X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=3aa5c1e18bb66bbc533b8b477b8f1d421f64f76c;hb=f8f79f57f467ffff4d31dc518ce3f6d2364090a0;hp=b9542c9bce811e8abac472e84401375afe1f9b06;hpb=e8e3b0fb0cb9321810604b285f562851680b538c;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/regcomp.h b/regcomp.h
index b9542c9..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. */
@@ -286,6 +293,14 @@ 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)
@@ -309,7 +324,8 @@ 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
 
@@ -331,7 +347,7 @@ 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,
@@ -350,6 +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;
@@ -359,7 +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 {
@@ -367,10 +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