Le 21/01/2011 08:35, Ryan Schmidt a écrit :
On Jan 20, 2011, at 12:40, Samuel Langlois wrote:
I would like to know (and hopefully set!) the encoding used by svn log.
(I know about the --xml option, which works fine, but it is not easily readable
by a human.)
From what I see, it outputs ISO8859-1, so European àcçeñts are OK, but
Japanese, Chinese, etc. characters are output as question marks (?).
Is there a way, configuring either svn or Windows, to make it produce UTF-8 ?
It seems somewhat linked to the mysterious "Language for non-Unicode programs"
setting in the Regional Settings, but I could not make it work.
My understanding was that Subversion log entries are always stored as UTF-8 in
the repository, and that the conversion to and from your system's character
encoding, if different from UTF-8, are handled by your client, in response to
things like (on UNIX-like operating systems) the LANG environment variable (not
sure what Windows uses for that).
Thank you for your answer.
Yes, that's exactly the point: what does Windows use for that... and is
svn complying to it?
On Unix, you are right: it works perfectly well when you play with the
LANG variable.
But on Windows, svn log always outputs ISO8859-1 characters.
I tried fiddling with the chcp command or using cmd /U or changing
things in the Regional Settings... no way.
Again, it works well with svn log --xml, so it is really a matter of svn
log deciding which encoding it speaks.
Thank you
Samuel
