X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/cad2e5aadfceb1a406f657488ea1c699f44a1487..389b72ffba57effeed96b76240b691bbd83408f7:/regcomp.h

diff --git a/regcomp.h b/regcomp.h
index 518add0..e30c8e7 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 {
@@ -133,6 +133,12 @@ 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)
@@ -154,31 +160,83 @@ struct regnode_2 {
 
 #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)
+/* Flags for first parameter byte [0] 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 */
+
+#define ANYOF_ALNUM	 0	/* \w, utf8::IsWord, isALNUM() */
+#define ANYOF_NALNUM	 1
+#define ANYOF_SPACE	 2
+#define ANYOF_NSPACE	 3
+#define ANYOF_DIGIT	 4
+#define ANYOF_NDIGIT	 5
+#define ANYOF_ALNUMC	 6	/* isalnum(3), utf8::IsAlnum, isALNUMC() */
+#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_MAX	31
+
+/* 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_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_FLAGS(p)		((p)[0])
+#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_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_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)
 
 /*
  * 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
@@ -189,14 +247,14 @@ struct regnode_2 {
 #define	FAIL(m) \
     STMT_START {							\
 	if (!SIZE_ONLY)							\
-	    SAVEDESTRUCTOR(S_clear_re,(void*)PL_regcomp_rx);		\
+	    SAVEDESTRUCTOR(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);		\
+	    SAVEDESTRUCTOR(clear_re,(void*)PL_regcomp_rx);		\
 	S_re_croak2(aTHX_ "/%.127s/: ",pat,PL_regprecomp,m);		\
     } STMT_END
 
@@ -211,22 +269,22 @@ 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,