On Mon, Jul 08, 2013 at 03:10:07PM -0400, Joey Hess wrote:
> Kirill Smelkov wrote:
> > Hello up there. I've tried to use git-annex with repo name in russian, and
> > though basic git operations (commit/push/pull) work, git-annex fails to
> > exchange annexed files - saying it cannot access remote repo. Details below.
>
> I have reproduced this bug, but not when using a UTF-8 capable locale
> like LANG=en_US.utf8
>
> So I reproduced it with LANG=C. What's your locale?
My locale is ru_RU.UTF-8:
kirr@mini:~$ locale
LANG=ru_RU.UTF-8
LANGUAGE=
LC_CTYPE="ru_RU.UTF-8"
LC_NUMERIC="ru_RU.UTF-8"
LC_TIME="ru_RU.UTF-8"
LC_COLLATE="ru_RU.UTF-8"
LC_MONETARY="ru_RU.UTF-8"
LC_MESSAGES="ru_RU.UTF-8"
LC_PAPER="ru_RU.UTF-8"
LC_NAME="ru_RU.UTF-8"
LC_ADDRESS="ru_RU.UTF-8"
LC_TELEPHONE="ru_RU.UTF-8"
LC_MEASUREMENT="ru_RU.UTF-8"
LC_IDENTIFICATION="ru_RU.UTF-8"
LC_ALL=
> > # NOTE here "мир" becomes "CCC8D0" which is wrong and that's why it
> > fails
>
> This is probably haskell's special encoding for unrepresentable characters,
> using a high unicode code page. git-annex is careful to ensure data read from
> git config --list is stored in this encoding (otherwise just reading it would
> crash with an encoding error).
>
> Normally, haskell automatically converts back from this encoding to bytes
> when accessing the filesystem. However, in this case, it's running a command
> instead and passing it in a parameter, which seems to bypass that conversion.
I see, thanks for explanation. If there is anything else on my side I
could assist with reproducing the bug - please ask.
Thanks again,
Kirill
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
