[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;

#ifdef PERL_OLD_COPY_ON_WRITE
#define SV_SAVED_COPY   SV *saved_copy; /* If non-NULL, SV which is COW from original */
#else
#define SV_SAVED_COPY
#endif

typedef struct regexp {
        /* Generic details */
	const struct regexp_engine* engine; /* what created this regexp? */
	I32 refcnt;             /* Refcount of this regexp */
        
        /* The original string as passed to the compilation routine */
	char *precomp;		/* pre-compilation regular expression */
	I32 prelen;		/* length of precomp */
        
	/* Used for generic optimisations by the perl core. 
	   All engines are expected to provide this information.  */
	U32 extflags;           /* Flags used both externally and internally */
	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 */
	U32 nparens;		/* number of capture buffers */
	HV *paren_names;	/* Optional hash of paren names */
        struct reg_substr_data *substrs; /* substring data about strings that must appear
                                   in the final match, used for optimisations */

        /* Data about the last/current match. Used by the core and therefore
           must be populated by all engines. */
	char *subbeg;		/* saved or original string 
				   so \digit works forever. */
	I32 sublen;		/* Length of string pointed by subbeg */
        I32 *startp;            /* Array of offsets from start of string (@-) */
	I32 *endp;              /* Array of offsets from start of string (@+) */
	
	SV_SAVED_COPY           /* If non-NULL, SV which is COW from original */
        U32 lastparen;		/* last open paren matched */
	U32 lastcloseparen;	/* last close paren matched */
	
        /* Perl Regex Engine specific data. Other engines shouldn't need 
           to touch this. Should be refactored out into a different structure
           and accessed via the *pprivate field. (except intflags) */
	U32 intflags;		/* Internal flags */
	void *pprivate;         /* Data private to the regex engine which 
                                   created this object. Perl will never mess with
                                   this member at all. */
        regexp_paren_ofs *swap; /* Swap copy of *startp / *endp */
	U32 *offsets;           /* offset annotations 20001228 MJD 
                                   data about mapping the program to the 
                                   string*/
        regnode *regstclass;    /* Optional startclass as identified or constructed
                                   by the optimiser */
        struct reg_data *data;	/* Additional miscellaneous data used by the program.
                                   Used to make it easier to clone and free arbitrary
                                   data that the regops need. Often the ARG field of
                                   a regop is an index into this structure */
	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;

/* 
 * Flags stored in regexp->intflags 
 * These are used only internally to the regexp engine
 */
#define PREGf_SKIP		0x00000001
#define PREGf_IMPLICIT		0x00000002 /* Converted .* to ^.* */
#define PREGf_NAUGHTY		0x00000004 /* how exponential is this pattern? */
#define PREGf_VERBARG_SEEN      0x00000008
#define PREGf_CUTGROUP_SEEN	0x00000010


/* Flags stored in regexp->extflags 
 * These are used by code external to the regexp engine
 */

/* Anchor and GPOS related stuff */
#define RXf_ANCH_BOL    	0x00000001
#define RXf_ANCH_MBOL   	0x00000002
#define RXf_ANCH_SBOL   	0x00000004
#define RXf_ANCH_GPOS   	0x00000008
#define RXf_GPOS_SEEN   	0x00000010
#define RXf_GPOS_FLOAT  	0x00000020
/* five bits here */
#define RXf_ANCH        	(RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL)
#define RXf_GPOS_CHECK          (RXf_GPOS_SEEN|RXf_ANCH_GPOS)
#define RXf_ANCH_SINGLE         (RXf_ANCH_SBOL|RXf_ANCH_GPOS)        
/* 
 * 0xF800 of extflags is used by PMf_COMPILETIME 
 * These are the regex equivelent of the PMf_xyz stuff defined 
 * in op.h
 */
#define RXf_PMf_LOCALE  	0x00000800
#define RXf_PMf_MULTILINE	0x00001000
#define RXf_PMf_SINGLELINE	0x00002000
#define RXf_PMf_FOLD    	0x00004000
#define RXf_PMf_EXTENDED	0x00008000
#define RXf_PMf_COMPILETIME	(RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_LOCALE|RXf_PMf_FOLD|RXf_PMf_EXTENDED)

/* What we have seen */
/* one bit here */
#define RXf_LOOKBEHIND_SEEN	0x00020000
#define RXf_EVAL_SEEN   	0x00040000
#define RXf_CANY_SEEN   	0x00080000

/* Special */
#define RXf_NOSCAN      	0x00100000
#define RXf_CHECK_ALL   	0x00200000

/* UTF8 related */
#define RXf_UTF8        	0x00400000
#define RXf_MATCH_UTF8  	0x00800000

/* Intuit related */
#define RXf_USE_INTUIT_NOML	0x01000000
#define RXf_USE_INTUIT_ML	0x02000000
#define RXf_INTUIT_TAIL 	0x04000000
/* one bit here */
#define RXf_USE_INTUIT		(RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML)

/* Copy and tainted info */
#define RXf_COPY_DONE   	0x10000000
#define RXf_TAINTED_SEEN	0x20000000
/* two bits here  */


#define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN)
#define RX_MATCH_TAINTED(prog)	((prog)->extflags & RXf_TAINTED_SEEN)
#define RX_MATCH_TAINTED_on(prog) ((prog)->extflags |= RXf_TAINTED_SEEN)
#define RX_MATCH_TAINTED_off(prog) ((prog)->extflags &= ~RXf_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)->extflags & RXf_COPY_DONE)
#define RX_MATCH_COPIED_on(prog)	((prog)->extflags |= RXf_COPY_DONE)
#define RX_MATCH_COPIED_off(prog)	((prog)->extflags &= ~RXf_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)->extflags & RXf_MATCH_UTF8)
#define RX_MATCH_UTF8_on(prog)		((prog)->extflags |= RXf_MATCH_UTF8)
#define RX_MATCH_UTF8_off(prog)		((prog)->extflags &= ~RXf_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;
bbe252da	34
c74340f9 YO	35	typedef struct regexp_paren_ofs {
	36	I32 *startp;
	37	I32 *endp;
	38	} regexp_paren_ofs;
f9f4320a	39
f8c7b90f	40	#ifdef PERL_OLD_COPY_ON_WRITE
bbe252da YO	41	#define SV_SAVED_COPY SV saved_copy; / If non-NULL, SV which is COW from original */
	42	#else
	43	#define SV_SAVED_COPY
ed252734	44	#endif
bbe252da YO	45
	46	typedef struct regexp {
	47	/* Generic details */
	48	const struct regexp_engine* engine; /* what created this regexp? */
	49	I32 refcnt; /* Refcount of this regexp */
	50
	51	/* The original string as passed to the compilation routine */
	52	char precomp; / pre-compilation regular expression */
	53	I32 prelen; /* length of precomp */
	54
	55	/* Used for generic optimisations by the perl core.
	56	All engines are expected to provide this information. */
	57	U32 extflags; /* Flags used both externally and internally */
de8c5301 YO	58	I32 minlen; /* mininum possible length of string to match */
de8c5301 YO	59	I32 minlenret; /* mininum possible length of $& */
58e23c8d	60	U32 gofs; /* chars left of pos that we search from */
bbe252da YO	61	U32 nparens; /* number of capture buffers */
	62	HV paren_names; / Optional hash of paren names */
	63	struct reg_substr_data substrs; / substring data about strings that must appear
	64	in the final match, used for optimisations */
	65
	66	/* Data about the last/current match. Used by the core and therefore
	67	must be populated by all engines. */
	68	char subbeg; / saved or original string
	69	so \digit works forever. */
	70	I32 sublen; /* Length of string pointed by subbeg */
	71	I32 startp; / Array of offsets from start of string (@-) */
	72	I32 endp; / Array of offsets from start of string (@+) */
	73
	74	SV_SAVED_COPY /* If non-NULL, SV which is COW from original */
	75	U32 lastparen; /* last open paren matched */
	76	U32 lastcloseparen; /* last close paren matched */
	77
	78	/* Perl Regex Engine specific data. Other engines shouldn't need
	79	to touch this. Should be refactored out into a different structure
	80	and accessed via the pprivate field. (except intflags) /
	81	U32 intflags; /* Internal flags */
	82	void pprivate; / Data private to the regex engine which
	83	created this object. Perl will never mess with
	84	this member at all. */
	85	regexp_paren_ofs swap; / Swap copy of startp / endp */
	86	U32 offsets; / offset annotations 20001228 MJD
	87	data about mapping the program to the
	88	string*/
	89	regnode regstclass; / Optional startclass as identified or constructed
	90	by the optimiser */
	91	struct reg_data data; / Additional miscellaneous data used by the program.
	92	Used to make it easier to clone and free arbitrary
	93	data that the regops need. Often the ARG field of
	94	a regop is an index into this structure */
c277df42	95	regnode program[1]; /* Unwarranted chumminess with compiler. */
378cc40b LW	96	} regexp;
378cc40b LW	97
f9f4320a YO	98
	99	typedef struct re_scream_pos_data_s
	100	{
	101	char *scream_olds; / match pos */
	102	I32 scream_pos; / Internal iterator of scream. */
	103	} re_scream_pos_data;
	104
	105	typedef struct regexp_engine {
f2f78491 YO	106	regexp* (comp) (pTHX_ char exp, char* xend, PMOP* pm);
f2f78491 YO	107	I32 (exec) (pTHX_ regexp prog, char* stringarg, char* strend,
f9f4320a YO	108	char* strbeg, I32 minend, SV* screamer,
f9f4320a YO	109	void* data, U32 flags);
f2f78491 YO	110	char* (intuit) (pTHX_ regexp prog, SV sv, char strpos,
	111	char *strend, U32 flags,
	112	struct re_scream_pos_data_s *data);
	113	SV* (checkstr) (pTHX_ regexp prog);
	114	void (free) (pTHX_ struct regexp r);
de8c5301	115	char* (as_str) (pTHX_ MAGIC mg, STRLEN lp, U32 flags, I32 *haseval);
f2f78491 YO	116	#ifdef USE_ITHREADS
f2f78491 YO	117	regexp* (dupe) (pTHX_ const regexp r, CLONE_PARAMS *param);
f9f4320a YO	118	#endif
	119	} regexp_engine;
	120
bbe252da YO	121	/*
	122	* Flags stored in regexp->intflags
	123	* These are used only internally to the regexp engine
	124	*/
	125	#define PREGf_SKIP 0x00000001
	126	#define PREGf_IMPLICIT 0x00000002 /* Converted .* to ^.* */
	127	#define PREGf_NAUGHTY 0x00000004 /* how exponential is this pattern? */
	128	#define PREGf_VERBARG_SEEN 0x00000008
	129	#define PREGf_CUTGROUP_SEEN 0x00000010
	130
	131
	132	/* Flags stored in regexp->extflags
	133	* These are used by code external to the regexp engine
	134	*/
	135
	136	/* Anchor and GPOS related stuff */
	137	#define RXf_ANCH_BOL 0x00000001
	138	#define RXf_ANCH_MBOL 0x00000002
	139	#define RXf_ANCH_SBOL 0x00000004
	140	#define RXf_ANCH_GPOS 0x00000008
	141	#define RXf_GPOS_SEEN 0x00000010
	142	#define RXf_GPOS_FLOAT 0x00000020
	143	/* five bits here */
	144	#define RXf_ANCH (RXf_ANCH_BOL\|RXf_ANCH_MBOL\|RXf_ANCH_GPOS\|RXf_ANCH_SBOL)
	145	#define RXf_GPOS_CHECK (RXf_GPOS_SEEN\|RXf_ANCH_GPOS)
	146	#define RXf_ANCH_SINGLE (RXf_ANCH_SBOL\|RXf_ANCH_GPOS)
	147	/*
	148	* 0xF800 of extflags is used by PMf_COMPILETIME
	149	* These are the regex equivelent of the PMf_xyz stuff defined
	150	* in op.h
	151	*/
	152	#define RXf_PMf_LOCALE 0x00000800
	153	#define RXf_PMf_MULTILINE 0x00001000
	154	#define RXf_PMf_SINGLELINE 0x00002000
	155	#define RXf_PMf_FOLD 0x00004000
	156	#define RXf_PMf_EXTENDED 0x00008000
	157	#define RXf_PMf_COMPILETIME (RXf_PMf_MULTILINE\|RXf_PMf_SINGLELINE\|RXf_PMf_LOCALE\|RXf_PMf_FOLD\|RXf_PMf_EXTENDED)
	158
	159	/* What we have seen */
	160	/* one bit here */
	161	#define RXf_LOOKBEHIND_SEEN 0x00020000
	162	#define RXf_EVAL_SEEN 0x00040000
	163	#define RXf_CANY_SEEN 0x00080000
	164
	165	/* Special */
	166	#define RXf_NOSCAN 0x00100000
	167	#define RXf_CHECK_ALL 0x00200000
	168
	169	/* UTF8 related */
	170	#define RXf_UTF8 0x00400000
	171	#define RXf_MATCH_UTF8 0x00800000
	172
	173	/* Intuit related */
	174	#define RXf_USE_INTUIT_NOML 0x01000000
	175	#define RXf_USE_INTUIT_ML 0x02000000
	176	#define RXf_INTUIT_TAIL 0x04000000
	177	/* one bit here */
	178	#define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML\|RXf_USE_INTUIT_ML)
	179
	180	/* Copy and tainted info */
	181	#define RXf_COPY_DONE 0x10000000
	182	#define RXf_TAINTED_SEEN 0x20000000
	183	/* two bits here */
	184
185
186	#define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN)
187	#define RX_MATCH_TAINTED(prog) ((prog)->extflags & RXf_TAINTED_SEEN)
188	#define RX_MATCH_TAINTED_on(prog) ((prog)->extflags \|= RXf_TAINTED_SEEN)
189	#define RX_MATCH_TAINTED_off(prog) ((prog)->extflags &= ~RXf_TAINTED_SEEN)
72311751 GS	190	#define RX_MATCH_TAINTED_set(prog, t) ((t) \
	191	? RX_MATCH_TAINTED_on(prog) \
	192	: RX_MATCH_TAINTED_off(prog))
c277df42	193
bbe252da YO	194	#define RX_MATCH_COPIED(prog) ((prog)->extflags & RXf_COPY_DONE)
	195	#define RX_MATCH_COPIED_on(prog) ((prog)->extflags \|= RXf_COPY_DONE)
	196	#define RX_MATCH_COPIED_off(prog) ((prog)->extflags &= ~RXf_COPY_DONE)
cf93c79d IZ	197	#define RX_MATCH_COPIED_set(prog,t) ((t) \
	198	? RX_MATCH_COPIED_on(prog) \
	199	: RX_MATCH_COPIED_off(prog))
f9f4320a	200
be8e71aa YO	201	#endif /* PLUGGABLE_RE_EXTENSION */
	202
	203	/* Stuff that needs to be included in the plugable extension goes below here */
	204
f8c7b90f	205	#ifdef PERL_OLD_COPY_ON_WRITE
ed252734 NC	206	#define RX_MATCH_COPY_FREE(rx) \
	207	STMT_START {if (rx->saved_copy) { \
	208	SV_CHECK_THINKFIRST_COW_DROP(rx->saved_copy); \
	209	} \
	210	if (RX_MATCH_COPIED(rx)) { \
	211	Safefree(rx->subbeg); \
	212	RX_MATCH_COPIED_off(rx); \
	213	}} STMT_END
	214	#else
	215	#define RX_MATCH_COPY_FREE(rx) \
	216	STMT_START {if (RX_MATCH_COPIED(rx)) { \
	217	Safefree(rx->subbeg); \
	218	RX_MATCH_COPIED_off(rx); \
	219	}} STMT_END
	220	#endif
	221
bbe252da YO	222	#define RX_MATCH_UTF8(prog) ((prog)->extflags & RXf_MATCH_UTF8)
	223	#define RX_MATCH_UTF8_on(prog) ((prog)->extflags \|= RXf_MATCH_UTF8)
	224	#define RX_MATCH_UTF8_off(prog) ((prog)->extflags &= ~RXf_MATCH_UTF8)
a30b2f1f RGS	225	#define RX_MATCH_UTF8_set(prog, t) ((t) \
	226	? (RX_MATCH_UTF8_on(prog), (PL_reg_match_utf8 = 1)) \
	227	: (RX_MATCH_UTF8_off(prog), (PL_reg_match_utf8 = 0)))
	228
cad2e5aa JH	229	#define REXEC_COPY_STR 0x01 /* Need to copy the string. */
	230	#define REXEC_CHECKED 0x02 /* check_substr already checked. */
	231	#define REXEC_SCREAM 0x04 /* use scream table. */
	232	#define REXEC_IGNOREPOS 0x08 /* \G matches at start. */
cf93c79d	233	#define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */
c277df42	234
155aba94	235	#define ReREFCNT_inc(re) ((void)(re && re->refcnt++), re)
f9f4320a	236	#define ReREFCNT_dec(re) CALLREGFREE(re)
cf93c79d IZ	237
cf93c79d IZ	238	#define FBMcf_TAIL_DOLLAR 1
cad2e5aa JH	239	#define FBMcf_TAIL_DOLLARM 2
	240	#define FBMcf_TAIL_Z 4
	241	#define FBMcf_TAIL_z 8
	242	#define FBMcf_TAIL (FBMcf_TAIL_DOLLAR\|FBMcf_TAIL_DOLLARM\|FBMcf_TAIL_Z\|FBMcf_TAIL_z)
cf93c79d IZ	243
cf93c79d IZ	244	#define FBMrf_MULTILINE 1
cad2e5aa	245
5d9a96ca DM	246	/* an accepting state/position*/
	247	struct _reg_trie_accepted {
	248	U8 *endpos;
	249	U16 wordnum;
	250	};
	251	typedef struct _reg_trie_accepted reg_trie_accepted;
	252
3b0527fe DM	253	/* some basic information about the current match that is created by
	254	* Perl_regexec_flags and then passed to regtry(), regmatch() etc */
	255
	256	typedef struct {
	257	regexp *prog;
	258	char *bol;
	259	char *till;
	260	SV *sv;
	261	char *ganch;
24b23f37	262	char *cutpoint;
3b0527fe DM	263	} regmatch_info;
3b0527fe DM	264
5d9a96ca DM	265
	266	/* structures for holding and saving the state maintained by regmatch() */
	267
6bda09f9 YO	268	#define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 50
6bda09f9 YO	269
5d9a96ca DM	270	typedef I32 CHECKPOINT;
5d9a96ca DM	271
a0374537	272	typedef struct regmatch_state {
40a82448	273	int resume_state; /* where to jump to on return */
24d3c4a9	274	char locinput; / where to backtrack in string on failure */
e822a8b4 DM	275
e822a8b4 DM	276	union {
77cb431f DM	277
	278	/* this is a fake union member that matches the first element
	279	* of each member that needs to store positive backtrack
	280	* information */
	281	struct {
	282	struct regmatch_state *prev_yes_state;
	283	} yes;
	284
e822a8b4	285	struct {
5d458dd8 YO	286	/* this first element must match u.yes */
5d458dd8 YO	287	struct regmatch_state *prev_yes_state;
e822a8b4	288	reg_trie_accepted *accept_buff;
166ba7cd	289	U32 accepted; /* how many accepting states we have seen */
7f69552c	290	U16 jump; / positive offsets from me */
166ba7cd	291	regnode B; / node following the trie */
7f69552c	292	regnode me; / Which node am I - needed for jump tries*/
e822a8b4 DM	293	} trie;
	294
	295	struct {
77cb431f DM	296	/* this first element must match u.yes */
77cb431f DM	297	struct regmatch_state *prev_yes_state;
faec1544 DM	298	struct regmatch_state *prev_eval;
faec1544 DM	299	struct regmatch_state *prev_curlyx;
aa283a38	300	regexp *prev_rex;
faec1544 DM	301	U32 toggle_reg_flags; /* what bits in PL_reg_flags to
	302	flip when transitioning between
	303	inner and outer rexen */
aa283a38 DM	304	CHECKPOINT cp; /* remember current savestack indexes */
aa283a38 DM	305	CHECKPOINT lastcp;
40a82448	306	regnode B; / the node following us */
6bda09f9	307	U32 close_paren; /* which close bracket is our end */
e822a8b4 DM	308	} eval;
	309
	310	struct {
c476f425	311	/* this first element must match u.yes */
262b90c4	312	struct regmatch_state *prev_yes_state;
c476f425 DM	313	struct regmatch_state prev_curlyx; / previous cur_curlyx */
	314	CHECKPOINT cp; /* remember current savestack index */
	315	bool minmod;
a0374537	316	int parenfloor;/* how far back to strip paren data */
c476f425 DM	317	int min; /* the minimal number of A's to match */
	318	int max; /* the maximal number of A's to match */
	319	regnode A, B; /* the nodes corresponding to /AB/ /
	320
	321	/* these two are modified by WHILEM */
	322	int count; /* how many instances of A we've matched */
	323	char lastloc;/ where previous A matched (0-len detect) */
e822a8b4 DM	324	} curlyx;
	325
	326	struct {
c476f425	327	/* this first element must match u.yes */
262b90c4	328	struct regmatch_state *prev_yes_state;
c476f425 DM	329	struct regmatch_state *save_curlyx;
	330	CHECKPOINT cp; /* remember current savestack indexes */
	331	CHECKPOINT lastcp;
	332	char save_lastloc; / previous curlyx.lastloc */
	333	I32 cache_offset;
	334	I32 cache_mask;
e822a8b4 DM	335	} whilem;
	336
	337	struct {
5d458dd8 YO	338	/* this first element must match u.yes */
5d458dd8 YO	339	struct regmatch_state *prev_yes_state;
3b6647e0	340	U32 lastparen;
40a82448 DM	341	regnode next_branch; / next branch node */
	342	CHECKPOINT cp;
	343	} branch;
	344
	345	struct {
dad79028 DM	346	/* this first element must match u.yes */
dad79028 DM	347	struct regmatch_state *prev_yes_state;
e822a8b4	348	I32 c1, c2; /* case fold search */
40a82448 DM	349	CHECKPOINT cp;
	350	I32 alen; /* length of first-matched A string */
	351	I32 count;
0cadcf80	352	bool minmod;
40a82448 DM	353	regnode A, B; /* the nodes corresponding to /AB/ /
40a82448 DM	354	regnode me; / the curlym node */
e822a8b4 DM	355	} curlym;
	356
	357	struct {
3b6647e0	358	U32 paren;
c255a977	359	CHECKPOINT cp;
e822a8b4	360	I32 c1, c2; /* case fold search */
c255a977 DM	361	char maxpos; / highest possible point in string to match */
c255a977 DM	362	char oldloc; / the previous locinput */
e822a8b4	363	int count;
c255a977 DM	364	int min, max; /* {m,n} */
	365	regnode A, B; /* the nodes corresponding to /AB/ /
	366	} curly; /* and CURLYN/PLUS/STAR */
dad79028 DM	367
	368	struct {
	369	/* this first element must match u.yes */
	370	struct regmatch_state *prev_yes_state;
	371	I32 wanted;
24d3c4a9	372	I32 logical; /* saved copy of 'logical' var */
40a82448	373	regnode me; / the IFMATCH/SUSPEND/UNLESSM node */
dad79028	374	} ifmatch; /* and SUSPEND/UNLESSM */
e2e6a0f1 YO	375
	376	struct {
	377	/* this first element must match u.yes */
	378	struct regmatch_state *prev_yes_state;
	379	struct regmatch_state *prev_mark;
	380	SV* mark_name;
	381	char *mark_loc;
	382	} mark;
d8319b27	383	} u;
5d9a96ca DM	384	} regmatch_state;
	385
	386	/* how many regmatch_state structs to allocate as a single slab.
	387	* We do it in 4K blocks for efficiency. The "3" is 2 for the next/prev
	388	* pointers, plus 1 for any mythical malloc overhead. */
	389
	390	#define PERL_REGMATCH_SLAB_SLOTS \
	391	((4096 - 3 * sizeof (void*)) / sizeof(regmatch_state))
	392
	393	typedef struct regmatch_slab {
	394	regmatch_state states[PERL_REGMATCH_SLAB_SLOTS];
	395	struct regmatch_slab prev, next;
	396	} regmatch_slab;
1ade1aa1	397
46ab3289 NC	398	#define PL_reg_flags PL_reg_state.re_state_reg_flags
	399	#define PL_bostr PL_reg_state.re_state_bostr
	400	#define PL_reginput PL_reg_state.re_state_reginput
46ab3289 NC	401	#define PL_regeol PL_reg_state.re_state_regeol
	402	#define PL_regstartp PL_reg_state.re_state_regstartp
	403	#define PL_regendp PL_reg_state.re_state_regendp
	404	#define PL_reglastparen PL_reg_state.re_state_reglastparen
	405	#define PL_reglastcloseparen PL_reg_state.re_state_reglastcloseparen
46ab3289 NC	406	#define PL_reg_start_tmp PL_reg_state.re_state_reg_start_tmp
	407	#define PL_reg_start_tmpl PL_reg_state.re_state_reg_start_tmpl
	408	#define PL_reg_eval_set PL_reg_state.re_state_reg_eval_set
46ab3289 NC	409	#define PL_reg_match_utf8 PL_reg_state.re_state_reg_match_utf8
	410	#define PL_reg_magic PL_reg_state.re_state_reg_magic
	411	#define PL_reg_oldpos PL_reg_state.re_state_reg_oldpos
	412	#define PL_reg_oldcurpm PL_reg_state.re_state_reg_oldcurpm
	413	#define PL_reg_curpm PL_reg_state.re_state_reg_curpm
	414	#define PL_reg_oldsaved PL_reg_state.re_state_reg_oldsaved
	415	#define PL_reg_oldsavedlen PL_reg_state.re_state_reg_oldsavedlen
	416	#define PL_reg_maxiter PL_reg_state.re_state_reg_maxiter
	417	#define PL_reg_leftiter PL_reg_state.re_state_reg_leftiter
	418	#define PL_reg_poscache PL_reg_state.re_state_reg_poscache
	419	#define PL_reg_poscache_size PL_reg_state.re_state_reg_poscache_size
	420	#define PL_regsize PL_reg_state.re_state_regsize
	421	#define PL_reg_starttry PL_reg_state.re_state_reg_starttry
	422	#define PL_nrs PL_reg_state.re_state_nrs
	423
1ade1aa1 NC	424	struct re_save_state {
	425	U32 re_state_reg_flags; /* from regexec.c */
	426	char *re_state_bostr;
	427	char re_state_reginput; / String-input pointer. */
1ade1aa1 NC	428	char re_state_regeol; / End of input, for $ check. */
	429	I32 re_state_regstartp; / Pointer to startp array. */
	430	I32 re_state_regendp; / Ditto for endp. */
	431	U32 re_state_reglastparen; / Similarly for lastparen. */
	432	U32 re_state_reglastcloseparen; / Similarly for lastcloseparen. */
1ade1aa1 NC	433	char *re_state_reg_start_tmp; / from regexec.c */
	434	U32 re_state_reg_start_tmpl; /* from regexec.c */
	435	I32 re_state_reg_eval_set; /* from regexec.c */
1ade1aa1 NC	436	bool re_state_reg_match_utf8; /* from regexec.c */
	437	MAGIC re_state_reg_magic; / from regexec.c */
	438	I32 re_state_reg_oldpos; /* from regexec.c */
	439	PMOP re_state_reg_oldcurpm; / from regexec.c */
	440	PMOP re_state_reg_curpm; / from regexec.c */
	441	char re_state_reg_oldsaved; / old saved substr during match */
	442	STRLEN re_state_reg_oldsavedlen; /* old length of saved substr during match */
	443	I32 re_state_reg_maxiter; /* max wait until caching pos */
	444	I32 re_state_reg_leftiter; /* wait until caching pos */
	445	char re_state_reg_poscache; / cache of pos of WHILEM */
	446	STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */
3b6647e0	447	U32 re_state_regsize; /* from regexec.c */
1ade1aa1 NC	448	char re_state_reg_starttry; / from regexec.c */
1ade1aa1 NC	449	#ifdef PERL_OLD_COPY_ON_WRITE
46ab3289	450	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	451	#endif
	452	};
	453
	454	#define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \
	455	(1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack)))
	456	/*
	457	* Local variables:
	458	* c-indentation-style: bsd
	459	* c-basic-offset: 4
	460	* indent-tabs-mode: t
	461	* End:
	462	*
	463	* ex: set ts=8 sts=4 sw=4 noet:
	464	*/