[p5sagit/p5-mst-13.2.git] / regexp.h

/*    regexp.h
 */

/*
 * Definitions etc. for regexp(3) routines.
 *
 * Caveat:  this is V8 regexp(3) [actually, a reimplementation thereof],
 * not the System V one.
 */


struct regnode {
    U8	flags;
    U8  type;
    U16 next_off;
};

typedef struct regnode regnode;

typedef struct regexp {
	I32 refcnt;
	char **startp;
	char **endp;
	regnode *regstclass;
	I32 minlen;		/* mininum possible length of $& */
	I32 prelen;		/* length of precomp */
	U32 nparens;		/* number of parentheses */
	U32 lastparen;		/* last paren matched */
	char *precomp;		/* pre-compilation regular expression */
	char *subbase;		/* saved string so \digit works forever */
	char *subbeg;		/* same, but not responsible for allocation */
	char *subend;		/* end of subbase */
	U16 naughty;		/* how exponential is this pattern? */
	U16 reganch;		/* Internal use only +
				   Tainted information used by regexec? */
        SV *anchored_substr;	/* Substring at fixed position wrt start. */
	I32 anchored_offset;	/* Position of it. */
        SV *float_substr;	/* Substring at variable position wrt start. */
	I32 float_min_offset;	/* Minimal position of it. */
	I32 float_max_offset;	/* Maximal position of it. */
        SV *check_substr;	/* Substring to check before matching. */
        I32 check_offset_min;	/* Offset of the above. */
        I32 check_offset_max;	/* Offset of the above. */
        struct reg_data *data;	/* Additional data. */
	regnode program[1];	/* Unwarranted chumminess with compiler. */
} regexp;

#define ROPT_ANCH		(ROPT_ANCH_BOL|ROPT_ANCH_MBOL|ROPT_ANCH_GPOS)
#define ROPT_ANCH_SINGLE	(ROPT_ANCH_BOL|ROPT_ANCH_GPOS)
#define ROPT_ANCH_BOL	 	1
#define ROPT_ANCH_MBOL	 	2
#define ROPT_ANCH_GPOS	 	4
#define ROPT_SKIP		8
#define ROPT_IMPLICIT		0x10	/* Converted .* to ^.* */
#define ROPT_NOSCAN		0x20	/* Check-string always at start. */
#define ROPT_GPOS_SEEN		0x40
#define ROPT_CHECK_ALL		0x80
#define ROPT_LOOKBEHIND_SEEN	0x100

#define ROPT_TAINTED_SEEN	0x8000

#define RX_MATCH_TAINTED(prog)	((prog)->reganch & ROPT_TAINTED_SEEN)
#define RX_MATCH_TAINTED_SET(prog, t) ((t) \
				       ? ((prog)->reganch |= ROPT_TAINTED_SEEN) \
				       : ((prog)->reganch &= ~ROPT_TAINTED_SEEN))

#define REXEC_COPY_STR	1		/* Need to copy the string. */
#define REXEC_CHECKED	2		/* check_substr already checked. */

#define ReREFCNT_inc(re) ((re && re->refcnt++), re)
#define ReREFCNT_dec(re) pregfree(re)
Commit	Line	Data
a0d0e21e	1	/* regexp.h
	2	*/
	3
378cc40b	4	/*
	5	* Definitions etc. for regexp(3) routines.
	6	*
	7	* Caveat: this is V8 regexp(3) [actually, a reimplementation thereof],
	8	* not the System V one.
	9	*/
	10
378cc40b	11
c277df42	12	struct regnode {
	13	U8 flags;
	14	U8 type;
	15	U16 next_off;
	16	};
	17
	18	typedef struct regnode regnode;
	19
378cc40b	20	typedef struct regexp {
c277df42	21	I32 refcnt;
fe14fcc3	22	char **startp;
fe14fcc3	23	char **endp;
c277df42	24	regnode *regstclass;
79072805	25	I32 minlen; /* mininum possible length of $& */
79072805	26	I32 prelen; /* length of precomp */
a0d0e21e	27	U32 nparens; /* number of parentheses */
a0d0e21e	28	U32 lastparen; /* last paren matched */
378cc40b	29	char precomp; / pre-compilation regular expression */
378cc40b	30	char subbase; / saved string so \digit works forever */
9ef589d8	31	char subbeg; / same, but not responsible for allocation */
00bf170e	32	char subend; / end of subbase */
a0d0e21e	33	U16 naughty; /* how exponential is this pattern? */
c277df42	34	U16 reganch; /* Internal use only +
	35	Tainted information used by regexec? */
	36	SV anchored_substr; / Substring at fixed position wrt start. */
	37	I32 anchored_offset; /* Position of it. */
	38	SV float_substr; / Substring at variable position wrt start. */
	39	I32 float_min_offset; /* Minimal position of it. */
	40	I32 float_max_offset; /* Maximal position of it. */
	41	SV check_substr; / Substring to check before matching. */
	42	I32 check_offset_min; /* Offset of the above. */
	43	I32 check_offset_max; /* Offset of the above. */
	44	struct reg_data data; / Additional data. */
	45	regnode program[1]; /* Unwarranted chumminess with compiler. */
378cc40b	46	} regexp;
378cc40b	47
c277df42	48	#define ROPT_ANCH (ROPT_ANCH_BOL\|ROPT_ANCH_MBOL\|ROPT_ANCH_GPOS)
	49	#define ROPT_ANCH_SINGLE (ROPT_ANCH_BOL\|ROPT_ANCH_GPOS)
	50	#define ROPT_ANCH_BOL 1
	51	#define ROPT_ANCH_MBOL 2
	52	#define ROPT_ANCH_GPOS 4
	53	#define ROPT_SKIP 8
	54	#define ROPT_IMPLICIT 0x10 /* Converted .* to ^.* */
	55	#define ROPT_NOSCAN 0x20 /* Check-string always at start. */
	56	#define ROPT_GPOS_SEEN 0x40
	57	#define ROPT_CHECK_ALL 0x80
	58	#define ROPT_LOOKBEHIND_SEEN 0x100
	59
	60	#define ROPT_TAINTED_SEEN 0x8000
	61
	62	#define RX_MATCH_TAINTED(prog) ((prog)->reganch & ROPT_TAINTED_SEEN)
	63	#define RX_MATCH_TAINTED_SET(prog, t) ((t) \
	64	? ((prog)->reganch \|= ROPT_TAINTED_SEEN) \
	65	: ((prog)->reganch &= ~ROPT_TAINTED_SEEN))
	66
	67	#define REXEC_COPY_STR 1 /* Need to copy the string. */
	68	#define REXEC_CHECKED 2 /* check_substr already checked. */
	69
	70	#define ReREFCNT_inc(re) ((re && re->refcnt++), re)
	71	#define ReREFCNT_dec(re) pregfree(re)