APItest/t/sv_streq.t: Generalize for EBCDIC

This test fails on EBCDIC systems, because it wants a non-ASCII
character, and the one it chose, E9, is ASCII on EBCDIC ('Z').

perlhacktips suggests B6 as a character to use in such tests, and this
commit changes to use that.

1 files changed, 7 insertions, 7 deletions

diff --git a/ext/XS-APItest/t/sv_streq.t b/ext/XS-APItest/t/sv_streq.t
index dca7309f4b..e2e6063046 100644
--- a/ext/XS-APItest/t/sv_streq.t
+++ b/ext/XS-APItest/t/sv_streq.t
@@ -8,19 +8,19 @@ ok  sv_streq($abc, "abc"), '$abc eq "abc"';
 ok !sv_streq($abc, "def"), '$abc ne "def"';
 
 {
-    # U+00E9 = LATIN SMALL LETTER E WITH ACUTE
-    #  UTF-8: \xc3 \xa9
-    my $cafe_LATIN1 = "caf\xe9";
-    utf8::upgrade(my $cafe_UNICODE = "caf\x{00e9}");
-    utf8::encode (my $cafe_UTF8    = "caf\x{00e9}");
+    # U+00B6 = PARAGRAPH SEPARATOR
+    #  UTF-8 on ASCII boxes: \xc2 \xb6
+    my $psep_LATIN1 = "psep\xb6";
+    utf8::upgrade(my $psep_UNICODE = "psep\x{00b6}");
+    utf8::encode (my $psep_UTF8    = "psep\x{00b6}");
 
     # Latin-1 and Unicode strings should compare equal despite containing
     # different underlying bytes in the SvPV
-    ok sv_streq($cafe_LATIN1, $cafe_UNICODE), 'sv_streq handles UTF8 strings';
+    ok sv_streq($psep_LATIN1, $psep_UNICODE), 'sv_streq handles UTF8 strings';
 
     # UTF-8 and Unicode strings should not compare equal, even though they
     # contain the same bytes in the SvPV
-    ok !sv_streq($cafe_UTF8, $cafe_UNICODE), 'sv_streq takes UTF8ness into account';
+    ok !sv_streq($psep_UTF8, $psep_UNICODE), 'sv_streq takes UTF8ness into account';
 }
 
 # GMAGIC