*** email-2.4.3/email/Header.py.orig	Tue Oct 15 01:50:57 2002
--- email-2.4.3/email/Header.py	Wed Dec 25 15:37:55 2002
***************
*** 4,9 ****
--- 4,10 ----
  """Header encoding and decoding functionality."""
  
  import re
+ import string
  from types import StringType, UnicodeType
  
  import email.quopriMIME
*************** ecre = re.compile(r'''
*** 47,52 ****
--- 48,56 ----
    \?=                   # literal ?=
    ''', re.VERBOSE | re.IGNORECASE)
  
+ nonasciiword = '[^\t ]*[^\t !-~][^\t ]*'
+ nonasciiwords = re.compile('^('+nonasciiword+'(?:[\t ]'+nonasciiword+')*)')
+ asciiwords = re.compile('^([\t !-~]*[\t ]|[\t !-~]+$)')
  
  
  # Helpers
*************** class Header:
*** 399,404 ****
--- 403,452 ----
          joiner = NL + self._continuation_ws
          return joiner.join(chunks)
  
+     def _split_nonascii(self, s, charset):
+         if charset == '8bit':
+             return [(s, charset)]
+         elif charset == 'us-ascii':
+             return [(s, charset)]
+         elif not isinstance(charset, Charset):
+             charset = Charset(charset)
+         splittable = charset.to_splittable(s)
+         chunks = []
+         while len(splittable):
+             # 'ASCII words' need not be encoded.  Its charset may be us-ascii.
+             match = asciiwords.search(splittable)
+             if match:
+                 splittable = splittable[match.end(1):]
+                 words = match.group(1)
+                 #if len(chunks):
+                 #    # prev. is 'non-ASCII words'.  SPACE is preserved.
+                 #    words = string.lstrip(words)
+                 if len(splittable):
+                     # next is 'non-ASCII words'.  SPACE is preserved.
+                     words = string.rstrip(words)
+                 encoded = charset.from_splittable(words, False)
+                 chunks.append((encoded, USASCII))
+                 continue
+             # 'non-ASCII words' must be encoded, where 'non-ASCII words'
+             # contain:
+             # - non-ASCII string.
+             # - non-SPACE ASCII string(s) concatinated with non-ASCII string,
+             #   as ASCII word and 'encoded-word' must be separated.
+             # - SPACE between 'non-ASCII words', as SPACE between 
+             #   'encoded-word's is not displayed.
+             match = nonasciiwords.search(splittable)
+             if match:
+                 splittable = splittable[match.end(1):]
+                 words = match.group(1)
+                 encoded = charset.from_splittable(words, False)
+                 chunks.append((encoded, charset))
+                 continue
+             # Broken?
+             encoded = charset.from_splittable(splittable, False)
+             chunks.append((encoded, charset))
+             break
+         return chunks
+ 
      def encode(self):
          """Encode a message header into an RFC-compliant format.
  
*************** class Header:
*** 419,423 ****
          """
          newchunks = []
          for s, charset in self._chunks:
!             newchunks += self._split(s, charset, True)
          return self._encode_chunks(newchunks)
--- 467,473 ----
          """
          newchunks = []
          for s, charset in self._chunks:
!             chunks = self._split_nonascii(s, charset)
!             for s, charset in chunks:
!                 newchunks += self._split(s, charset, True)
          return self._encode_chunks(newchunks)