[perl5.git] / regexp.h

/*    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.
 *
 */

/*
 * Definitions etc. for regexp(3) routines.
 *
 * 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;
    U8  type;
    U16 next_off;
};

typedef struct regnode regnode;

struct reg_substr_data;

struct reg_data;

struct regexp_engine;
typedef struct regexp_paren_ofs {
    I32 *startp;
    I32 *endp;
} regexp_paren_ofs;

typedef struct regexp {
        I32 *startp;
	I32 *endp;
        regexp_paren_ofs *swap;
	regnode *regstclass;
        struct reg_substr_data *substrs;
	char *precomp;		/* pre-compilation regular expression */
        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;
	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 */
	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? */
	HV *paren_names;	/* Paren names */
	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);
    char*   (*as_str)   (pTHX_ MAGIC *mg, STRLEN *lp, U32 *flags,  I32 *haseval);
#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	 	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)

/* 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 ROPT_VERBARG_SEEN       0x20000000
#define ROPT_CUTGROUP_SEEN      0x40000000
#define ROPT_GPOS_FLOAT         0x80000000

#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_HAS_CUTGROUP(prog) ((prog)->reganch & ROPT_CUTGROUP_SEEN)
#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)
#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_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 */

#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 ReREFCNT_inc(re) ((void)(re && re->refcnt++), re)
#define ReREFCNT_dec(re) CALLREGFREE(re)

#define FBMcf_TAIL_DOLLAR	1
#define FBMcf_TAIL_DOLLARM	2
#define FBMcf_TAIL_Z		4
#define FBMcf_TAIL_z		8
#define FBMcf_TAIL		(FBMcf_TAIL_DOLLAR|FBMcf_TAIL_DOLLARM|FBMcf_TAIL_Z|FBMcf_TAIL_z)

#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;
    char *bol;
    char *till;
    SV *sv;
    char *ganch;
    char *cutpoint;
} regmatch_info;
 

/* structures for holding and saving the state maintained by regmatch() */

#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 50

typedef I32 CHECKPOINT;

typedef struct regmatch_state {
    int resume_state;		/* where to jump to on return */
    char *locinput;		/* where to backtrack in string on failure */

    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 {
	    /* 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 */
	    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;
	    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 {
	    /* 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 {
	    /* 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;
	    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 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 {
	    U32 paren;
	    CHECKPOINT cp;
	    I32 c1, c2;		/* case fold search */
	    char *maxpos;	/* highest possible point in string to match */
	    char *oldloc;	/* the previous locinput */
	    int count;
	    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;
    } 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_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 */
    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 */
#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:
 */
Commit	Line	Data
a0d0e21e	1	/* regexp.h
d6376244	2	*
4bb101f2 JH	3	* Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003,
4bb101f2 JH	4	* by Larry Wall and others
d6376244 JH	5	*
	6	* You may distribute under the terms of either the GNU General Public
	7	* License or the Artistic License, as specified in the README file.
	8	*
a0d0e21e LW	9	*/
a0d0e21e LW	10
378cc40b LW	11	/*
	12	* Definitions etc. for regexp(3) routines.
	13	*
	14	* Caveat: this is V8 regexp(3) [actually, a reimplementation thereof],
	15	* not the System V one.
	16	*/
be8e71aa YO	17	#ifndef PLUGGABLE_RE_EXTENSION
	18	/* we don't want to include this stuff if we are inside Nicholas'
	19	* pluggable regex engine code */
378cc40b	20
c277df42 IZ	21	struct regnode {
	22	U8 flags;
	23	U8 type;
	24	U16 next_off;
	25	};
	26
	27	typedef struct regnode regnode;
	28
cad2e5aa	29	struct reg_substr_data;
2779dcf1	30
0ee3c8d0 JH	31	struct reg_data;
0ee3c8d0 JH	32
f9f4320a	33	struct regexp_engine;
c74340f9 YO	34	typedef struct regexp_paren_ofs {
	35	I32 *startp;
	36	I32 *endp;
	37	} regexp_paren_ofs;
f9f4320a	38
378cc40b	39	typedef struct regexp {
c74340f9	40	I32 *startp;
cf93c79d	41	I32 *endp;
c74340f9	42	regexp_paren_ofs *swap;
c277df42	43	regnode *regstclass;
2779dcf1	44	struct reg_substr_data *substrs;
cf93c79d	45	char precomp; / pre-compilation regular expression */
c277df42	46	struct reg_data data; / Additional data. */
cf93c79d IZ	47	char subbeg; / saved or original string
cf93c79d IZ	48	so \digit works forever. */
f8c7b90f	49	#ifdef PERL_OLD_COPY_ON_WRITE
ed252734 NC	50	SV saved_copy; / If non-NULL, SV which is COW from original */
ed252734 NC	51	#endif
fac92740	52	U32 offsets; / offset annotations 20001228 MJD */
cf93c79d IZ	53	I32 sublen; /* Length of string pointed by subbeg */
cf93c79d IZ	54	I32 refcnt;
de8c5301 YO	55	I32 minlen; /* mininum possible length of string to match */
de8c5301 YO	56	I32 minlenret; /* mininum possible length of $& */
58e23c8d	57	U32 gofs; /* chars left of pos that we search from */
cf93c79d IZ	58	I32 prelen; /* length of precomp */
	59	U32 nparens; /* number of parentheses */
	60	U32 lastparen; /* last paren matched */
a01268b5	61	U32 lastcloseparen; /* last paren matched */
cf93c79d IZ	62	U32 reganch; /* Internal use only +
cf93c79d IZ	63	Tainted information used by regexec? */
81714fb9 YO	64	HV paren_names; / Paren names */
81714fb9 YO	65	const struct regexp_engine* engine;
c277df42	66	regnode program[1]; /* Unwarranted chumminess with compiler. */
378cc40b LW	67	} regexp;
378cc40b LW	68
f9f4320a YO	69
	70	typedef struct re_scream_pos_data_s
	71	{
	72	char *scream_olds; / match pos */
	73	I32 scream_pos; / Internal iterator of scream. */
	74	} re_scream_pos_data;
	75
	76	typedef struct regexp_engine {
f2f78491 YO	77	regexp* (comp) (pTHX_ char exp, char* xend, PMOP* pm);
f2f78491 YO	78	I32 (exec) (pTHX_ regexp prog, char* stringarg, char* strend,
f9f4320a YO	79	char* strbeg, I32 minend, SV* screamer,
f9f4320a YO	80	void* data, U32 flags);
f2f78491 YO	81	char* (intuit) (pTHX_ regexp prog, SV sv, char strpos,
	82	char *strend, U32 flags,
	83	struct re_scream_pos_data_s *data);
	84	SV* (checkstr) (pTHX_ regexp prog);
	85	void (free) (pTHX_ struct regexp r);
de8c5301	86	char* (as_str) (pTHX_ MAGIC mg, STRLEN lp, U32 flags, I32 *haseval);
f2f78491 YO	87	#ifdef USE_ITHREADS
f2f78491 YO	88	regexp* (dupe) (pTHX_ const regexp r, CLONE_PARAMS *param);
f9f4320a YO	89	#endif
	90	} regexp_engine;
	91
cad2e5aa JH	92	#define ROPT_ANCH (ROPT_ANCH_BOL\|ROPT_ANCH_MBOL\|ROPT_ANCH_GPOS\|ROPT_ANCH_SBOL)
cad2e5aa JH	93	#define ROPT_ANCH_SINGLE (ROPT_ANCH_SBOL\|ROPT_ANCH_GPOS)
f2278c82 YO	94	#define ROPT_ANCH_BOL 0x00000001
	95	#define ROPT_ANCH_MBOL 0x00000002
	96	#define ROPT_ANCH_SBOL 0x00000004
	97	#define ROPT_ANCH_GPOS 0x00000008
	98	#define ROPT_SKIP 0x00000010
	99	#define ROPT_IMPLICIT 0x00000020 /* Converted .* to ^.* */
	100	#define ROPT_NOSCAN 0x00000040 /* Check-string always at start. */
	101	#define ROPT_GPOS_SEEN 0x00000080
	102	#define ROPT_CHECK_ALL 0x00000100
	103	#define ROPT_LOOKBEHIND_SEEN 0x00000200
	104	#define ROPT_EVAL_SEEN 0x00000400
	105	#define ROPT_CANY_SEEN 0x00000800
cce850e4	106	#define ROPT_SANY_SEEN ROPT_CANY_SEEN /* src bckwrd cmpt */
f9f4320a	107	#define ROPT_GPOS_CHECK (ROPT_GPOS_SEEN\|ROPT_ANCH_GPOS)
a0ed51b3	108
5d458dd8	109	/* 0xF800 of reganch is used by PMf_COMPILETIME */
c277df42	110
f2278c82 YO	111	#define ROPT_UTF8 0x00010000
	112	#define ROPT_NAUGHTY 0x00020000 /* how exponential is this pattern? */
	113	#define ROPT_COPY_DONE 0x00040000 /* subbeg is a copy of the string */
	114	#define ROPT_TAINTED_SEEN 0x00080000
a30b2f1f	115	#define ROPT_MATCH_UTF8 0x10000000 /* subbeg is utf-8 */
5d458dd8 YO	116	#define ROPT_VERBARG_SEEN 0x20000000
5d458dd8 YO	117	#define ROPT_CUTGROUP_SEEN 0x40000000
58e23c8d	118	#define ROPT_GPOS_FLOAT 0x80000000
a0ed51b3	119
f2278c82 YO	120	#define RE_USE_INTUIT_NOML 0x00100000 /* Best to intuit before matching */
	121	#define RE_USE_INTUIT_ML 0x00200000
	122	#define REINT_AUTORITATIVE_NOML 0x00400000 /* Can trust a positive answer */
be8e71aa	123	#define REINT_AUTORITATIVE_ML 0x00800000
f2278c82 YO	124	#define REINT_ONCE_NOML 0x01000000 /* Intuit can succed once only. */
	125	#define REINT_ONCE_ML 0x02000000
	126	#define RE_INTUIT_ONECHAR 0x04000000
	127	#define RE_INTUIT_TAIL 0x08000000
	128
cad2e5aa JH	129
	130	#define RE_USE_INTUIT (RE_USE_INTUIT_NOML\|RE_USE_INTUIT_ML)
	131	#define REINT_AUTORITATIVE (REINT_AUTORITATIVE_NOML\|REINT_AUTORITATIVE_ML)
	132	#define REINT_ONCE (REINT_ONCE_NOML\|REINT_ONCE_ML)
	133
5d458dd8	134	#define RX_HAS_CUTGROUP(prog) ((prog)->reganch & ROPT_CUTGROUP_SEEN)
c277df42	135	#define RX_MATCH_TAINTED(prog) ((prog)->reganch & ROPT_TAINTED_SEEN)
72311751 GS	136	#define RX_MATCH_TAINTED_on(prog) ((prog)->reganch \|= ROPT_TAINTED_SEEN)
	137	#define RX_MATCH_TAINTED_off(prog) ((prog)->reganch &= ~ROPT_TAINTED_SEEN)
	138	#define RX_MATCH_TAINTED_set(prog, t) ((t) \
	139	? RX_MATCH_TAINTED_on(prog) \
	140	: RX_MATCH_TAINTED_off(prog))
c277df42	141
cf93c79d IZ	142	#define RX_MATCH_COPIED(prog) ((prog)->reganch & ROPT_COPY_DONE)
	143	#define RX_MATCH_COPIED_on(prog) ((prog)->reganch \|= ROPT_COPY_DONE)
	144	#define RX_MATCH_COPIED_off(prog) ((prog)->reganch &= ~ROPT_COPY_DONE)
	145	#define RX_MATCH_COPIED_set(prog,t) ((t) \
	146	? RX_MATCH_COPIED_on(prog) \
	147	: RX_MATCH_COPIED_off(prog))
f9f4320a	148
be8e71aa YO	149	#endif /* PLUGGABLE_RE_EXTENSION */
	150
	151	/* Stuff that needs to be included in the plugable extension goes below here */
	152
f8c7b90f	153	#ifdef PERL_OLD_COPY_ON_WRITE
ed252734 NC	154	#define RX_MATCH_COPY_FREE(rx) \
	155	STMT_START {if (rx->saved_copy) { \
	156	SV_CHECK_THINKFIRST_COW_DROP(rx->saved_copy); \
	157	} \
	158	if (RX_MATCH_COPIED(rx)) { \
	159	Safefree(rx->subbeg); \
	160	RX_MATCH_COPIED_off(rx); \
	161	}} STMT_END
	162	#else
	163	#define RX_MATCH_COPY_FREE(rx) \
	164	STMT_START {if (RX_MATCH_COPIED(rx)) { \
	165	Safefree(rx->subbeg); \
	166	RX_MATCH_COPIED_off(rx); \
	167	}} STMT_END
	168	#endif
	169
a30b2f1f RGS	170	#define RX_MATCH_UTF8(prog) ((prog)->reganch & ROPT_MATCH_UTF8)
	171	#define RX_MATCH_UTF8_on(prog) ((prog)->reganch \|= ROPT_MATCH_UTF8)
	172	#define RX_MATCH_UTF8_off(prog) ((prog)->reganch &= ~ROPT_MATCH_UTF8)
	173	#define RX_MATCH_UTF8_set(prog, t) ((t) \
	174	? (RX_MATCH_UTF8_on(prog), (PL_reg_match_utf8 = 1)) \
	175	: (RX_MATCH_UTF8_off(prog), (PL_reg_match_utf8 = 0)))
	176
cad2e5aa JH	177	#define REXEC_COPY_STR 0x01 /* Need to copy the string. */
	178	#define REXEC_CHECKED 0x02 /* check_substr already checked. */
	179	#define REXEC_SCREAM 0x04 /* use scream table. */
	180	#define REXEC_IGNOREPOS 0x08 /* \G matches at start. */
cf93c79d	181	#define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */
c277df42	182
155aba94	183	#define ReREFCNT_inc(re) ((void)(re && re->refcnt++), re)
f9f4320a	184	#define ReREFCNT_dec(re) CALLREGFREE(re)
cf93c79d IZ	185
cf93c79d IZ	186	#define FBMcf_TAIL_DOLLAR 1
cad2e5aa JH	187	#define FBMcf_TAIL_DOLLARM 2
	188	#define FBMcf_TAIL_Z 4
	189	#define FBMcf_TAIL_z 8
	190	#define FBMcf_TAIL (FBMcf_TAIL_DOLLAR\|FBMcf_TAIL_DOLLARM\|FBMcf_TAIL_Z\|FBMcf_TAIL_z)
cf93c79d IZ	191
cf93c79d IZ	192	#define FBMrf_MULTILINE 1
cad2e5aa	193
5d9a96ca DM	194	/* an accepting state/position*/
	195	struct _reg_trie_accepted {
	196	U8 *endpos;
	197	U16 wordnum;
	198	};
	199	typedef struct _reg_trie_accepted reg_trie_accepted;
	200
3b0527fe DM	201	/* some basic information about the current match that is created by
	202	* Perl_regexec_flags and then passed to regtry(), regmatch() etc */
	203
	204	typedef struct {
	205	regexp *prog;
	206	char *bol;
	207	char *till;
	208	SV *sv;
	209	char *ganch;
24b23f37	210	char *cutpoint;
3b0527fe DM	211	} regmatch_info;
3b0527fe DM	212
5d9a96ca DM	213
	214	/* structures for holding and saving the state maintained by regmatch() */
	215
6bda09f9 YO	216	#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 50
6bda09f9 YO	217
5d9a96ca DM	218	typedef I32 CHECKPOINT;
5d9a96ca DM	219
a0374537	220	typedef struct regmatch_state {
40a82448	221	int resume_state; /* where to jump to on return */
24d3c4a9	222	char locinput; / where to backtrack in string on failure */
e822a8b4 DM	223
e822a8b4 DM	224	union {
77cb431f DM	225
	226	/* this is a fake union member that matches the first element
	227	* of each member that needs to store positive backtrack
	228	* information */
	229	struct {
	230	struct regmatch_state *prev_yes_state;
	231	} yes;
	232
e822a8b4	233	struct {
5d458dd8 YO	234	/* this first element must match u.yes */
5d458dd8 YO	235	struct regmatch_state *prev_yes_state;
e822a8b4	236	reg_trie_accepted *accept_buff;
166ba7cd	237	U32 accepted; /* how many accepting states we have seen */
7f69552c	238	U16 jump; / positive offsets from me */
166ba7cd	239	regnode B; / node following the trie */
7f69552c	240	regnode me; / Which node am I - needed for jump tries*/
e822a8b4 DM	241	} trie;
	242
	243	struct {
77cb431f DM	244	/* this first element must match u.yes */
77cb431f DM	245	struct regmatch_state *prev_yes_state;
faec1544 DM	246	struct regmatch_state *prev_eval;
faec1544 DM	247	struct regmatch_state *prev_curlyx;
aa283a38	248	regexp *prev_rex;
faec1544 DM	249	U32 toggle_reg_flags; /* what bits in PL_reg_flags to
	250	flip when transitioning between
	251	inner and outer rexen */
aa283a38 DM	252	CHECKPOINT cp; /* remember current savestack indexes */
aa283a38 DM	253	CHECKPOINT lastcp;
40a82448	254	regnode B; / the node following us */
6bda09f9	255	U32 close_paren; /* which close bracket is our end */
e822a8b4 DM	256	} eval;
	257
	258	struct {
c476f425	259	/* this first element must match u.yes */
262b90c4	260	struct regmatch_state *prev_yes_state;
c476f425 DM	261	struct regmatch_state prev_curlyx; / previous cur_curlyx */
	262	CHECKPOINT cp; /* remember current savestack index */
	263	bool minmod;
a0374537	264	int parenfloor;/* how far back to strip paren data */
c476f425 DM	265	int min; /* the minimal number of A's to match */
	266	int max; /* the maximal number of A's to match */
	267	regnode A, B; /* the nodes corresponding to /AB/ /
	268
	269	/* these two are modified by WHILEM */
	270	int count; /* how many instances of A we've matched */
	271	char lastloc;/ where previous A matched (0-len detect) */
e822a8b4 DM	272	} curlyx;
	273
	274	struct {
c476f425	275	/* this first element must match u.yes */
262b90c4	276	struct regmatch_state *prev_yes_state;
c476f425 DM	277	struct regmatch_state *save_curlyx;
	278	CHECKPOINT cp; /* remember current savestack indexes */
	279	CHECKPOINT lastcp;
	280	char save_lastloc; / previous curlyx.lastloc */
	281	I32 cache_offset;
	282	I32 cache_mask;
e822a8b4 DM	283	} whilem;
	284
	285	struct {
5d458dd8 YO	286	/* this first element must match u.yes */
5d458dd8 YO	287	struct regmatch_state *prev_yes_state;
3b6647e0	288	U32 lastparen;
40a82448 DM	289	regnode next_branch; / next branch node */
	290	CHECKPOINT cp;
	291	} branch;
	292
	293	struct {
dad79028 DM	294	/* this first element must match u.yes */
dad79028 DM	295	struct regmatch_state *prev_yes_state;
e822a8b4	296	I32 c1, c2; /* case fold search */
40a82448 DM	297	CHECKPOINT cp;
	298	I32 alen; /* length of first-matched A string */
	299	I32 count;
0cadcf80	300	bool minmod;
40a82448 DM	301	regnode A, B; /* the nodes corresponding to /AB/ /
40a82448 DM	302	regnode me; / the curlym node */
e822a8b4 DM	303	} curlym;
	304
	305	struct {
3b6647e0	306	U32 paren;
c255a977	307	CHECKPOINT cp;
e822a8b4	308	I32 c1, c2; /* case fold search */
c255a977 DM	309	char maxpos; / highest possible point in string to match */
c255a977 DM	310	char oldloc; / the previous locinput */
e822a8b4	311	int count;
c255a977 DM	312	int min, max; /* {m,n} */
	313	regnode A, B; /* the nodes corresponding to /AB/ /
	314	} curly; /* and CURLYN/PLUS/STAR */
dad79028 DM	315
	316	struct {
	317	/* this first element must match u.yes */
	318	struct regmatch_state *prev_yes_state;
	319	I32 wanted;
24d3c4a9	320	I32 logical; /* saved copy of 'logical' var */
40a82448	321	regnode me; / the IFMATCH/SUSPEND/UNLESSM node */
dad79028	322	} ifmatch; /* and SUSPEND/UNLESSM */
e2e6a0f1 YO	323
	324	struct {
	325	/* this first element must match u.yes */
	326	struct regmatch_state *prev_yes_state;
	327	struct regmatch_state *prev_mark;
	328	SV* mark_name;
	329	char *mark_loc;
	330	} mark;
d8319b27	331	} u;
5d9a96ca DM	332	} regmatch_state;
	333
	334	/* how many regmatch_state structs to allocate as a single slab.
	335	* We do it in 4K blocks for efficiency. The "3" is 2 for the next/prev
	336	* pointers, plus 1 for any mythical malloc overhead. */
	337
	338	#define PERL_REGMATCH_SLAB_SLOTS \
	339	((4096 - 3 * sizeof (void*)) / sizeof(regmatch_state))
	340
	341	typedef struct regmatch_slab {
	342	regmatch_state states[PERL_REGMATCH_SLAB_SLOTS];
	343	struct regmatch_slab prev, next;
	344	} regmatch_slab;
1ade1aa1	345
46ab3289 NC	346	#define PL_reg_flags PL_reg_state.re_state_reg_flags
	347	#define PL_bostr PL_reg_state.re_state_bostr
	348	#define PL_reginput PL_reg_state.re_state_reginput
46ab3289 NC	349	#define PL_regeol PL_reg_state.re_state_regeol
	350	#define PL_regstartp PL_reg_state.re_state_regstartp
	351	#define PL_regendp PL_reg_state.re_state_regendp
	352	#define PL_reglastparen PL_reg_state.re_state_reglastparen
	353	#define PL_reglastcloseparen PL_reg_state.re_state_reglastcloseparen
46ab3289 NC	354	#define PL_reg_start_tmp PL_reg_state.re_state_reg_start_tmp
	355	#define PL_reg_start_tmpl PL_reg_state.re_state_reg_start_tmpl
	356	#define PL_reg_eval_set PL_reg_state.re_state_reg_eval_set
46ab3289 NC	357	#define PL_reg_match_utf8 PL_reg_state.re_state_reg_match_utf8
	358	#define PL_reg_magic PL_reg_state.re_state_reg_magic
	359	#define PL_reg_oldpos PL_reg_state.re_state_reg_oldpos
	360	#define PL_reg_oldcurpm PL_reg_state.re_state_reg_oldcurpm
	361	#define PL_reg_curpm PL_reg_state.re_state_reg_curpm
	362	#define PL_reg_oldsaved PL_reg_state.re_state_reg_oldsaved
	363	#define PL_reg_oldsavedlen PL_reg_state.re_state_reg_oldsavedlen
	364	#define PL_reg_maxiter PL_reg_state.re_state_reg_maxiter
	365	#define PL_reg_leftiter PL_reg_state.re_state_reg_leftiter
	366	#define PL_reg_poscache PL_reg_state.re_state_reg_poscache
	367	#define PL_reg_poscache_size PL_reg_state.re_state_reg_poscache_size
	368	#define PL_regsize PL_reg_state.re_state_regsize
	369	#define PL_reg_starttry PL_reg_state.re_state_reg_starttry
	370	#define PL_nrs PL_reg_state.re_state_nrs
	371
1ade1aa1 NC	372	struct re_save_state {
	373	U32 re_state_reg_flags; /* from regexec.c */
	374	char *re_state_bostr;
	375	char re_state_reginput; / String-input pointer. */
1ade1aa1 NC	376	char re_state_regeol; / End of input, for $ check. */
	377	I32 re_state_regstartp; / Pointer to startp array. */
	378	I32 re_state_regendp; / Ditto for endp. */
	379	U32 re_state_reglastparen; / Similarly for lastparen. */
	380	U32 re_state_reglastcloseparen; / Similarly for lastcloseparen. */
1ade1aa1 NC	381	char *re_state_reg_start_tmp; / from regexec.c */
	382	U32 re_state_reg_start_tmpl; /* from regexec.c */
	383	I32 re_state_reg_eval_set; /* from regexec.c */
1ade1aa1 NC	384	bool re_state_reg_match_utf8; /* from regexec.c */
	385	MAGIC re_state_reg_magic; / from regexec.c */
	386	I32 re_state_reg_oldpos; /* from regexec.c */
	387	PMOP re_state_reg_oldcurpm; / from regexec.c */
	388	PMOP re_state_reg_curpm; / from regexec.c */
	389	char re_state_reg_oldsaved; / old saved substr during match */
	390	STRLEN re_state_reg_oldsavedlen; /* old length of saved substr during match */
	391	I32 re_state_reg_maxiter; /* max wait until caching pos */
	392	I32 re_state_reg_leftiter; /* wait until caching pos */
	393	char re_state_reg_poscache; / cache of pos of WHILEM */
	394	STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */
3b6647e0	395	U32 re_state_regsize; /* from regexec.c */
1ade1aa1 NC	396	char re_state_reg_starttry; / from regexec.c */
1ade1aa1 NC	397	#ifdef PERL_OLD_COPY_ON_WRITE
46ab3289	398	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 */
1ade1aa1 NC	399	#endif
	400	};
	401
	402	#define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \
	403	(1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack)))
	404	/*
	405	* Local variables:
	406	* c-indentation-style: bsd
	407	* c-basic-offset: 4
	408	* indent-tabs-mode: t
	409	* End:
	410	*
	411	* ex: set ts=8 sts=4 sw=4 noet:
	412	*/