From 67f8f2fe2a93081aaac8ddc1409df4b05daf4fab Mon Sep 17 00:00:00 2001
From: Barry Warsaw <barry@python.org>
Date: Mon, 14 Oct 2002 16:52:41 +0000
Subject: [PATCH] append(): Fixing the test for convertability after
 consultation with Ben.  If s is a byte string, make sure it can be converted
 to unicode with the input codec, and from unicode with the output codec, or
 raise a UnicodeError exception early.  Skip this test (and the unicode->byte
 string conversion) when the charset is our faux 8bit raw charset.

---
 Lib/email/Header.py | 42 ++++++++++++++++++++++++++++--------------
 1 file changed, 28 insertions(+), 14 deletions(-)

diff --git a/Lib/email/Header.py b/Lib/email/Header.py
index 9bbc32fd84c..0ceacc7bf30 100644
--- a/Lib/email/Header.py
+++ b/Lib/email/Header.py
@@ -218,20 +218,34 @@ def append(self, s, charset=None):
             charset = self._charset
         elif not isinstance(charset, Charset):
             charset = Charset(charset)
-        # Normalize and check the string
-        if isinstance(s, StringType):
-            # Possibly raise UnicodeError if it can't be encoded
-            unicode(s, charset.get_output_charset())
-        elif isinstance(s, UnicodeType):
-            # Convert Unicode to byte string for later concatenation
-            for charset in USASCII, charset, UTF8:
-                try:
-                    s = s.encode(charset.get_output_charset())
-                    break
-                except UnicodeError:
-                    pass
-            else:
-                assert False, 'Could not encode to utf-8'
+        # If the charset is our faux 8bit charset, leave the string unchanged
+        if charset <> '8bit':
+            # We need to test that the string can be converted to unicode and
+            # back to a byte string, given the input and output codecs of the
+            # charset.
+            if isinstance(s, StringType):
+                # Possibly raise UnicodeError if the byte string can't be
+                # converted to a unicode with the input codec of the charset.
+                incodec = charset.input_codec or 'us-ascii'
+                ustr = unicode(s, incodec)
+                # Now make sure that the unicode could be converted back to a
+                # byte string with the output codec, which may be different
+                # than the iput coded.  Still, use the original byte string.
+                outcodec = charset.output_codec or 'us-ascii'
+                ustr.encode(outcodec)
+            elif isinstance(s, UnicodeType):
+                # Now we have to be sure the unicode string can be converted
+                # to a byte string with a reasonable output codec.  We want to
+                # use the byte string in the chunk.
+                for charset in USASCII, charset, UTF8:
+                    try:
+                        outcodec = charset.output_codec or 'us-ascii'
+                        s = s.encode(outcodec)
+                        break
+                    except UnicodeError:
+                        pass
+                else:
+                    assert False, 'utf-8 conversion failed'
         self._chunks.append((s, charset))
 
     def _split(self, s, charset, firstline=False):