[MERGE] Make annotate behave in a non-ASCII world

[Goffredo Baroncelli]

On Wednesday 11 July 2007, John Arbash Meinel wrote:
> So the problem is that we need to write some things is sys.stdout.encoding
> (with errors='replace') and some things as 8-bit strings.
> 

Yes, this is the same problem of the diff (the paths are unicode, the file 
content is raw 8-bit data).

I encountered the same problem in testing my patch "[RFC][PATCH] Show the diff 
in the commit messages": the diff function uses a stream to pass the data to 
the higher function.

I think that it is time to develop a common wrapper for this kind of 
situation: something like

   class UnicodeStringIO(StringIO.StringIO):
	def __init__(self, buf='', decoding='utf8'):
		StringIO.StringIO.__init__(self, buf)
		self._usio_decoding = decoding

        def write(self, s):
            if not isinstance(s, unicode):
                s = s.decode(self._usio_decoding, "replace")
            StringIO.StringIO.write(self, s)

this class returns always an unicode string. If the string passed to this 
wrapper is an internal bazaar [meta]data (as user name or path which are 
unicode string), the data are stored "as is". If the data is a string, the 
data are _decoded_ with the selected encoding, then it is stored. 

Any thougts

-- 
gpg key@ keyserver.linux.it: Goffredo Baroncelli (ghigo) <kreijack at inwind.it>
Key fingerprint = CE3C 7E01 6782 30A3 5B87  87C0 BB86 505C 6B2A CFF9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: unicodestringio.py
Type: application/x-python
Size: 1634 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20070711/a0e2ad90/attachment.bin 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20070711/a0e2ad90/attachment.pgp