Output regex above-Unicode matching in syn strt class

A warning is supposed to be raised under some conditions when matching
an above-Unicode code point against a Unicode property.  Prior to this
patch, if the synthetic start class excluded the code point, the warning
would be skipped, even though it was attempted to be matched.

diff --git a/regcomp.c b/regcomp.c
index c24c7d0..46da08b 100644 (file)
--- a/regcomp.c
+++ b/regcomp.c
@@ -1107,6 +1107,25 @@ S_ssc_and(pTHX_ const RExC_state_t *pRExC_state, regnode_ssc *ssc,
     if (OP(and_with) == ANYOF_SYNTHETIC) {
         anded_cp_list = and_with->invlist;
         anded_flags = ANYOF_FLAGS(and_with);
+
+        /* XXX This is a kludge around what appears to be deficiencies in the
+         * optimizer.  If we make S_ssc_anything() add in the WARN_SUPER flag,
+         * there are paths through the optimizer where it doesn't get weeded
+         * out when it should.  And if we don't make some extra provision for
+         * it like the code just below, it doesn't get added when it should.
+         * This solution is to add it only when AND'ing, which is here, and
+         * only when what is being AND'ed is the pristine, original node
+         * matching anything.  Thus it is like adding it to ssc_anything() but
+         * only when the result is to be AND'ed.  Probably the same solution
+         * could be adopted for the same problem we have with /l matching,
+         * which is solved differently in S_ssc_init(), and that would lead to
+         * fewer false positives than that solution has.  But if this solution
+         * creates bugs, the consequences are only that a warning isn't raised
+         * that should be; while the consequences for having /l bugs is
+         * incorrect matches */
+        if (ssc_is_anything(and_with)) {
+            anded_flags |= ANYOF_WARN_SUPER;
+        }
     }
     else {
         anded_cp_list = get_ANYOF_cp_list_for_ssc(pRExC_state,

diff --git a/regcomp.h b/regcomp.h
index d87865f..9b0d7da 100644 (file)
--- a/regcomp.h
+++ b/regcomp.h
@@ -375,7 +375,7 @@ struct regnode_ssc {
 /* These are the flags that apply to both regular ANYOF nodes and synthetic
  * start class nodes during construction of the SSC.  During finalization of
  * the SSC, other of the flags could be added to it */
-#define ANYOF_COMMON_FLAGS    (ANYOF_LOCALE_FLAGS)
+#define ANYOF_COMMON_FLAGS    (ANYOF_LOCALE_FLAGS | ANYOF_WARN_SUPER)
 
 /* Character classes for node->classflags of ANYOF */
 /* Should be synchronized with a table in regprop() */

diff --git a/t/lib/warnings/utf8 b/t/lib/warnings/utf8
index aefa653..7c387b5 100644 (file)
--- a/t/lib/warnings/utf8
+++ b/t/lib/warnings/utf8
@@ -643,7 +643,7 @@ Code point 0x110048 is not Unicode, all \p{} matches fail; all \P{} matches succ
 Code point 0x110049 is not Unicode, all \p{} matches fail; all \P{} matches succeed at - line 72.
 Code point 0x110049 is not Unicode, all \p{} matches fail; all \P{} matches succeed at - line 72.
 ########
-# TODO Matching Unicode property against above-Unicode code point outputs a warning even if optimizer rejects the match (in synthetic start class)
+# NAME Matching Unicode property against above-Unicode code point outputs a warning even if optimizer rejects the match (in synthetic start class)
 use warnings 'non_unicode';
 "\x{110000}" =~ /b?\p{Space}/;
 EXPECT