Encode non-ASCII document IDs in multipart headers with RFC 2047 (issue 179).

kxepal · kxepal · commit dcf407310366 · 2013-04-25T09:46:55.000+02:00
diff --git a/couchdb/multipart.py b/couchdb/multipart.py
@@ -10,6 +10,7 @@
 
 from base64 import b64encode
 from cgi import parse_header
+from email import header
 try:
     from hashlib import md5
 except ImportError:
@@ -67,8 +68,13 @@ def _current_part():
         if in_headers:
             line = line.replace(CRLF, '\n')
             if line != '\n':
-                name, value = line.split(':', 1)
-                headers[name.lower().strip()] = value.strip()
+                name, value = [item.strip() for item in line.split(':', 1)]
+                name = name.lower()
+                value, charset = header.decode_header(value)[0]
+                if charset is None:
+                    headers[name] = value
+                else:
+                    headers[name] = value.decode(charset)
             else:
                 in_headers = False
                 mimetype, params = parse_header(headers.get('content-type'))
@@ -170,9 +176,12 @@ def _make_boundary(self):
     def _write_headers(self, headers):
         if headers:
             for name in sorted(headers.keys()):
+                value = headers[name]
+                if isinstance(value, unicode):
+                    value = str(header.make_header([(value, 'utf-8')]))
                 self.fileobj.write(name)
                 self.fileobj.write(': ')
-                self.fileobj.write(headers[name])
+                self.fileobj.write(value)
                 self.fileobj.write(CRLF)
         self.fileobj.write(CRLF)
 
diff --git a/couchdb/tests/multipart.py b/couchdb/tests/multipart.py
@@ -148,6 +148,23 @@ def test_nested(self):
             num += 1
         self.assertEqual(num, 3)
 
+    def test_unicode_headers(self):
+        # http://code.google.com/p/couchdb-python/issues/detail?id=179
+        dump = '''Content-Type: multipart/mixed; boundary="==123456789=="
+
+--==123456789==
+Content-ID: =?utf-8?b?5paH5qGj?=
+Content-Length: 63
+Content-MD5: Cpw3iC3xPua8YzKeWLzwvw==
+Content-Type: application/json
+
+{"_rev": "3-bc27b6930ca514527d8954c7c43e6a09", "_id": "文档"}
+'''
+        parts = multipart.read_multipart(StringIO(dump))
+        for headers, is_multipart, payload in parts:
+            self.assertEqual(headers['content-id'], u'文档')
+            break
+
 
 class WriteMultipartTestCase(unittest.TestCase):
 
@@ -173,6 +190,25 @@ def test_unicode_content_ascii(self):
         self.assertRaises(UnicodeEncodeError, envelope.add,
                           'text/plain;charset=ascii', u'Iñtërnâtiônàlizætiøn')
 
+    def test_unicode_headers(self):
+        # http://code.google.com/p/couchdb-python/issues/detail?id=179
+        buf = StringIO()
+        envelope = multipart.write_multipart(buf, boundary='==123456789==')
+        envelope.add('application/json',
+                     '{"_rev": "3-bc27b6930ca514527d8954c7c43e6a09",'
+                     ' "_id": "文档"}',
+                     headers={'Content-ID': u"文档"})
+        self.assertEqual('''Content-Type: multipart/mixed; boundary="==123456789=="
+
+--==123456789==
+Content-ID: =?utf-8?b?5paH5qGj?=
+Content-Length: 63
+Content-MD5: Cpw3iC3xPua8YzKeWLzwvw==
+Content-Type: application/json
+
+{"_rev": "3-bc27b6930ca514527d8954c7c43e6a09", "_id": "文档"}
+''', buf.getvalue().replace('\r\n', '\n'))
+
 
 def suite():
     suite = unittest.TestSuite()
