X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=11c09f05819703c23340eb0310a8275ec9b42778;hb=4c79ee7a1e7564ef83d0ac25d6677fdebb3ec7aa;hp=b292ba36d547cd4c44bf966f816a4f2dd4c8cfcb;hpb=19103b91374ccd8680065ad4812871fe2d75144e;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/regcomp.h b/regcomp.h
index b292ba3..11c09f0 100644
--- a/regcomp.h
+++ b/regcomp.h
@@ -1,4 +1,10 @@
 /*    regcomp.h
+ *
+ *    Copyright (c) 1997-2003, Larry Wall
+ *
+ *    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. */
@@ -132,6 +138,10 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #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))
@@ -139,6 +149,9 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #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)
@@ -151,6 +164,11 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 
 #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))
@@ -159,6 +177,11 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #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)
@@ -269,6 +292,14 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #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)
@@ -292,7 +323,8 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #define REG_SEEN_LOOKBEHIND	 2
 #define REG_SEEN_GPOS		 4
 #define REG_SEEN_EVAL		 8
-#define REG_SEEN_SANY		16
+#define REG_SEEN_CANY		16
+#define REG_SEEN_SANY		REG_SEEN_CANY /* src bckwrd cmpt */
 
 START_EXTERN_C
 
@@ -314,7 +346,7 @@ EXTCONST U8 PL_varies[] = {
 EXTCONST U8 PL_simple[];
 #else
 EXTCONST U8 PL_simple[] = {
-    REG_ANY,	SANY,
+    REG_ANY,	SANY,	CANY,
     ANYOF,
     ALNUM,	ALNUML,
     NALNUM,	NALNUML,
@@ -333,6 +365,18 @@ 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;
@@ -342,7 +386,8 @@ 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 {
@@ -350,10 +395,13 @@ 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