X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=e30e67f30f4be15b707e3edbfc6db73bfb0fe047;hb=422af00a325a1ad9ccc3d4fb7a0c04b9ab88ade6;hp=c679ca4d46a7e700eca9fc0a711422bd4e06be19;hpb=b8c5462f6edbb2dd616e1733df011beee816eee1;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regcomp.h b/regcomp.h index c679ca4..e30e67f 100644 --- a/regcomp.h +++ b/regcomp.h @@ -66,10 +66,10 @@ 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 +87,24 @@ struct regnode_2 { U16 arg2; }; +#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */ +#define ANYOF_CLASSBITMAP_SIZE 4 + +struct regnode_charclass { + U8 flags; + U8 type; + U16 next_off; + char bitmap[ANYOF_BITMAP_SIZE]; +}; + +struct regnode_charclass_class { + U8 flags; + U8 type; + U16 next_off; + char bitmap[ANYOF_BITMAP_SIZE]; + char classflags[ANYOF_CLASSBITMAP_SIZE]; +}; + /* XXX fix this description. Impose a limit of REG_INFTY on various pattern matching operations to limit stack growth and to avoid "infinite" recursions. @@ -133,6 +151,12 @@ struct regnode_2 { #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) + #define NODE_ALIGN(node) #define ARG_LOC(p) (((struct regnode_1 *)p)->arg1) #define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1) @@ -154,18 +178,23 @@ struct regnode_2 { #define SIZE_ONLY (PL_regcode == &PL_regdummy) -/* Flags for first parameter byte [0] of ANYOF */ +/* Flags for node->flags of ANYOF */ #define ANYOF_CLASS 0x08 #define ANYOF_INVERT 0x04 #define ANYOF_FOLD 0x02 #define ANYOF_LOCALE 0x01 -/* Character classes for bytes [1..4] of ANYOF */ +/* Used for regstclass only */ +#define ANYOF_EOS 0x10 /* Can match an empty string too */ + +/* 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, utf8::IsWord, isALNUM() */ #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 @@ -189,8 +218,12 @@ 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 +#define ANYOF_NBLANK 29 /* GNU extension: space and tab */ -#define ANYOF_MAX 31 +#define ANYOF_MAX 32 /* Backward source code compatibility. */ @@ -201,36 +234,38 @@ 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(p) ((p)->flags) #define ANYOF_FLAGS_ALL 0xff #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_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 @@ -238,20 +273,6 @@ 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 @@ -263,22 +284,22 @@ 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[] = { +EXTCONST U8 PL_simple[] = { REG_ANY, ANYUTF8, SANY, SANYUTF8, ANYOF, ANYOFUTF8, ALNUM, ALNUMUTF8, ALNUML, ALNUMLUTF8, NALNUM, NALNUMUTF8, NALNUML, NALNUMLUTF8,