Add test that to/from native character set works

For non-ASCII systems, there are character set translation tables.  This
makes sure the two accessible ones are inverses of each other.  If not,
nothing can be expected to work right.

2 files changed, 28 insertions, 0 deletions

diff --git a/MANIFEST b/MANIFEST
index d7da8c45ab..3291137c1f 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -4917,6 +4917,7 @@ t/base/num.t			See if numbers work
 t/base/pat.t			See if pattern matching works
 t/base/rs.t			See if record-read works
 t/base/term.t			See if various terms work
+t/base/translate.t		See if character set translation works
 t/base/while.t			See if while work
 t/benchmark/rt26188-speed-up-keys-on-empty-hash.t	Benchmark if keys on empty hashes is fast enough
 t/bigmem/index.t		Check that index() handles large offsets
diff --git a/t/base/translate.t b/t/base/translate.t
new file mode 100644
index 0000000000..614f22c9c6
--- /dev/null
+++ b/t/base/translate.t
@@ -0,0 +1,27 @@
+#!./perl
+
+# Verify round trip of translations from the native character set to unicode
+# and back work.  If this is wrong, nothing will be reliable.
+
+print "1..257\n";   # 0-255 plus one beyond
+
+for my $i (0 .. 255) {
+    my $uni = utf8::native_to_unicode($i);
+    if ($uni < 0 || $uni >= 256) {
+        print "not ";
+    }
+    elsif (utf8::unicode_to_native(utf8::native_to_unicode($i)) != $i) {
+        print "not ";
+    }
+    print "ok ";
+    print $i + 1;
+    print "\n";
+}
+
+# Choose a largish number that might cause a seg fault if inappropriate array
+# lookup
+if (utf8::unicode_to_native(utf8::native_to_unicode(100000)) != 100000) {
+    print "not ";
+}
+print "ok ";
+print "257\n";