X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.h;h=2ac1be11e995394266a2fc276249e4d1a0e8a601;hb=9469eb4a69f2671f0af6c58baf2e0a7e08975fc1;hp=645e386bdeaed280d15737657e9a978422b3b5ab;hpb=e9a8c0991ec6634f51266c0ad53e02c328bc0761;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/regcomp.h b/regcomp.h
index 645e386..2ac1be1 100644
--- a/regcomp.h
+++ b/regcomp.h
@@ -18,6 +18,24 @@ typedef OP OP_4tree;			/* Will be redefined later. */
 /* Be really agressive about optimising patterns with trie sequences? */
 #define PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION 1
 
+/* Use old style unicode mappings for perl and posix character classes
+ *
+ * NOTE: Enabling this essentially breaks character class matching against unicode 
+ * strings, so that POSIX char classes match when they shouldn't, and \d matches 
+ * way more than 10 characters, and sometimes a charclass and its complement either
+ * both match or neither match.
+ * NOTE: Disabling this will cause various backwards compatibility issues to rear 
+ * their head, and tests to fail. However it will make the charclass behaviour 
+ * consistant regardless of internal string type, and make character class inversions
+ * consistant. The tests that fail in the regex engine are basically broken tests.
+ *
+ * Personally I think 5.12 should disable this for sure. Its a bit more debatable for
+ * 5.10, so for now im leaving it enabled.
+ *
+ * -demerphq
+ */
+#define PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS 1
+
 /* Should the optimiser take positive assertions into account? */
 #define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 0
 
@@ -317,9 +335,9 @@ struct regnode_charclass_class {	/* has [[:blah:]] classes */
 #define ANYOF_NALNUM	 1
 #define ANYOF_SPACE	 2	/* \s */
 #define ANYOF_NSPACE	 3
-#define ANYOF_DIGIT	 4
+#define ANYOF_DIGIT	 4	/* \d */
 #define ANYOF_NDIGIT	 5
-#define ANYOF_ALNUMC	 6	/* isalnum(3), utf8::IsAlnum, ALNUMC */
+#define ANYOF_ALNUMC	 6	/* [[:alnum:]] isalnum(3), utf8::IsAlnum, ALNUMC */
 #define ANYOF_NALNUMC	 7
 #define ANYOF_ALPHA	 8
 #define ANYOF_NALPHA	 9
@@ -528,7 +546,10 @@ struct reg_data {
 #define check_offset_max substrs->data[2].max_offset
 #define check_end_shift substrs->data[2].end_shift
 
-
+#define RX_ANCHORED_SUBSTR(rx)	(((struct regexp *)SvANY(rx))->anchored_substr)
+#define RX_ANCHORED_UTF8(rx)	(((struct regexp *)SvANY(rx))->anchored_utf8)
+#define RX_FLOAT_SUBSTR(rx)	(((struct regexp *)SvANY(rx))->float_substr)
+#define RX_FLOAT_UTF8(rx)	(((struct regexp *)SvANY(rx))->float_utf8)
 
 /* trie related stuff */