Re: Thoughts about the transcoding of path names in SVN - it's incontradiction to CVS, SVK and Mercurial...

Hi Ryan,

On Wed, 25 Jun 2008, Ryan Schmidt wrote:
> Subversion stores all filenames in UTF-8 in the repository. It converts

I noticed that.

> between the repository's UTF-8 and the client's character encoding by
> relying on the user to properly set the LANG environment variable to a
> locale that makes sense for their system.

Well, it works without trouble with my linux and windows clients. :)

 
> I don't see how any software could be thought to function correctly
> today if it did not take into consideration the character encoding of
> its data.

Well, there are obviously different opinions about it. I could point you
to a thread on mercurials user mailing list which goes on about that, in
case you are interested in the arguments of mercurial developers...

> In fact, Subversion's UTF-8 support is overly simplistic, which causes
> nearly insurmountable issues for Mac users who must deal with files with
> non-ASCII names which were originally committed from Windows or Linux.

No problem for me, since I'll never come into touch with Macs.

> Subversion must go further than just identifying the character encoding;
> it must also normalize composed/decomposed UTF-8 characters. See this
> issue:
>
> http://subversion.tigris.org/issues/show_bug.cgi?id=2464

Interesting link!

So, I see that this issue is everywhere not so easy to handle. But at
least SVN does a good job for me between WXP using TortoiseSVN and linux.
:) Who knows, perhaps it would be different if I'd try to use svn from a
windows console. So I am happy that TortoiseSVN handles it properly... :)

Thanks for this clarifying post.

Regards,
Marko

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: users-help_at_subversion.tigris.org
Received on 2008-06-25 13:40:57 CEST