Fix "[perl #20667] unicode regex vs non-unicode regex".

The PL_regdata was not set up (S_cache_re()) before it
was needed (S_reginclass()).  The test put in fresh_perl.t
because we need a pristine state to start.

p4raw-id: //depot/perl@18653

1 files changed, 12 insertions, 0 deletions

diff --git a/t/run/fresh_perl.t b/t/run/fresh_perl.t
index 9c3e55c4ef..a07396f419 100644
--- a/t/run/fresh_perl.t
+++ b/t/run/fresh_perl.t
@@ -807,3 +807,15 @@ utf8::upgrade($_); # the original code used a UTF-8 locale (affects STDIN)
 # is what matters.
 /^([[:digit:]]+)/;
 EXPECT
+######## [perl #20667] unicode regex vs non-unicode regex
+$toto = 'Hello';
+$toto =~ /\w/; # this line provokes the problem!
+$name = 'A B';
+# utf8::upgrade($name) if @ARGV;
+if ($name =~ /(\p{IsUpper}) (\p{IsUpper})/){
+    print "It's good! >$1< >$2<\n";
+} else {
+    print "It's not good...\n";
+}
+EXPECT
+It's good! >A< >B<