X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=066e31f01d2f51d4693a87709b1e4e48cf5b3f92;hb=ed6b3797850720f78128f390c7ef98bfdebafac0;hp=d6ee2f1be4844e33904012d98da4127e623bdbb3;hpb=c76ac1ee16c5c9591d8ffad5a0651b7f81738b68;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regcomp.h b/regcomp.h index d6ee2f1..066e31f 100644 --- a/regcomp.h +++ b/regcomp.h @@ -87,6 +87,26 @@ 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]; +}; + +struct regnode_charclass_class { + U8 flags; + U8 type; + U16 next_off; + U32 arg1; + 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. @@ -158,24 +178,38 @@ 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 */ + +#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 */ + +/* There is a character or a range past 0xff */ +#define ANYOF_UNICODE 0x20 +#define ANYOF_UNICODE_ALL 0x40 /* Can match any char past 0xff */ -/* Flags for first parameter byte [0] of ANYOF */ +/* Are there any runtime flags on in this node? */ +#define ANYOF_RUNTIME(s) (ANYOF_FLAGS(s) & 0x0f) -#define ANYOF_CLASS 0x08 -#define ANYOF_INVERT 0x04 -#define ANYOF_FOLD 0x02 -#define ANYOF_LOCALE 0x01 +#define ANYOF_FLAGS_ALL 0xff -/* Character classes for bytes [1..4] of ANYOF */ +/* 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_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 @@ -195,8 +229,12 @@ 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. */ @@ -207,29 +245,30 @@ 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_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. @@ -244,20 +283,6 @@ struct regnode_2 { #define UCHARAT(p) PL_regdummy #endif /* lint */ -#define FAIL(m) \ - STMT_START { \ - if (!SIZE_ONLY) \ - SAVEDESTRUCTOR_X(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_X(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 @@ -285,12 +310,14 @@ EXTCONST U8 PL_varies[] = { EXTCONST U8 PL_simple[]; #else EXTCONST U8 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 + REG_ANY, SANY, + ANYOF, + ALNUM, ALNUML, + NALNUM, NALNUML, + SPACE, SPACEL, + NSPACE, NSPACEL, + DIGIT, NDIGIT, + 0 }; #endif