Move the test for 078c425b721ea35d from fresh_perl.t to pat_rt_report.t

[perl5.git] / regcomp.h

diff --git a/regcomp.h b/regcomp.h
index 48e270f..8e96b75 100644 (file)
--- a/regcomp.h
+++ b/regcomp.h
@@ -18,25 +18,6 @@ typedef OP OP_4tree;                 /* Will be redefined later. */
 /* Be really aggressive 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 
- * consistent regardless of internal string type, and make character class inversions
- * consistent. 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.
- * XXX: It is now enabled for 5.11/5.12
- *
- * -demerphq
- */
-#define PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS 1
-
 /* Should the optimiser take positive assertions into account? */
 #define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 0
 
@@ -310,8 +291,21 @@ struct regnode_charclass_class {
 
 #define SIZE_ONLY (RExC_emit == &PL_regdummy)
 
-/* Flags for node->flags of several of the node types */
-#define USE_UNI                0x01
+/* If the bitmap doesn't fully represent what this ANYOF node can match, the
+ * ARG is set to this special value (since 0, 1, ... are legal, but will never
+ * reach this high). */
+#define ANYOF_NONBITMAP_EMPTY  ((U32) -1)
+
+/* The information used to be stored as as combination of the ANYOF_UTF8 and
+ * ANYOF_NONBITMAP_NON_UTF8 bits in the flags field, but was moved out of there
+ * to free up a bit for other uses.  This tries to hide the change from
+ * existing code as much as possible.  Now, the data structure that goes in ARG
+ * is not allocated unless it is needed, and that is what is used to determine
+ * if there is something outside the bitmap.  The code now assumes that if
+ * that structure exists, that any UTF-8 encoded string should be tried against
+ * it, but a non-UTF8-encoded string will be tried only if the
+ * ANYOF_NONBITMAP_NON_UTF8 bit is also set. */
+#define ANYOF_NONBITMAP(node)  (ARG(node) != ANYOF_NONBITMAP_EMPTY)
 
 /* Flags for node->flags of ANYOF.  These are in short supply, so some games
  * are done to share them, as described below.  If necessary, the ANYOF_LOCALE
@@ -338,28 +332,28 @@ struct regnode_charclass_class {
 #define ANYOF_INVERT            0x04
 
 /* EOS, meaning that it can match an empty string too, is used for the
- * synthetic start class (ssc) only.  It can share the INVERT bit, as the ssc
- * is never inverted.  The bit just needs to be turned off before regexec.c
- * gets a hold of it so that regexec.c doesn't think it's inverted, but this
- * happens automatically, as if the ssc can match an EOS, the ssc is discarded,
- * and never passed to regexec.c */
-#define ANYOF_EOS              ANYOF_INVERT
+ * synthetic start class (ssc) only.  It looks like it could share the INVERT
+ * bit, as the ssc is never inverted.  But doing that caused this reges to
+ * not match:
+ * 'foo/file.fob' =~ m,^(?=[^\.])[^/]* /(?=[^\.])[^/]*\.fo[^/]$,;
+ * (except the space between the * and the / above shouldn't be there; it was
+ * inserted to make this comment continue on.)
+ * Rather than try to figure out what was going on in the optimizer, I (khw)
+ * found a way to save a different bit.  But my original line of reasoning was
+ * "The bit just needs to be turned off before regexec.c gets a hold of it so
+ * that regexec.c doesn't think it's inverted, but this happens automatically,
+ * as if the ssc can match an EOS, the ssc is discarded, and never passed to
+ * regexec.c" */
+#define ANYOF_EOS              0x10
 
 /* CLASS is never set unless LOCALE is too: has runtime \d, \w, [:posix:], ...
  * The non-locale ones are resolved at compile-time */
 #define ANYOF_CLASS     0x08
 #define ANYOF_LARGE      ANYOF_CLASS    /* Same; name retained for back compat */
 
-/* Can match something outside the bitmap that is expressible only in utf8 */
-#define ANYOF_UTF8             0x10
-
 /* Can match something outside the bitmap that isn't in utf8 */
 #define ANYOF_NONBITMAP_NON_UTF8 0x20
 
-/* Set if the bitmap doesn't fully represent what this node can match */
-#define ANYOF_NONBITMAP                (ANYOF_UTF8|ANYOF_NONBITMAP_NON_UTF8)
-#define ANYOF_UNICODE          ANYOF_NONBITMAP /* old name, for back compat */
-
 /* Matches every code point 0x100 and above*/
 #define ANYOF_UNICODE_ALL      0x40