Re: UnicodeDecodeError getting revision log with Python

Jack Bates wrote:
>> David James wrote on Thu, 28 May 2009 at 15:58 -0700:
>>> Subversion standardizes on UTF-8 internally, but I don't
>>> think we verify that log messages are UTF-8 -- I think that users can
>>> encode their log messages using any encoding they want.
>> As of 1.6, we do. See validate_prop() in libsvn_repos/fs-wrap.c.
>
> May I file a feature request for,
>
> svn.fs.revision_prop(..., svn.core.SVN_PROP_REVISION_LOG)
>
> - to return a unicode object instead of a bytestream?

No.

Subversion doesn't make any promises about the revision properties except
that they are bytestreams. Some of them (the ones managed by Subversion ...
svn:log, svn:author, etc.) are "supposed to be" UTF-8 strings in order for
Subversion to function properly. But those are requirements of API
functions in levels higher than that of the svn.fs module (or its matching
svn_fs_* underlying C functions).

-- 
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2357592