {
- CHECKPOINT cp; /* remember current savestack indexes */
- CHECKPOINT lastcp;
+ /* 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;
+ 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 */
- CURCUR *savecc;
+ /* 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 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;
- CURCUR *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 {
- I32 paren;
+ I32 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;
+ 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 {
- CHECKPOINT cp; /* remember current savestack indexes */
- CHECKPOINT lastcp;
- CURCUR *savecc;
- re_cc_state *cur_call_cc;
- regexp *end_re;
- } end;
- };
-
- re_cc_state *reg_call_cc; /* saved value of PL_reg_call_cc */
+ /* this first element must match u.yes */
+ struct regmatch_state *prev_yes_state;
+ I32 wanted;
+ regnode *me; /* the IFMATCH/SUSPEND/UNLESSM node */
+ } ifmatch; /* and SUSPEND/UNLESSM */
+ } u;
} regmatch_state;
/* how many regmatch_state structs to allocate as a single 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_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 */
+ 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:
+ */
https://perl5.git.perl.org / perl5.git / blobdiff