'use utf8' should imply /u regex matching

This should be true even if the pattern isn't in utf8.

diff --git a/regcomp.c b/regcomp.c
index ab5fc04..bcd159c 100644 (file)
--- a/regcomp.c
+++ b/regcomp.c
@@ -6289,7 +6289,10 @@ Perl_re_op_compile(pTHX_ SV ** const patternp, int pat_count,
 
     /* ignore the utf8ness if the pattern is 0 length */
     RExC_utf8 = RExC_orig_utf8 = (plen == 0 || IN_BYTES) ? 0 : SvUTF8(pat);
-    RExC_uni_semantics = 0;
+
+    /* 'use utf8' in the program indicates Unicode rules are wanted */
+    RExC_uni_semantics = (PL_hints & HINT_UTF8);
+
     RExC_contains_locale = 0;
     RExC_contains_i = 0;
     pRExC_state->runtime_code_qr = NULL;
@@ -6340,10 +6343,13 @@ Perl_re_op_compile(pTHX_ SV ** const patternp, int pat_count,
     if (initial_charset == REGEX_LOCALE_CHARSET) {
        RExC_contains_locale = 1;
     }
-    else if (RExC_utf8 && initial_charset == REGEX_DEPENDS_CHARSET) {
+    else if ((RExC_utf8 || RExC_uni_semantics)
+              && initial_charset == REGEX_DEPENDS_CHARSET)
+    {
 
-       /* Set to use unicode semantics if the pattern is in utf8 and has the
-        * 'depends' charset specified, as it means unicode when utf8  */
+        /* Set to use unicode semantics if has the 'depends' charset specified,
+         * and either the pattern is in utf8 (as it means unicode when utf8),
+         * or we already know we want unicode rules  */
        set_regex_charset(&rx_flags, REGEX_UNICODE_CHARSET);
     }
 

diff --git a/t/re/pat.t b/t/re/pat.t
index 91274e6..6211065 100644 (file)
--- a/t/re/pat.t
+++ b/t/re/pat.t
@@ -20,7 +20,7 @@ BEGIN {
     require './test.pl';
 }
 
-plan tests => 712;  # Update this when adding/deleting tests.
+plan tests => 713;  # Update this when adding/deleting tests.
 
 run_tests() unless caller;
 
@@ -1538,6 +1538,10 @@ EOP
         like "\x{AA}", qr/a?[\W_]/d, "\\W with /d synthetic start class works";
     }
 
+    {
+        use utf8;
+        unlike("\xe0", qr/\W/, "'use utf8' implies /u");
+    }
 
 
 } # End of sub run_tests