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

/*    regcomp.h
 *
 *    Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 *    2000, 2001, 2002, 2003, 2005 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.
 *
 */

typedef OP OP_4tree;			/* Will be redefined later. */

/*
 * The "internal use only" fields in regexp.h are present to pass info from
 * compile to execute that permits the execute phase to run lots faster on
 * simple cases.  They are:
 *
 * regstart	sv that must begin a match; NULL if none obvious
 * reganch	is the match anchored (at beginning-of-line only)?
 * regmust	string (pointer into program) that match must include, or NULL
 *  [regmust changed to SV* for bminstr()--law]
 * regmlen	length of regmust string
 *  [regmlen not used currently]
 *
 * Regstart and reganch permit very fast decisions on suitable starting points
 * for a match, cutting down the work a lot.  Regmust permits fast rejection
 * of lines that cannot possibly match.  The regmust tests are costly enough
 * that pregcomp() supplies a regmust only if the r.e. contains something
 * potentially expensive (at present, the only such thing detected is * or +
 * at the start of the r.e., which can involve a lot of backup).  Regmlen is
 * supplied because the test in pregexec() needs it and pregcomp() is computing
 * it anyway.
 * [regmust is now supplied always.  The tests that use regmust have a
 * heuristic that disables the test if it usually matches.]
 *
 * [In fact, we now use regmust in many cases to locate where the search
 * starts in the string, so if regback is >= 0, the regmust search is never
 * wasted effort.  The regback variable says how many characters back from
 * where regmust matched is the earliest possible start of the match.
 * For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.]
 */

/*
 * Structure for regexp "program".  This is essentially a linear encoding
 * of a nondeterministic finite-state machine (aka syntax charts or
 * "railroad normal form" in parsing technology).  Each node is an opcode
 * plus a "next" pointer, possibly plus an operand.  "Next" pointers of
 * all nodes except BRANCH implement concatenation; a "next" pointer with
 * a BRANCH on both ends of it is connecting two alternatives.  (Here we
 * have one of the subtle syntax dependencies:  an individual BRANCH (as
 * opposed to a collection of them) is never concatenated with anything
 * because of operator precedence.)  The operand of some types of node is
 * a literal string; for others, it is a node leading into a sub-FSM.  In
 * particular, the operand of a BRANCH node is the first node of the branch.
 * (NB this is *not* a tree structure:  the tail of the branch connects
 * to the thing following the set of BRANCHes.)  The opcodes are defined
 * in regnodes.h which is generated from regcomp.sym by regcomp.pl.
 */

/*
 * A node is one char of opcode followed by two chars of "next" pointer.
 * "Next" pointers are stored as two 8-bit pieces, high order first.  The
 * value is a positive offset from the opcode of the node containing it.
 * An operand, if any, simply follows the node.  (Note that much of the
 * code generation knows about this implicit relationship.)
 *
 * Using two bytes for the "next" pointer is vast overkill for most things,
 * but allows patterns to get big without disasters.
 *
 * [The "next" pointer is always aligned on an even
 * boundary, and reads the offset directly as a short.  Also, there is no
 * special test to reverse the sign of BACK pointers since the offset is
 * stored negative.]
 */

struct regnode_string {
    U8	str_len;
    U8  type;
    U16 next_off;
    char string[1];
};

struct regnode_1 {
    U8	flags;
    U8  type;
    U16 next_off;
    U32 arg1;
};

struct regnode_2 {
    U8	flags;
    U8  type;
    U16 next_off;
    U16 arg1;
    U16 arg2;
};

#define ANYOF_BITMAP_SIZE	32	/* 256 b/(8 b/B) */
#define ANYOF_CLASSBITMAP_SIZE	 4	/* up to 32 (8*4) named classes */

struct regnode_charclass {
    U8	flags;
    U8  type;
    U16 next_off;
    U32 arg1;
    char bitmap[ANYOF_BITMAP_SIZE];	/* only compile-time */
};

struct regnode_charclass_class {	/* has [[:blah:]] classes */
    U8	flags;				/* should have ANYOF_CLASS here */
    U8  type;
    U16 next_off;
    U32 arg1;
    char bitmap[ANYOF_BITMAP_SIZE];		/* both compile-time */
    char classflags[ANYOF_CLASSBITMAP_SIZE];	/* and run-time */
};

/* XXX fix this description.
   Impose a limit of REG_INFTY on various pattern matching operations
   to limit stack growth and to avoid "infinite" recursions.
*/
/* The default size for REG_INFTY is I16_MAX, which is the same as
   SHORT_MAX (see perl.h).  Unfortunately I16 isn't necessarily 16 bits
   (see handy.h).  On the Cray C90, sizeof(short)==4 and hence I16_MAX is
   ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
   ((1<<63)-1).  To limit stack growth to reasonable sizes, supply a
   smaller default.
	--Andy Dougherty  11 June 1998
*/
#if SHORTSIZE > 2
#  ifndef REG_INFTY
#    define REG_INFTY ((1<<15)-1)
#  endif
#endif

#ifndef REG_INFTY
#  define REG_INFTY I16_MAX
#endif

#define ARG_VALUE(arg) (arg)
#define ARG__SET(arg,val) ((arg) = (val))

#undef ARG
#undef ARG1
#undef ARG2

#define ARG(p) ARG_VALUE(ARG_LOC(p))
#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))

#undef NEXT_OFF
#undef NODE_ALIGN

#define NEXT_OFF(p) ((p)->next_off)
#define NODE_ALIGN(node)
#define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */

#define SIZE_ALIGN NODE_ALIGN

#undef OP
#undef OPERAND
#undef MASK
#undef STRING

#define	OP(p)		((p)->type)
#define	OPERAND(p)	(((struct regnode_string *)p)->string)
#define MASK(p)		((char*)OPERAND(p))
#define	STR_LEN(p)	(((struct regnode_string *)p)->str_len)
#define	STRING(p)	(((struct regnode_string *)p)->string)
#define STR_SZ(l)	((l + sizeof(regnode) - 1) / sizeof(regnode))
#define NODE_SZ_STR(p)	(STR_SZ(STR_LEN(p))+1)

#undef NODE_ALIGN
#undef ARG_LOC
#undef NEXTOPER
#undef PREVOPER

#define	NODE_ALIGN(node)
#define	ARG_LOC(p)	(((struct regnode_1 *)p)->arg1)
#define	ARG1_LOC(p)	(((struct regnode_2 *)p)->arg1)
#define	ARG2_LOC(p)	(((struct regnode_2 *)p)->arg2)
#define NODE_STEP_REGNODE	1	/* sizeof(regnode)/sizeof(regnode) */
#define EXTRA_STEP_2ARGS	EXTRA_SIZE(struct regnode_2)

#define NODE_STEP_B	4

#define	NEXTOPER(p)	((p) + NODE_STEP_REGNODE)
#define	PREVOPER(p)	((p) - NODE_STEP_REGNODE)

#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
    (ptr)->type = op;    (ptr)->next_off = 0;   (ptr)++; } STMT_END
#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
    ARG_SET(ptr, arg);  FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END

#define REG_MAGIC 0234

#define SIZE_ONLY (RExC_emit == &PL_regdummy)

/* Flags for node->flags of ANYOF */

#define ANYOF_CLASS		0x08	/* has [[:blah:]] classes */
#define ANYOF_INVERT		0x04
#define ANYOF_FOLD		0x02
#define ANYOF_LOCALE		0x01

/* Used for regstclass only */
#define ANYOF_EOS		0x10		/* Can match an empty string too */

/* There is a character or a range past 0xff */
#define ANYOF_UNICODE		0x20
#define ANYOF_UNICODE_ALL	0x40	/* Can match any char past 0xff */

/* size of node is large (includes class pointer) */
#define ANYOF_LARGE 		0x80

/* Are there any runtime flags on in this node? */
#define ANYOF_RUNTIME(s)	(ANYOF_FLAGS(s) & 0x0f)

#define ANYOF_FLAGS_ALL		0xff

/* Character classes for node->classflags of ANYOF */
/* Should be synchronized with a table in regprop() */
/* 2n should pair with 2n+1 */

#define ANYOF_ALNUM	 0	/* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
#define ANYOF_NALNUM	 1
#define ANYOF_SPACE	 2	/* \s */
#define ANYOF_NSPACE	 3
#define ANYOF_DIGIT	 4
#define ANYOF_NDIGIT	 5
#define ANYOF_ALNUMC	 6	/* isalnum(3), utf8::IsAlnum, ALNUMC */
#define ANYOF_NALNUMC	 7
#define ANYOF_ALPHA	 8
#define ANYOF_NALPHA	 9
#define ANYOF_ASCII	10
#define ANYOF_NASCII	11
#define ANYOF_CNTRL	12
#define ANYOF_NCNTRL	13
#define ANYOF_GRAPH	14
#define ANYOF_NGRAPH	15
#define ANYOF_LOWER	16
#define ANYOF_NLOWER	17
#define ANYOF_PRINT	18
#define ANYOF_NPRINT	19
#define ANYOF_PUNCT	20
#define ANYOF_NPUNCT	21
#define ANYOF_UPPER	22
#define ANYOF_NUPPER	23
#define ANYOF_XDIGIT	24
#define ANYOF_NXDIGIT	25
#define ANYOF_PSXSPC	26	/* POSIX space: \s plus the vertical tab */
#define ANYOF_NPSXSPC	27
#define ANYOF_BLANK	28	/* GNU extension: space and tab: non-vertical space */
#define ANYOF_NBLANK	29

#define ANYOF_MAX	32

/* Backward source code compatibility. */

#define ANYOF_ALNUML	 ANYOF_ALNUM
#define ANYOF_NALNUML	 ANYOF_NALNUM
#define ANYOF_SPACEL	 ANYOF_SPACE
#define ANYOF_NSPACEL	 ANYOF_NSPACE

/* Utility macros for the bitmap and classes of ANYOF */

#define ANYOF_SIZE		(sizeof(struct regnode_charclass))
#define ANYOF_CLASS_SIZE	(sizeof(struct regnode_charclass_class))

#define ANYOF_FLAGS(p)		((p)->flags)

#define ANYOF_BIT(c)		(1 << ((c) & 7))

#define ANYOF_CLASS_BYTE(p, c)	(((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3])
#define ANYOF_CLASS_SET(p, c)	(ANYOF_CLASS_BYTE(p, c) |=  ANYOF_BIT(c))
#define ANYOF_CLASS_CLEAR(p, c)	(ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
#define ANYOF_CLASS_TEST(p, c)	(ANYOF_CLASS_BYTE(p, c) &   ANYOF_BIT(c))

#define ANYOF_CLASS_ZERO(ret)	Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char)
#define ANYOF_BITMAP_ZERO(ret)	Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)

#define ANYOF_BITMAP(p)		(((struct regnode_charclass*)(p))->bitmap)
#define ANYOF_BITMAP_BYTE(p, c)	(ANYOF_BITMAP(p)[((c) >> 3) & 31])
#define ANYOF_BITMAP_SET(p, c)	(ANYOF_BITMAP_BYTE(p, c) |=  ANYOF_BIT(c))
#define ANYOF_BITMAP_CLEAR(p,c)	(ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
#define ANYOF_BITMAP_TEST(p, c)	(ANYOF_BITMAP_BYTE(p, c) &   ANYOF_BIT(c))

#define ANYOF_BITMAP_SETALL(p)		\
	memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE)
#define ANYOF_BITMAP_CLEARALL(p)	\
	Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE)
/* Check that all 256 bits are all set.  Used in S_cl_is_anything()  */
#define ANYOF_BITMAP_TESTALLSET(p)	\
	memEQ (ANYOF_BITMAP(p), "\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377", ANYOF_BITMAP_SIZE)

#define ANYOF_SKIP		((ANYOF_SIZE - 1)/sizeof(regnode))
#define ANYOF_CLASS_SKIP	((ANYOF_CLASS_SIZE - 1)/sizeof(regnode))
#define ANYOF_CLASS_ADD_SKIP	(ANYOF_CLASS_SKIP - ANYOF_SKIP)

/*
 * Utility definitions.
 */
#ifndef CHARMASK
#  define UCHARAT(p)	((int)*(const U8*)(p))
#else
#  define UCHARAT(p)	((int)*(p)&CHARMASK)
#endif

#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))

#define REG_SEEN_ZERO_LEN	 1
#define REG_SEEN_LOOKBEHIND	 2
#define REG_SEEN_GPOS		 4
#define REG_SEEN_EVAL		 8
#define REG_SEEN_CANY		16
#define REG_SEEN_SANY		REG_SEEN_CANY /* src bckwrd cmpt */

START_EXTERN_C

#include "regnodes.h"

/* The following have no fixed length. U8 so we can do strchr() on it. */
#ifndef DOINIT
EXTCONST U8 PL_varies[];
#else
EXTCONST U8 PL_varies[] = {
    BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL,
    WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0
};
#endif

/* The following always have a length of 1. U8 we can do strchr() on it. */
/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
#ifndef DOINIT
EXTCONST U8 PL_simple[];
#else
EXTCONST U8 PL_simple[] = {
    REG_ANY,	SANY,	CANY,
    ANYOF,
    ALNUM,	ALNUML,
    NALNUM,	NALNUML,
    SPACE,	SPACEL,
    NSPACE,	NSPACEL,
    DIGIT,	NDIGIT,
    0
};
#endif

END_EXTERN_C

typedef struct re_scream_pos_data_s
{
    char **scream_olds;		/* match pos */
    I32 *scream_pos;		/* Internal iterator of scream. */
} re_scream_pos_data;

/* .what is a character array with one character for each member of .data
 * The character describes the function of the corresponding .data item:
 *   f - start-class data for regstclass optimization
 *   n - Root of op tree for (?{EVAL}) item
 *   o - Start op for (?{EVAL}) item
 *   p - Pad for (?{EVAL} item
 *   s - swash for unicode-style character class, and the multicharacter
 *       strings resulting from casefolding the single-character entries
 *       in the character class
 *   t - trie struct
 * 20010712 mjd@plover.com
 * (Remember to update re_dup() and pregfree() if you add any items.)
 */
struct reg_data {
    U32 count;
    U8 *what;
    void* data[1];
};

struct reg_substr_datum {
    I32 min_offset;
    I32 max_offset;
    SV *substr;		/* non-utf8 variant */
    SV *utf8_substr;	/* utf8 variant */
};

struct reg_substr_data {
    struct reg_substr_datum data[3];	/* Actual array */
};

#define anchored_substr substrs->data[0].substr
#define anchored_utf8 substrs->data[0].utf8_substr
#define anchored_offset substrs->data[0].min_offset
#define float_substr substrs->data[1].substr
#define float_utf8 substrs->data[1].utf8_substr
#define float_min_offset substrs->data[1].min_offset
#define float_max_offset substrs->data[1].max_offset
#define check_substr substrs->data[2].substr
#define check_utf8 substrs->data[2].utf8_substr
#define check_offset_min substrs->data[2].min_offset
#define check_offset_max substrs->data[2].max_offset


/* trie related stuff */

/* a transition record for the state machine. the
   check field determines which state "owns" the
   transition. the char the transition is for is
   determined by offset from the owning states base
   field.  the next field determines which state
   is to be transitioned to if any.
*/
struct _reg_trie_trans {
  U32 next;
  U32 check;
};

/* a transition list element for the list based representation */
struct _reg_trie_trans_list_elem {
    U16 forid;
    U32 newstate;
};
typedef struct _reg_trie_trans_list_elem reg_trie_trans_le;

/* a state for compressed nodes. base is an offset
  into an array of reg_trie_trans array. If wordnum is
  nonzero the state is accepting. if base is zero then
  the state has no children (and will be accepting)
*/
struct _reg_trie_state {
  U16 wordnum;
  union {
    U32                base;
    reg_trie_trans_le* list;
  } trans;
};


typedef struct _reg_trie_state    reg_trie_state;
typedef struct _reg_trie_trans    reg_trie_trans;


/* anything in here that needs to be freed later
should be dealt with in pregfree */
struct _reg_trie_data {
    U16              uniquecharcount;
    U16              wordcount;
    STRLEN           charcount;
    U32              laststate;
    U32              lasttrans;
    U16              *charmap;
    HV               *widecharmap;
    reg_trie_state   *states;
    reg_trie_trans   *trans;
    U32              refcount;
#ifdef DEBUGGING
    AV               *words;
    AV               *revcharmap;
#endif
};

typedef struct _reg_trie_data reg_trie_data;

/* these defines assume uniquecharcount is the correct variable, and state may be evaluated twice */
#define TRIE_NODENUM(state) (((state)-1)/(trie->uniquecharcount)+1)
#define SAFE_TRIE_NODENUM(state) ((state) ? (((state)-1)/(trie->uniquecharcount)+1) : (state))
#define TRIE_NODEIDX(state) ((state) ? (((state)-1)*(trie->uniquecharcount)+1) : (state))

#define DO_TRIE 1
#define TRIE_DEBUG 1

#define RE_TRIE_MAXBUF_INIT 65536
#define RE_TRIE_MAXBUF_NAME "\022E_TRIE_MAXBUF"
#define RE_DEBUG_FLAGS "\022E_DEBUG_FLAGS"

/* If you change these be sure to update ext/re/re.pm as well */
#define RE_DEBUG_COMPILE       1
#define RE_DEBUG_EXECUTE       2
#define RE_DEBUG_TRIE_COMPILE  4
#define RE_DEBUG_TRIE_EXECUTE  8
#define RE_DEBUG_TRIE_MORE    16
#define RE_DEBUG_OPTIMISE     32
#define RE_DEBUG_OFFSETS      64

#define DEBUG_OPTIMISE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OPTIMISE) x  )
#define DEBUG_EXECUTE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_EXECUTE) x  )
#define DEBUG_COMPILE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_COMPILE) x  )
#define DEBUG_OFFSETS_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OFFSETS) x  )
#define DEBUG_TRIE_r(x) DEBUG_r( \
   if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE       \
       || SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE )  \
   x  \
)
#define DEBUG_TRIE_EXECUTE_r(x) \
    DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE) x )

#define DEBUG_TRIE_COMPILE_r(x) \
    DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE) x )

#define DEBUG_TRIE_EXECUTE_MORE_r(x) \
    DEBUG_TRIE_EXECUTE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x )

#define DEBUG_TRIE_COMPILE_MORE_r(x) \
    DEBUG_TRIE_COMPILE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x )

#define GET_RE_DEBUG_FLAGS DEBUG_r( \
        re_debug_flags=get_sv(RE_DEBUG_FLAGS, 1); \
        if (!SvIOK(re_debug_flags)) { \
            sv_setiv(re_debug_flags, RE_DEBUG_COMPILE | RE_DEBUG_EXECUTE | RE_DEBUG_OFFSETS); \
        } \
    )


#ifdef DEBUGGING
#define GET_RE_DEBUG_FLAGS_DECL SV *re_debug_flags = NULL; GET_RE_DEBUG_FLAGS;
#else
#define GET_RE_DEBUG_FLAGS_DECL
#endif
Commit	Line	Data
a0d0e21e	1	/* regcomp.h
d6376244	2	*
4bb101f2	3	* Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
1d325971	4	* 2000, 2001, 2002, 2003, 2005 by Larry Wall and others
d6376244	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	*
a687059c	9	*/
a687059c	10
c277df42	11	typedef OP OP_4tree; /* Will be redefined later. */
c277df42	12
a687059c	13	/*
	14	* The "internal use only" fields in regexp.h are present to pass info from
	15	* compile to execute that permits the execute phase to run lots faster on
	16	* simple cases. They are:
	17	*
bd61b366	18	* regstart sv that must begin a match; NULL if none obvious
a687059c	19	* reganch is the match anchored (at beginning-of-line only)?
a687059c	20	* regmust string (pointer into program) that match must include, or NULL
79072805	21	* [regmust changed to SV* for bminstr()--law]
a687059c	22	* regmlen length of regmust string
	23	* [regmlen not used currently]
	24	*
	25	* Regstart and reganch permit very fast decisions on suitable starting points
	26	* for a match, cutting down the work a lot. Regmust permits fast rejection
	27	* of lines that cannot possibly match. The regmust tests are costly enough
e50aee73	28	* that pregcomp() supplies a regmust only if the r.e. contains something
a687059c	29	* potentially expensive (at present, the only such thing detected is * or +
a687059c	30	* at the start of the r.e., which can involve a lot of backup). Regmlen is
e50aee73	31	* supplied because the test in pregexec() needs it and pregcomp() is computing
a687059c	32	* it anyway.
	33	* [regmust is now supplied always. The tests that use regmust have a
	34	* heuristic that disables the test if it usually matches.]
	35	*
	36	* [In fact, we now use regmust in many cases to locate where the search
	37	* starts in the string, so if regback is >= 0, the regmust search is never
	38	* wasted effort. The regback variable says how many characters back from
	39	* where regmust matched is the earliest possible start of the match.
	40	* For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.]
	41	*/
	42
	43	/*
	44	* Structure for regexp "program". This is essentially a linear encoding
	45	* of a nondeterministic finite-state machine (aka syntax charts or
	46	* "railroad normal form" in parsing technology). Each node is an opcode
	47	* plus a "next" pointer, possibly plus an operand. "Next" pointers of
	48	* all nodes except BRANCH implement concatenation; a "next" pointer with
	49	* a BRANCH on both ends of it is connecting two alternatives. (Here we
	50	* have one of the subtle syntax dependencies: an individual BRANCH (as
	51	* opposed to a collection of them) is never concatenated with anything
	52	* because of operator precedence.) The operand of some types of node is
	53	* a literal string; for others, it is a node leading into a sub-FSM. In
	54	* particular, the operand of a BRANCH node is the first node of the branch.
	55	* (NB this is not a tree structure: the tail of the branch connects
a3621e74	56	* to the thing following the set of BRANCHes.) The opcodes are defined
a3621e74	57	* in regnodes.h which is generated from regcomp.sym by regcomp.pl.
a687059c	58	*/
a687059c	59
a687059c	60	/*
	61	* A node is one char of opcode followed by two chars of "next" pointer.
	62	* "Next" pointers are stored as two 8-bit pieces, high order first. The
	63	* value is a positive offset from the opcode of the node containing it.
	64	* An operand, if any, simply follows the node. (Note that much of the
	65	* code generation knows about this implicit relationship.)
	66	*
	67	* Using two bytes for the "next" pointer is vast overkill for most things,
	68	* but allows patterns to get big without disasters.
	69	*
e91177ed	70	* [The "next" pointer is always aligned on an even
a687059c	71	* boundary, and reads the offset directly as a short. Also, there is no
	72	* special test to reverse the sign of BACK pointers since the offset is
	73	* stored negative.]
	74	*/
	75
c277df42	76	struct regnode_string {
cd439c50	77	U8 str_len;
c277df42	78	U8 type;
c277df42	79	U16 next_off;
cd439c50	80	char string[1];
c277df42	81	};
a687059c	82
c277df42	83	struct regnode_1 {
	84	U8 flags;
	85	U8 type;
	86	U16 next_off;
	87	U32 arg1;
	88	};
	89
	90	struct regnode_2 {
	91	U8 flags;
	92	U8 type;
	93	U16 next_off;
	94	U16 arg1;
	95	U16 arg2;
	96	};
	97
936ed897	98	#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */
ffc61ed2	99	#define ANYOF_CLASSBITMAP_SIZE 4 /* up to 32 (84) named classes /
936ed897	100
	101	struct regnode_charclass {
	102	U8 flags;
	103	U8 type;
	104	U16 next_off;
ffc61ed2	105	U32 arg1;
19103b91	106	char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */
936ed897	107	};
936ed897	108
19103b91	109	struct regnode_charclass_class { /* has [[:blah:]] classes */
19103b91	110	U8 flags; /* should have ANYOF_CLASS here */
936ed897	111	U8 type;
936ed897	112	U16 next_off;
ffc61ed2	113	U32 arg1;
19103b91	114	char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time */
19103b91	115	char classflags[ANYOF_CLASSBITMAP_SIZE]; /* and run-time */
936ed897	116	};
936ed897	117
9c7e81e8	118	/* XXX fix this description.
	119	Impose a limit of REG_INFTY on various pattern matching operations
	120	to limit stack growth and to avoid "infinite" recursions.
	121	*/
	122	/* The default size for REG_INFTY is I16_MAX, which is the same as
	123	SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
	124	(see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
	125	((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
	126	((1<<63)-1). To limit stack growth to reasonable sizes, supply a
	127	smaller default.
	128	--Andy Dougherty 11 June 1998
	129	*/
	130	#if SHORTSIZE > 2
	131	# ifndef REG_INFTY
	132	# define REG_INFTY ((1<<15)-1)
	133	# endif
	134	#endif
	135
	136	#ifndef REG_INFTY
83cfe48c	137	# define REG_INFTY I16_MAX
83cfe48c	138	#endif
a687059c	139
e91177ed	140	#define ARG_VALUE(arg) (arg)
e91177ed	141	#define ARG__SET(arg,val) ((arg) = (val))
c277df42	142
e8e3b0fb	143	#undef ARG
	144	#undef ARG1
	145	#undef ARG2
	146
c277df42	147	#define ARG(p) ARG_VALUE(ARG_LOC(p))
	148	#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
	149	#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
	150	#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
	151	#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
	152	#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
	153
e8e3b0fb	154	#undef NEXT_OFF
	155	#undef NODE_ALIGN
	156
075abff3	157	#define NEXT_OFF(p) ((p)->next_off)
	158	#define NODE_ALIGN(node)
	159	#define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */
a687059c	160
c277df42	161	#define SIZE_ALIGN NODE_ALIGN
c277df42	162
e8e3b0fb	163	#undef OP
	164	#undef OPERAND
	165	#undef MASK
	166	#undef STRING
	167
e91177ed	168	#define OP(p) ((p)->type)
e91177ed	169	#define OPERAND(p) (((struct regnode_string *)p)->string)
cd439c50	170	#define MASK(p) ((char*)OPERAND(p))
	171	#define STR_LEN(p) (((struct regnode_string *)p)->str_len)
	172	#define STRING(p) (((struct regnode_string *)p)->string)
	173	#define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode))
	174	#define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1)
	175
e8e3b0fb	176	#undef NODE_ALIGN
	177	#undef ARG_LOC
	178	#undef NEXTOPER
	179	#undef PREVOPER
	180
e91177ed	181	#define NODE_ALIGN(node)
	182	#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
	183	#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
	184	#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
	185	#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
	186	#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
	187
	188	#define NODE_STEP_B 4
c277df42	189
	190	#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
	191	#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
	192
e91177ed	193	#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42	194	(ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed	195	#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42	196	ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
a687059c	197
22c35a8c	198	#define REG_MAGIC 0234
a687059c	199
830247a4	200	#define SIZE_ONLY (RExC_emit == &PL_regdummy)
c277df42	201
936ed897	202	/* Flags for node->flags of ANYOF */
b8c5462f	203
19103b91	204	#define ANYOF_CLASS 0x08 /* has [[:blah:]] classes */
ffc61ed2	205	#define ANYOF_INVERT 0x04
	206	#define ANYOF_FOLD 0x02
	207	#define ANYOF_LOCALE 0x01
b8c5462f	208
653099ff	209	/* Used for regstclass only */
ffc61ed2	210	#define ANYOF_EOS 0x10 /* Can match an empty string too */
	211
	212	/* There is a character or a range past 0xff */
	213	#define ANYOF_UNICODE 0x20
1aa99e6b	214	#define ANYOF_UNICODE_ALL 0x40 /* Can match any char past 0xff */
ffc61ed2	215
4f66b38d	216	/* size of node is large (includes class pointer) */
	217	#define ANYOF_LARGE 0x80
	218
ffc61ed2	219	/* Are there any runtime flags on in this node? */
	220	#define ANYOF_RUNTIME(s) (ANYOF_FLAGS(s) & 0x0f)
	221
	222	#define ANYOF_FLAGS_ALL 0xff
653099ff	223
936ed897	224	/* Character classes for node->classflags of ANYOF */
653099ff	225	/* Should be synchronized with a table in regprop() */
653099ff	226	/* 2n should pair with 2n+1 */
b8c5462f	227
289d4f09	228	#define ANYOF_ALNUM 0 /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
b8c5462f	229	#define ANYOF_NALNUM 1
aaa51d5e	230	#define ANYOF_SPACE 2 /* \s */
b8c5462f	231	#define ANYOF_NSPACE 3
	232	#define ANYOF_DIGIT 4
	233	#define ANYOF_NDIGIT 5
289d4f09	234	#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */
b8c5462f	235	#define ANYOF_NALNUMC 7
	236	#define ANYOF_ALPHA 8
	237	#define ANYOF_NALPHA 9
	238	#define ANYOF_ASCII 10
	239	#define ANYOF_NASCII 11
	240	#define ANYOF_CNTRL 12
	241	#define ANYOF_NCNTRL 13
	242	#define ANYOF_GRAPH 14
	243	#define ANYOF_NGRAPH 15
	244	#define ANYOF_LOWER 16
	245	#define ANYOF_NLOWER 17
	246	#define ANYOF_PRINT 18
	247	#define ANYOF_NPRINT 19
	248	#define ANYOF_PUNCT 20
	249	#define ANYOF_NPUNCT 21
	250	#define ANYOF_UPPER 22
	251	#define ANYOF_NUPPER 23
	252	#define ANYOF_XDIGIT 24
	253	#define ANYOF_NXDIGIT 25
aaa51d5e	254	#define ANYOF_PSXSPC 26 /* POSIX space: \s plus the vertical tab */
aaa51d5e	255	#define ANYOF_NPSXSPC 27
ffc61ed2	256	#define ANYOF_BLANK 28 /* GNU extension: space and tab: non-vertical space */
289d4f09	257	#define ANYOF_NBLANK 29
b8c5462f	258
aaa51d5e	259	#define ANYOF_MAX 32
b8c5462f	260
	261	/* Backward source code compatibility. */
	262
	263	#define ANYOF_ALNUML ANYOF_ALNUM
	264	#define ANYOF_NALNUML ANYOF_NALNUM
	265	#define ANYOF_SPACEL ANYOF_SPACE
	266	#define ANYOF_NSPACEL ANYOF_NSPACE
	267
	268	/* Utility macros for the bitmap and classes of ANYOF */
	269
936ed897	270	#define ANYOF_SIZE (sizeof(struct regnode_charclass))
936ed897	271	#define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class))
b8c5462f	272
936ed897	273	#define ANYOF_FLAGS(p) ((p)->flags)
b8c5462f	274
	275	#define ANYOF_BIT(c) (1 << ((c) & 7))
	276
936ed897	277	#define ANYOF_CLASS_BYTE(p, c) (((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3])
b8c5462f	278	#define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) \|= ANYOF_BIT(c))
	279	#define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
	280	#define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c))
	281
936ed897	282	#define ANYOF_CLASS_ZERO(ret) Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char)
	283	#define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)
	284
	285	#define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap)
	286	#define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[((c) >> 3) & 31])
b8c5462f	287	#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) \|= ANYOF_BIT(c))
	288	#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
	289	#define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
	290
f8bef550	291	#define ANYOF_BITMAP_SETALL(p) \
	292	memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE)
	293	#define ANYOF_BITMAP_CLEARALL(p) \
	294	Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE)
	295	/* Check that all 256 bits are all set. Used in S_cl_is_anything() */
	296	#define ANYOF_BITMAP_TESTALLSET(p) \
	297	memEQ (ANYOF_BITMAP(p), "\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377", ANYOF_BITMAP_SIZE)
	298
936ed897	299	#define ANYOF_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode))
	300	#define ANYOF_CLASS_SKIP ((ANYOF_CLASS_SIZE - 1)/sizeof(regnode))
	301	#define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP)
c277df42	302
a687059c	303	/*
	304	* Utility definitions.
	305	*/
2304df62	306	#ifndef CHARMASK
075abff3	307	# define UCHARAT(p) ((int)(const U8)(p))
a687059c	308	#else
075abff3	309	# define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c	310	#endif
a687059c	311
c277df42	312	#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
c277df42	313
3baa4c62	314	#define REG_SEEN_ZERO_LEN 1
	315	#define REG_SEEN_LOOKBEHIND 2
	316	#define REG_SEEN_GPOS 4
	317	#define REG_SEEN_EVAL 8
cce850e4	318	#define REG_SEEN_CANY 16
cce850e4	319	#define REG_SEEN_SANY REG_SEEN_CANY /* src bckwrd cmpt */
d09b2d29	320
73c4f7a1	321	START_EXTERN_C
73c4f7a1	322
d09b2d29	323	#include "regnodes.h"
d09b2d29	324
4d61ec05	325	/* The following have no fixed length. U8 so we can do strchr() on it. */
d09b2d29	326	#ifndef DOINIT
4d61ec05	327	EXTCONST U8 PL_varies[];
d09b2d29	328	#else
4d61ec05	329	EXTCONST U8 PL_varies[] = {
0111c4fd	330	BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL,
a0ed51b3	331	WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0
c277df42	332	};
d09b2d29	333	#endif
c277df42	334
4d61ec05	335	/* The following always have a length of 1. U8 we can do strchr() on it. */
4d61ec05	336	/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
d09b2d29	337	#ifndef DOINIT
4d61ec05	338	EXTCONST U8 PL_simple[];
d09b2d29	339	#else
4d61ec05	340	EXTCONST U8 PL_simple[] = {
f33976b4	341	REG_ANY, SANY, CANY,
ffc61ed2	342	ANYOF,
	343	ALNUM, ALNUML,
	344	NALNUM, NALNUML,
	345	SPACE, SPACEL,
	346	NSPACE, NSPACEL,
	347	DIGIT, NDIGIT,
	348	0
c277df42	349	};
	350	#endif
	351
73c4f7a1	352	END_EXTERN_C
cad2e5aa	353
	354	typedef struct re_scream_pos_data_s
	355	{
	356	char *scream_olds; / match pos */
	357	I32 scream_pos; / Internal iterator of scream. */
	358	} re_scream_pos_data;
	359
261faec3	360	/* .what is a character array with one character for each member of .data
261faec3	361	* The character describes the function of the corresponding .data item:
0111c4fd	362	* f - start-class data for regstclass optimization
261faec3	363	* n - Root of op tree for (?{EVAL}) item
	364	* o - Start op for (?{EVAL}) item
	365	* p - Pad for (?{EVAL} item
9e55ce06	366	* s - swash for unicode-style character class, and the multicharacter
	367	* strings resulting from casefolding the single-character entries
	368	* in the character class
a3621e74	369	* t - trie struct
261faec3	370	* 20010712 mjd@plover.com
c3613ed7	371	* (Remember to update re_dup() and pregfree() if you add any items.)
261faec3	372	*/
cad2e5aa	373	struct reg_data {
	374	U32 count;
	375	U8 *what;
	376	void* data[1];
	377	};
	378
	379	struct reg_substr_datum {
	380	I32 min_offset;
	381	I32 max_offset;
33b8afdf	382	SV substr; / non-utf8 variant */
33b8afdf	383	SV utf8_substr; / utf8 variant */
cad2e5aa	384	};
	385
	386	struct reg_substr_data {
	387	struct reg_substr_datum data[3]; /* Actual array */
	388	};
	389
	390	#define anchored_substr substrs->data[0].substr
33b8afdf	391	#define anchored_utf8 substrs->data[0].utf8_substr
cad2e5aa	392	#define anchored_offset substrs->data[0].min_offset
cad2e5aa	393	#define float_substr substrs->data[1].substr
33b8afdf	394	#define float_utf8 substrs->data[1].utf8_substr
cad2e5aa	395	#define float_min_offset substrs->data[1].min_offset
	396	#define float_max_offset substrs->data[1].max_offset
	397	#define check_substr substrs->data[2].substr
33b8afdf	398	#define check_utf8 substrs->data[2].utf8_substr
cad2e5aa	399	#define check_offset_min substrs->data[2].min_offset
cad2e5aa	400	#define check_offset_max substrs->data[2].max_offset
a3621e74	401
	402
	403
	404	/* trie related stuff */
5d9a96ca	405
a3621e74	406	/* a transition record for the state machine. the
	407	check field determines which state "owns" the
	408	transition. the char the transition is for is
	409	determined by offset from the owning states base
	410	field. the next field determines which state
	411	is to be transitioned to if any.
	412	*/
	413	struct _reg_trie_trans {
	414	U32 next;
	415	U32 check;
	416	};
	417
	418	/* a transition list element for the list based representation */
	419	struct _reg_trie_trans_list_elem {
	420	U16 forid;
	421	U32 newstate;
	422	};
	423	typedef struct _reg_trie_trans_list_elem reg_trie_trans_le;
	424
	425	/* a state for compressed nodes. base is an offset
	426	into an array of reg_trie_trans array. If wordnum is
	427	nonzero the state is accepting. if base is zero then
	428	the state has no children (and will be accepting)
	429	*/
	430	struct _reg_trie_state {
	431	U16 wordnum;
	432	union {
	433	U32 base;
	434	reg_trie_trans_le* list;
	435	} trans;
	436	};
	437
	438
	439
a3621e74	440	typedef struct _reg_trie_state reg_trie_state;
	441	typedef struct _reg_trie_trans reg_trie_trans;
	442
	443
	444	/* anything in here that needs to be freed later
	445	should be dealt with in pregfree */
	446	struct _reg_trie_data {
	447	U16 uniquecharcount;
	448	U16 wordcount;
	449	STRLEN charcount;
	450	U32 laststate;
cc601c31	451	U32 lasttrans;
a3621e74	452	U16 *charmap;
	453	HV *widecharmap;
	454	reg_trie_state *states;
	455	reg_trie_trans *trans;
	456	U32 refcount;
	457	#ifdef DEBUGGING
	458	AV *words;
	459	AV *revcharmap;
	460	#endif
	461	};
	462
	463	typedef struct _reg_trie_data reg_trie_data;
	464
	465	/* these defines assume uniquecharcount is the correct variable, and state may be evaluated twice */
	466	#define TRIE_NODENUM(state) (((state)-1)/(trie->uniquecharcount)+1)
	467	#define SAFE_TRIE_NODENUM(state) ((state) ? (((state)-1)/(trie->uniquecharcount)+1) : (state))
	468	#define TRIE_NODEIDX(state) ((state) ? (((state)-1)*(trie->uniquecharcount)+1) : (state))
	469
	470	#define DO_TRIE 1
	471	#define TRIE_DEBUG 1
	472
0111c4fd	473	#define RE_TRIE_MAXBUF_INIT 65536
0111c4fd	474	#define RE_TRIE_MAXBUF_NAME "\022E_TRIE_MAXBUF"
a3621e74	475	#define RE_DEBUG_FLAGS "\022E_DEBUG_FLAGS"
	476
	477	/* If you change these be sure to update ext/re/re.pm as well */
	478	#define RE_DEBUG_COMPILE 1
	479	#define RE_DEBUG_EXECUTE 2
	480	#define RE_DEBUG_TRIE_COMPILE 4
	481	#define RE_DEBUG_TRIE_EXECUTE 8
	482	#define RE_DEBUG_TRIE_MORE 16
	483	#define RE_DEBUG_OPTIMISE 32
	484	#define RE_DEBUG_OFFSETS 64
	485
	486	#define DEBUG_OPTIMISE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OPTIMISE) x )
	487	#define DEBUG_EXECUTE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_EXECUTE) x )
	488	#define DEBUG_COMPILE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_COMPILE) x )
	489	#define DEBUG_OFFSETS_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OFFSETS) x )
	490	#define DEBUG_TRIE_r(x) DEBUG_r( \
	491	if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE \
	492	\|\| SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE ) \
	493	x \
	494	)
	495	#define DEBUG_TRIE_EXECUTE_r(x) \
	496	DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE) x )
	497
	498	#define DEBUG_TRIE_COMPILE_r(x) \
	499	DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE) x )
	500
	501	#define DEBUG_TRIE_EXECUTE_MORE_r(x) \
	502	DEBUG_TRIE_EXECUTE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x )
	503
	504	#define DEBUG_TRIE_COMPILE_MORE_r(x) \
	505	DEBUG_TRIE_COMPILE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x )
	506
	507	#define GET_RE_DEBUG_FLAGS DEBUG_r( \
	508	re_debug_flags=get_sv(RE_DEBUG_FLAGS, 1); \
	509	if (!SvIOK(re_debug_flags)) { \
	510	sv_setiv(re_debug_flags, RE_DEBUG_COMPILE \| RE_DEBUG_EXECUTE \| RE_DEBUG_OFFSETS); \
	511	} \
	512	)
	513
	514
	515	#ifdef DEBUGGING
e1ec3a88	516	#define GET_RE_DEBUG_FLAGS_DECL SV *re_debug_flags = NULL; GET_RE_DEBUG_FLAGS;
a3621e74	517	#else
	518	#define GET_RE_DEBUG_FLAGS_DECL
	519	#endif
	520
	521