/* regcomp.h
*
* Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
- * 2000, 2001, 2002, 2003, 2005 by Larry Wall and others
+ * 2000, 2001, 2002, 2003, 2005, 2006, 2007, 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.
*
*/
+#include "regcharclass.h"
typedef OP OP_4tree; /* Will be redefined later. */
/* Convert branch sequences to more efficient trie ops? */
#define PERL_ENABLE_TRIE_OPTIMISATION 1
-/* Be really agressive about optimising patterns with trie sequences? */
+/* Be really aggressive about optimising patterns with trie sequences? */
#define PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION 1
/* Should the optimiser take positive assertions into account? */
-#define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 1
+#define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 0
/* Not for production use: */
#define PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS 0
+/* Activate offsets code - set to if 1 to enable */
+#ifdef DEBUGGING
+#define RE_TRACK_PATTERN_OFFSETS
+#endif
+
/* Unless the next line is uncommented it is illegal to combine lazy
matching with possessive matching. Frankly it doesn't make much sense
to allow it as X*?+ matches nothing, X+?+ matches a single char only,
* stored negative.]
*/
+/* This is the stuff that used to live in regexp.h that was truly
+ private to the engine itself. It now lives here. */
+
+
+
+ typedef struct regexp_internal {
+ int name_list_idx; /* Optional data index of an array of paren names */
+ union {
+ U32 *offsets; /* offset annotations 20001228 MJD
+ data about mapping the program to the
+ string -
+ offsets[0] is proglen when this is used
+ */
+ U32 proglen;
+ } u;
+
+ 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_internal;
+
+#define RXi_SET(x,y) (x)->pprivate = (void*)(y)
+#define RXi_GET(x) ((regexp_internal *)((x)->pprivate))
+#define RXi_GET_DECL(r,ri) regexp_internal *ri = RXi_GET(r)
+/*
+ * Flags stored in regexp->intflags
+ * These are used only internally to the regexp engine
+ *
+ * See regexp.h for flags used externally 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
+
+
+/* this is where the old regcomp.h started */
+
struct regnode_string {
U8 str_len;
U8 type;
U8 flags;
U8 type;
U16 next_off;
- U32 arg1;
+ U32 arg1; /* used as ptr in S_regclass */
char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */
};
-struct regnode_charclass_class { /* has [[:blah:]] classes */
- U8 flags; /* should have ANYOF_CLASS here */
+/* has runtime (locale) \d, \w, ..., [:posix:] classes */
+struct regnode_charclass_class {
+ U8 flags; /* ANYOF_CLASS bit must go here */
U8 type;
U16 next_off;
- U32 arg1;
+ U32 arg1; /* used as ptr in S_regclass */
char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time */
char classflags[ANYOF_CLASSBITMAP_SIZE]; /* and run-time */
};
#undef STRING
#define OP(p) ((p)->type)
+#define FLAGS(p) ((p)->flags) /* Caution: Doesn't apply to all \
+ regnode types. For some, it's the \
+ character set of the regnode */
#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 SIZE_ONLY (RExC_emit == &PL_regdummy)
-/* Flags for node->flags of ANYOF */
+/* If the bitmap doesn't fully represent what this ANYOF node can match, the
+ * ARG is set to this special value (since 0, 1, ... are legal, but will never
+ * reach this high). */
+#define ANYOF_NONBITMAP_EMPTY ((U32) -1)
+
+/* The information used to be stored as as combination of the ANYOF_UTF8 and
+ * ANYOF_NONBITMAP_NON_UTF8 bits in the flags field, but was moved out of there
+ * to free up a bit for other uses. This tries to hide the change from
+ * existing code as much as possible. Now, the data structure that goes in ARG
+ * is not allocated unless it is needed, and that is what is used to determine
+ * if there is something outside the bitmap. The code now assumes that if
+ * that structure exists, that any UTF-8 encoded string should be tried against
+ * it, but a non-UTF8-encoded string will be tried only if the
+ * ANYOF_NONBITMAP_NON_UTF8 bit is also set. */
+#define ANYOF_NONBITMAP(node) (ARG(node) != ANYOF_NONBITMAP_EMPTY)
+
+/* Flags for node->flags of ANYOF. These are in short supply, so some games
+ * are done to share them, as described below. If necessary, the ANYOF_LOCALE
+ * and ANYOF_CLASS bits could be shared with a space penalty for locale nodes,
+ * but this isn't quite so easy, as the optimizer also uses ANYOF_CLASS.
+ * Once the planned change to compile all the above-latin1 code points is done,
+ * then the UNICODE_ALL bit can be freed up, with a small performance penalty.
+ * If flags need to be added that are applicable to the synthetic start class
+ * only, with some work, they could be put in the next-node field, or in an
+ * unused bit of the classflags field. */
+
+#define ANYOF_LOCALE 0x01 /* /l modifier */
+
+/* The fold is calculated and stored in the bitmap where possible at compile
+ * time. However there are two cases where it isn't possible. These share
+ * this bit: 1) under locale, where the actual folding varies depending on
+ * what the locale is at the time of execution; and 2) where the folding is
+ * specified in a swash, not the bitmap, such as characters which aren't
+ * specified in the bitmap, or properties that aren't looked at at compile time
+ */
+#define ANYOF_LOC_NONBITMAP_FOLD 0x02
+
+#define ANYOF_INVERT 0x04
+
+/* Set if this is a struct regnode_charclass_class vs a regnode_charclass. This
+ * is used for runtime \d, \w, [:posix:], ..., which are used only in locale
+ * and the optimizer's synthetic start class. Non-locale \d, etc are resolved
+ * at compile-time */
+#define ANYOF_CLASS 0x08
+#define ANYOF_LARGE ANYOF_CLASS /* Same; name retained for back compat */
-#define ANYOF_CLASS 0x08 /* has [[:blah:]] classes */
-#define ANYOF_INVERT 0x04
-#define ANYOF_FOLD 0x02
-#define ANYOF_LOCALE 0x01
+/* EOS, meaning that it can match an empty string too, is used for the
+ * synthetic start class only. */
+#define ANYOF_EOS 0x10
-/* Used for regstclass only */
-#define ANYOF_EOS 0x10 /* Can match an empty string too */
+/* ? Is this node the synthetic start class (ssc). This bit is shared with
+ * ANYOF_EOS, as the latter is used only for the ssc, and then not used by
+ * regexec.c. And, the code is structured so that if it is set, the ssc is
+ * not used, so it is guaranteed to be 0 for the ssc by the time regexec.c
+ * gets executed, and 0 for a non-ssc ANYOF node, as it only ever gets set for
+ * a potential ssc candidate. Thus setting it to 1 after it has been
+ * determined that the ssc will be used is not ambiguous */
+#define ANYOF_IS_SYNTHETIC ANYOF_EOS
-/* There is a character or a range past 0xff */
-#define ANYOF_UNICODE 0x20
-#define ANYOF_UNICODE_ALL 0x40 /* Can match any char past 0xff */
+/* Can match something outside the bitmap that isn't in utf8 */
+#define ANYOF_NONBITMAP_NON_UTF8 0x20
-/* size of node is large (includes class pointer) */
-#define ANYOF_LARGE 0x80
+/* Matches every code point 0x100 and above*/
+#define ANYOF_UNICODE_ALL 0x40
-/* Are there any runtime flags on in this node? */
-#define ANYOF_RUNTIME(s) (ANYOF_FLAGS(s) & 0x0f)
+/* Match all Latin1 characters that aren't ASCII when the target string is not
+ * in utf8. */
+#define ANYOF_NON_UTF8_LATIN1_ALL 0x80
#define ANYOF_FLAGS_ALL 0xff
+/* These are the flags that ANYOF_INVERT being set or not doesn't affect
+ * whether they are operative or not. e.g., the node still has LOCALE
+ * regardless of being inverted; whereas ANYOF_UNICODE_ALL means something
+ * different if inverted */
+#define INVERSION_UNAFFECTED_FLAGS (ANYOF_LOCALE \
+ |ANYOF_LOC_NONBITMAP_FOLD \
+ |ANYOF_CLASS \
+ |ANYOF_EOS \
+ |ANYOF_NONBITMAP_NON_UTF8)
+
/* Character classes for node->classflags of ANYOF */
/* Should be synchronized with a table in regprop() */
/* 2n should pair with 2n+1 */
#define ANYOF_NALNUM 1
#define ANYOF_SPACE 2 /* \s */
#define ANYOF_NSPACE 3
-#define ANYOF_DIGIT 4
+#define ANYOF_DIGIT 4 /* \d */
#define ANYOF_NDIGIT 5
-#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */
+#define ANYOF_ALNUMC 6 /* [[:alnum:]] isalnum(3), utf8::IsAlnum, ALNUMC */
#define ANYOF_NALNUMC 7
#define ANYOF_ALPHA 8
#define ANYOF_NALPHA 9
#define ANYOF_MAX 32
+/* pseudo classes, not stored in the class bitmap, but used as flags
+ during compilation of char classes */
+
+#define ANYOF_VERTWS (ANYOF_MAX+1)
+#define ANYOF_NVERTWS (ANYOF_MAX+2)
+#define ANYOF_HORIZWS (ANYOF_MAX+3)
+#define ANYOF_NHORIZWS (ANYOF_MAX+4)
+
/* Backward source code compatibility. */
#define ANYOF_ALNUML ANYOF_ALNUM
#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_CLASS_SETALL(ret) \
+ memset (((struct regnode_charclass_class*)(ret))->classflags, 255, ANYOF_CLASSBITMAP_SIZE)
#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_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) \
+#define ANYOF_BITMAP_TESTALLSET(p) /* Assumes sizeof(p) == 32 */ \
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)
+
+#if ANYOF_CLASSBITMAP_SIZE != 4
+# error ANYOF_CLASSBITMAP_SIZE is expected to be 4
+#endif
+#define ANYOF_CLASS_TEST_ANY_SET(p) ((ANYOF_FLAGS(p) & ANYOF_CLASS) \
+ && memNE (((struct regnode_charclass_class*)(p))->classflags, \
+ "\0\0\0\0", ANYOF_CLASSBITMAP_SIZE))
+/*#define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP)
+ * */
/*
#define REG_TOP_LEVEL_BRANCHES 0x00000040
#define REG_SEEN_VERBARG 0x00000080
#define REG_SEEN_CUTGROUP 0x00000100
+#define REG_SEEN_RUN_ON_COMMENT 0x00000200
START_EXTERN_C
#include "regnodes.h"
#endif
-/* 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
-
#ifndef PLUGGABLE_RE_EXTENSION
#ifndef DOINIT
EXTCONST regexp_engine PL_core_reg_engine;
#else /* DOINIT */
EXTCONST regexp_engine PL_core_reg_engine = {
- Perl_pregcomp,
- Perl_regexec_flags,
+ Perl_re_compile,
+ Perl_regexec_flags,
Perl_re_intuit_start,
Perl_re_intuit_string,
- Perl_pregfree,
- Perl_reg_stringify,
+ Perl_regfree_internal,
+ Perl_reg_numbered_buff_fetch,
+ Perl_reg_numbered_buff_store,
+ Perl_reg_numbered_buff_length,
+ Perl_reg_named_buff,
+ Perl_reg_named_buff_iter,
+ Perl_reg_qr_package,
#if defined(USE_ITHREADS)
- Perl_regdupe
+ Perl_regdupe_internal
#endif
};
#endif /* DOINIT */
/* .what is a character array with one character for each member of .data
* The character describes the function of the corresponding .data item:
+ * a - AV for paren_name_list under DEBUGGING
* 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
+ * 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
void* data[1];
};
-struct reg_substr_datum {
- I32 min_offset;
- I32 max_offset;
- SV *substr; /* non-utf8 variant */
- SV *utf8_substr; /* utf8 variant */
- I32 end_shift;
-};
-
-struct reg_substr_data {
- struct reg_substr_datum data[3]; /* Actual array */
-};
-
+/* Code in S_to_utf8_substr() and S_to_byte_substr() in regexec.c accesses
+ anchored* and float* via array indexes 0 and 1. */
#define anchored_substr substrs->data[0].substr
#define anchored_utf8 substrs->data[0].utf8_substr
#define anchored_offset substrs->data[0].min_offset
#define check_offset_max substrs->data[2].max_offset
#define check_end_shift substrs->data[2].end_shift
-
+#define RX_ANCHORED_SUBSTR(rx) (((struct regexp *)SvANY(rx))->anchored_substr)
+#define RX_ANCHORED_UTF8(rx) (((struct regexp *)SvANY(rx))->anchored_utf8)
+#define RX_FLOAT_SUBSTR(rx) (((struct regexp *)SvANY(rx))->float_substr)
+#define RX_FLOAT_UTF8(rx) (((struct regexp *)SvANY(rx))->float_utf8)
/* trie related stuff */
} trans;
};
+/* info per word; indexed by wordnum */
+typedef struct {
+ U16 prev; /* previous word in acceptance chain; eg in
+ * zzz|abc|ab/ after matching the chars abc, the
+ * accepted word is #2, and the previous accepted
+ * word is #3 */
+ U32 len; /* how many chars long is this word? */
+ U32 accept; /* accept state for this word */
+} reg_trie_wordinfo;
typedef struct _reg_trie_state reg_trie_state;
optimisation in Perl_regdupe. */
struct _reg_trie_data {
U32 refcount; /* number of times this trie is referenced */
- U16 uniquecharcount; /* unique chars in trie (width of trans table) */
U32 lasttrans; /* last valid transition element */
U16 *charmap; /* byte to charid lookup array */
reg_trie_state *states; /* state data */
reg_trie_trans *trans; /* array of transition elements */
char *bitmap; /* stclass bitmap */
+ U16 *jump; /* optional 1 indexed array of offsets before tail
+ for the node following a given word. */
+ reg_trie_wordinfo *wordinfo; /* array of info per word */
+ U16 uniquecharcount; /* unique chars in trie (width of trans table) */
U32 startstate; /* initial state - used for common prefix optimisation */
STRLEN minlen; /* minimum length of words in trie - build/opt only? */
STRLEN maxlen; /* maximum length of words in trie - build/opt only? */
- U32 *wordlen; /* array of lengths of words */
- U16 *jump; /* optional 1 indexed array of offsets before tail
- for the node following a given word. */
- U16 *nextword; /* optional 1 indexed array to support linked list
- of duplicate wordnums */
+ U32 prefixlen; /* #chars in common prefix */
U32 statecount; /* Build only - number of states in the states array
(including the unused zero state) */
U32 wordcount; /* Build only */
optimisation in Perl_regdupe. */
struct _reg_ac_data {
U32 refcount;
+ U32 trie;
U32 *fail;
reg_trie_state *states;
- U32 trie;
};
typedef struct _reg_ac_data reg_ac_data;
-/* ANY_BIT doesnt use the structure, so we can borrow it here.
+/* ANY_BIT doesn't use the structure, so we can borrow it here.
This is simpler than refactoring all of it as wed end up with
three different sets... */
#define RE_DEBUG_COMPILE_OPTIMISE 0x000002
#define RE_DEBUG_COMPILE_TRIE 0x000004
#define RE_DEBUG_COMPILE_DUMP 0x000008
+#define RE_DEBUG_COMPILE_FLAGS 0x000010
/* Execute */
#define RE_DEBUG_EXECUTE_MASK 0x00FF00
#define RE_DEBUG_EXTRA_OFFDEBUG 0x040000
#define RE_DEBUG_EXTRA_STATE 0x080000
#define RE_DEBUG_EXTRA_OPTIMISE 0x100000
+#define RE_DEBUG_EXTRA_BUFFERS 0x400000
+#define RE_DEBUG_EXTRA_GPOS 0x800000
/* combined */
#define RE_DEBUG_EXTRA_STACK 0x280000
if (re_debug_flags & RE_DEBUG_COMPILE_DUMP) x )
#define DEBUG_TRIE_COMPILE_r(x) DEBUG_r( \
if (re_debug_flags & RE_DEBUG_COMPILE_TRIE) x )
-
+#define DEBUG_FLAGS_r(x) DEBUG_r( \
+ if (re_debug_flags & RE_DEBUG_COMPILE_FLAGS) x )
/* Execute */
#define DEBUG_EXECUTE_r(x) DEBUG_r( \
if (re_debug_flags & RE_DEBUG_EXECUTE_MASK) x )
if (re_debug_flags & RE_DEBUG_EXTRA_STATE) x )
#define DEBUG_STACK_r(x) DEBUG_r( \
if (re_debug_flags & RE_DEBUG_EXTRA_STACK) x )
+#define DEBUG_BUFFERS_r(x) DEBUG_r( \
+ if (re_debug_flags & RE_DEBUG_EXTRA_BUFFERS) x )
+
#define DEBUG_OPTIMISE_MORE_r(x) DEBUG_r( \
if ((RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE) == \
(re_debug_flags & (RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE)) ) x )
#define DEBUG_TRIE_r(x) DEBUG_r( \
if (re_debug_flags & (RE_DEBUG_COMPILE_TRIE \
| RE_DEBUG_EXECUTE_TRIE )) x )
+#define DEBUG_GPOS_r(x) DEBUG_r( \
+ if (re_debug_flags & RE_DEBUG_EXTRA_GPOS) x )
/* initialization */
-/* get_sv() can return NULL during global destruction. */
+/* get_sv() can return NULL during global destruction. re_debug_flags can get
+ * clobbered by a longjmp, so must be initialized */
#define GET_RE_DEBUG_FLAGS DEBUG_r({ \
SV * re_debug_flags_sv = NULL; \
+ re_debug_flags = 0; \
re_debug_flags_sv = get_sv(RE_DEBUG_FLAGS, 1); \
if (re_debug_flags_sv) { \
if (!SvIOK(re_debug_flags_sv)) \
#ifdef DEBUGGING
-#define GET_RE_DEBUG_FLAGS_DECL IV re_debug_flags = 0; GET_RE_DEBUG_FLAGS;
+#define GET_RE_DEBUG_FLAGS_DECL VOL IV re_debug_flags \
+ __attribute__unused__ = 0; GET_RE_DEBUG_FLAGS;
#define RE_PV_COLOR_DECL(rpv,rlen,isuni,dsv,pv,l,m,c1,c2) \
const char * const rpv = \
pv_pretty((dsv), (pv), (l), (m), \
PL_colors[(c1)],PL_colors[(c2)], \
- ((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \
+ PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \
const int rlen = SvCUR(dsv)
#define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m) \
const char * const rpv = \
pv_pretty((dsv), (SvPV_nolen_const(sv)), (SvCUR(sv)), (m), \
PL_colors[(c1)],PL_colors[(c2)], \
- ((isuni) ? PERL_PV_ESCAPE_UNI : 0) )
+ PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) )
#define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m) \
const char * const rpv = \
pv_pretty((dsv), (pv), (l), (m), \
PL_colors[0], PL_colors[1], \
- ( PERL_PV_PRETTY_QUOTE | PERL_PV_PRETTY_ELIPSES | \
+ ( PERL_PV_PRETTY_QUOTE | PERL_PV_ESCAPE_RE | PERL_PV_ESCAPE_NONASCII | PERL_PV_PRETTY_ELLIPSES | \
((isuni) ? PERL_PV_ESCAPE_UNI : 0)) \
- )
+ )
#define RE_SV_DUMPLEN(ItEm) (SvCUR(ItEm) - (SvTAIL(ItEm)!=0))
#define RE_SV_TAIL(ItEm) (SvTAIL(ItEm) ? "$" : "")
#endif /* DEBUG RELATED DEFINES */
+/*
+ * Local variables:
+ * c-indentation-style: bsd
+ * c-basic-offset: 4
+ * indent-tabs-mode: t
+ * End:
+ *
+ * ex: set ts=8 sts=4 sw=4 noet:
+ */