X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/b8c5462f6edbb2dd616e1733df011beee816eee1..090f71652b598badef32b379c0f36c2c6e4a9542:/regcomp.h diff --git a/regcomp.h b/regcomp.h index c679ca4..3aa5c1e 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, 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. */ @@ -66,10 +73,10 @@ 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 +94,26 @@ 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 +139,10 @@ 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)) @@ -119,6 +150,9 @@ 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)) +#undef NEXT_OFF +#undef NODE_ALIGN + #ifndef lint # define NEXT_OFF(p) ((p)->next_off) # define NODE_ALIGN(node) @@ -131,8 +165,24 @@ struct regnode_2 { #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) @@ -152,24 +202,41 @@ struct regnode_2 { #define REG_MAGIC 0234 -#define SIZE_ONLY (PL_regcode == &PL_regdummy) +#define SIZE_ONLY (RExC_emit == &PL_regdummy) + +/* Flags for node->flags of ANYOF */ -/* Flags for first parameter byte [0] of ANYOF */ +#define ANYOF_CLASS 0x08 /* has [[:blah:]] classes */ +#define ANYOF_INVERT 0x04 +#define ANYOF_FOLD 0x02 +#define ANYOF_LOCALE 0x01 -#define ANYOF_CLASS 0x08 -#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 */ -/* Character classes for bytes [1..4] of ANYOF */ +/* 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 -#define ANYOF_ALNUM 0 /* \w, utf8::IsWord, isALNUM() */ +/* 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 +#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, isALNUMC() */ +#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */ #define ANYOF_NALNUMC 7 #define ANYOF_ALPHA 8 #define ANYOF_NALPHA 9 @@ -189,8 +256,12 @@ struct regnode_2 { #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 31 +#define ANYOF_MAX 32 /* Backward source code compatibility. */ @@ -201,36 +272,45 @@ struct regnode_2 { /* Utility macros for the bitmap and classes of ANYOF */ -#define ANYOF_OPND_SIZE 1 -#define ANYOF_CLASS_SIZE 4 -#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */ -#define ANYOF_SIZE (ANYOF_OPND_SIZE+ANYOF_CLASS_SIZE+ANYOF_BITMAP_SIZE) +#define ANYOF_SIZE (sizeof(struct regnode_charclass)) +#define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class)) -#define ANYOF_FLAGS(p) ((p)[0]) -#define ANYOF_FLAGS_ALL 0xff +#define ANYOF_FLAGS(p) ((p)->flags) #define ANYOF_BIT(c) (1 << ((c) & 7)) -#define ANYOF_CLASS_OFFSET ANYOF_OPND_SIZE -#define ANYOF_CLASS_BYTE(p, c) ((p)[ANYOF_CLASS_OFFSET + (((c) >> 3) & 3)]) +#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_BITMAP_OFFSET (ANYOF_CLASS_OFFSET+ANYOF_CLASS_SIZE) -#define ANYOF_BITMAP_BYTE(p, c) ((p)[ANYOF_BITMAP_OFFSET + (((c) >> 3) & 31)]) +#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 ANY_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode) + 1) +#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)*(U8*)(p)) #else #define UCHARAT(p) ((int)*(p)&CHARMASK) #endif @@ -238,53 +318,43 @@ struct regnode_2 { #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[] = { +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 @@ -296,6 +366,18 @@ 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 + * 20010712 mjd@plover.com + * (Remember to update re_dup() and pregfree() if you add any items.) + */ struct reg_data { U32 count; U8 *what; @@ -305,7 +387,8 @@ 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 { @@ -313,10 +396,13 @@ 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