X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/f722798beaa437498e1f879b01996be44f2fdc1a..7acbb8ec149b2b0481081c92fe4351d46ae0ae3a:/regcomp.h diff --git a/regcomp.h b/regcomp.h index 518add0..46a114d 100644 --- a/regcomp.h +++ b/regcomp.h @@ -1,4 +1,11 @@ /* 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. */ @@ -8,7 +15,7 @@ typedef OP OP_4tree; /* Will be redefined later. */ * compile to execute that permits the execute phase to run lots faster on * simple cases. They are: * - * regstart sv that must begin a match; Nullch if none obvious + * 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] @@ -46,7 +53,8 @@ typedef OP OP_4tree; /* Will be redefined later. */ * 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: + * to the thing following the set of BRANCHes.) The opcodes are defined + * in regnodes.h which is generated from regcomp.sym by regcomp.pl. */ /* @@ -66,10 +74,10 @@ typedef OP OP_4tree; /* Will be redefined later. */ */ struct regnode_string { - U8 flags; + U8 str_len; U8 type; U16 next_off; - U8 string[1]; + char string[1]; }; struct regnode_1 { @@ -87,6 +95,26 @@ struct regnode_2 { 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. @@ -112,6 +140,10 @@ struct regnode_2 { #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)) @@ -119,20 +151,33 @@ struct regnode_2 { #define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val)) #define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val)) -#ifndef lint -# define NEXT_OFF(p) ((p)->next_off) -# define NODE_ALIGN(node) -# define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */ -#else /* lint */ -# define NEXT_OFF(p) 0 -# define NODE_ALIGN(node) -# define NODE_ALIGN_FILL(node) -#endif /* lint */ +#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) @@ -152,87 +197,155 @@ struct regnode_2 { #define REG_MAGIC 0234 -#define SIZE_ONLY (PL_regcode == &PL_regdummy) - -/* Flags for first parameter byte of ANYOF */ -#define ANYOF_INVERT 0x40 -#define ANYOF_FOLD 0x20 -#define ANYOF_LOCALE 0x10 -#define ANYOF_ISA 0x0F -#define ANYOF_ALNUML 0x08 -#define ANYOF_NALNUML 0x04 -#define ANYOF_SPACEL 0x02 -#define ANYOF_NSPACEL 0x01 - -/* Utility macros for bitmap of ANYOF */ -#define ANYOF_BYTE(p,c) (p)[1 + (((c) >> 3) & 31)] -#define ANYOF_BIT(c) (1 << ((c) & 7)) -#define ANYOF_SET(p,c) (ANYOF_BYTE(p,c) |= ANYOF_BIT(c)) -#define ANYOF_CLEAR(p,c) (ANYOF_BYTE(p,c) &= ~ANYOF_BIT(c)) -#define ANYOF_TEST(p,c) (ANYOF_BYTE(p,c) & ANYOF_BIT(c)) - -#define ANY_SKIP ((33 - 1)/sizeof(regnode) + 1) +#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 lint #ifndef CHARMASK -#define UCHARAT(p) ((int)*(unsigned char *)(p)) +# define UCHARAT(p) ((int)*(const U8*)(p)) #else -#define UCHARAT(p) ((int)*(p)&CHARMASK) +# define UCHARAT(p) ((int)*(p)&CHARMASK) #endif -#else /* lint */ -#define UCHARAT(p) PL_regdummy -#endif /* lint */ - -#define FAIL(m) \ - STMT_START { \ - if (!SIZE_ONLY) \ - SAVEDESTRUCTOR(S_clear_re,(void*)PL_regcomp_rx); \ - Perl_croak(aTHX_ "/%.127s/: %s", PL_regprecomp,m); \ - } STMT_END - -#define FAIL2(pat,m) \ - STMT_START { \ - if (!SIZE_ONLY) \ - SAVEDESTRUCTOR(S_clear_re,(void*)PL_regcomp_rx); \ - S_re_croak2(aTHX_ "/%.127s/: ",pat,PL_regprecomp,m); \ - } STMT_END #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_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. char* since we do strchr on it. */ +/* The following have no fixed length. U8 so we can do strchr() on it. */ #ifndef DOINIT -EXTCONST char PL_varies[]; +EXTCONST U8 PL_varies[]; #else -EXTCONST char PL_varies[] = { - BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL, +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. char* since we do strchr on it. */ -/* (Note that lenght 1 means "one character" under UTF8, not "one octet".) */ +/* 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 char PL_simple[]; +EXTCONST U8 PL_simple[]; #else -EXTCONST char PL_simple[] = { - REG_ANY, ANYUTF8, SANY, SANYUTF8, ANYOF, ANYOFUTF8, - ALNUM, ALNUMUTF8, ALNUML, ALNUMLUTF8, - NALNUM, NALNUMUTF8, NALNUML, NALNUMLUTF8, - SPACE, SPACEUTF8, SPACEL, SPACELUTF8, - NSPACE, NSPACEUTF8, NSPACEL, NSPACELUTF8, - DIGIT, DIGITUTF8, NDIGIT, NDIGITUTF8, 0 +EXTCONST U8 PL_simple[] = { + REG_ANY, SANY, CANY, + ANYOF, + ALNUM, ALNUML, + NALNUM, NALNUML, + SPACE, SPACEL, + NSPACE, NSPACEL, + DIGIT, NDIGIT, + 0 }; #endif @@ -244,6 +357,19 @@ typedef struct re_scream_pos_data_s 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; @@ -253,7 +379,8 @@ struct reg_data { struct reg_substr_datum { I32 min_offset; I32 max_offset; - SV *substr; + SV *substr; /* non-utf8 variant */ + SV *utf8_substr; /* utf8 variant */ }; struct reg_substr_data { @@ -261,10 +388,139 @@ struct reg_substr_data { }; #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 */ +/* an accepting state/position*/ +struct _reg_trie_accepted { + U8 *endpos; + U16 wordnum; +}; +/* 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_accepted reg_trie_accepted; +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 + +