utf8: Add imap4_utf_7 codec

Add code to reencode IMAP folder names to regular utf-8.
This starts an implementation that will add a new config option
`utf8foldernames` on account level which will fix #299 and on the
long run replace the current `decodefoldernames` option.

This commit introduces code to register an `imap4_utf_7` codec
on which two-way conversion methods will later be built.

Original code by
(https://www.blogger.com/profile/16648963337079496096),
taken from
http://piao-tech.blogspot.no/2010/03/get-offlineimap-working-with-non-ascii.html

In the comment
http://piao-tech.blogspot.com/2010/03/get-offlineimap-working-with-non-ascii.html?showComment=1316041409339#c669880170006851138
indicates that this code is expected to be incorporated into offlineIMAP and therefore the author implicitly agrees to put it under this license.

Signed-off-by: Urs Liska <git@ursliska.de>
Signed-off-by: Nicolas Sebrecht <nicolas.s-dev@laposte.net>
This commit is contained in:
Urs Liska 2017-10-02 01:06:18 +02:00 committed by Nicolas Sebrecht
parent 1ce596d713
commit 032376efad

View File

@ -17,6 +17,8 @@
import re
import string
import binascii
import codecs
from offlineimap.ui import getglobalui
@ -370,3 +372,73 @@ def decode_mailbox_name(name):
return ret.decode('utf-7').encode('utf-8')
except (UnicodeDecodeError, UnicodeEncodeError):
return name
# Functionality to convert folder names encoded in IMAP_utf_7 to utf_8.
# This is achieved by defining 'imap4_utf_7' as a proper encoding scheme.
def modified_base64(s):
s = s.encode('utf-16be')
return binascii.b2a_base64(s).rstrip('\n=').replace('/', ',')
def doB64(_in, r):
if _in:
r.append('&%s-' % modified_base64(''.join(_in)))
del _in[:]
def encoder(s):
r = []
_in = []
for c in s:
ordC = ord(c)
if 0x20 <= ordC <= 0x25 or 0x27 <= ordC <= 0x7e:
doB64(_in, r)
r.append(c)
elif c == '&':
doB64(_in, r)
r.append('&-')
else:
_in.append(c)
doB64(_in, r)
return (str(''.join(r)), len(s))
# decoding
def modified_unbase64(s):
b = binascii.a2b_base64(s.replace(',', '/') + '===')
return unicode(b, 'utf-16be')
def decoder(s):
r = []
decode = []
for c in s:
if c == '&' and not decode:
decode.append('&')
elif c == '-' and decode:
if len(decode) == 1:
r.append('&')
else:
r.append(modified_unbase64(''.join(decode[1:])))
decode = []
elif decode:
decode.append(c)
else:
r.append(c)
if decode:
r.append(modified_unbase64(''.join(decode[1:])))
bin_str = ''.join(r)
return (bin_str, len(s))
class StreamReader(codecs.StreamReader):
def decode(self, s, errors='strict'):
return decoder(s)
class StreamWriter(codecs.StreamWriter):
def decode(self, s, errors='strict'):
return encoder(s)
def imap4_utf_7(name):
if name == 'imap4-utf-7':
return (encoder, decoder, StreamReader, StreamWriter)
codecs.register(imap4_utf_7)