Test by Martin v. Loewis for the new UTF-16 codec handling of BOM

marks.

1 files changed, 25 insertions, 0 deletions

diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
new file mode 100644
index 0000000..6d9e31d
--- /dev/null
+++ b/Lib/test/test_codecs.py
@@ -0,0 +1,25 @@
+import test_support,unittest
+import codecs
+import StringIO
+
+class UTF16Test(unittest.TestCase):
+
+    spamle = '\xff\xfes\x00p\x00a\x00m\x00s\x00p\x00a\x00m\x00'
+    spambe = '\xfe\xff\x00s\x00p\x00a\x00m\x00s\x00p\x00a\x00m'
+
+    def test_only_one_bom(self):
+        _,_,reader,writer = codecs.lookup("utf-16")
+        # encode some stream
+        s = StringIO.StringIO()
+        f = writer(s)
+        f.write(u"spam")
+        f.write(u"spam")
+        d = s.getvalue()
+        # check whether there is exactly one BOM in it
+        self.assert_(d == self.spamle or d == self.spambe)
+        # try to read it back
+        s = StringIO.StringIO(d)
+        f = reader(s)
+        self.assertEquals(f.read(), u"spamspam")
+
+test_support.run_unittest(UTF16Test)