You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Johann Glaser <gl...@ict.tuwien.ac.at> on 2006/02/02 10:04:12 UTC

Non UTF-8 filename problem

Hi!

I've got a problem with a filename, which contains an UTF-8 incompatible
character. When I checkout the repository via https (mod_dav_svn, Debian
package libapache2-svn version 1.1.4-2, Apache 2.0.54-5, logfile says
"SVN/1.2.3 (r15833) neon/0.24.7") or when I do an update for the
half-checkedout repository, I get the following error message:

svn: REPORT request failed on '/svn/conas/!svn/vcc/default'
svn: The REPORT request returned invalid XML in the response: XML parse error at line 10: Input is not proper UTF-8, indicate encoding !
Bytes: 0xFC 0x62 0x65 0x72
. (/svn/conas/!svn/vcc/default)

The file name is 'Fleetnet-Internet_on_the Road_Projektübersicht.pdf'
and the 'ü' character in 'Projektübersicht' is hex 0xFC. 

When I do a checkout via a file URL, the error message says:

A  PDFs/Fleetnet-Internet_on_the Road_Projektübersicht.pdf
svn: Malformed XML: not well-formed (invalid token) at line 3

Only when checking out with a file url, the file .svn/log is written. It
holds the filename several times with the wrong UTF-8 character.
Therefore svn commands which use this file (e.g. cleanup) also complain
about wrong XML.

Fortunately the file can be accessed with a normale web browser via
HTTPS.

Some search fo this problem revealed some bugs and discussions, but all
issues are resolved.
  http://svn.haxx.se/dev/archive-2004-12/0027.shtml
  http://subversion.tigris.org/issues/show_bug.cgi?id=2157
  http://svn.haxx.se/dev/archive-2004-12/0014.shtml

What is wrong? Why could that file be added to the repository (this was
done on 6th December 2005)? How can I rename the file within the
repository without using XML?

Thanks
  Hansi



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Non UTF-8 filename problem

Posted by kf...@collab.net.
Johann Glaser <gl...@ict.tuwien.ac.at> writes:
> I'm sorry I can't test this issue with another server, because this is a
> production machine. I'm also sure that we do have some more files with
> umlauts in their names which didn't make any problems so far. But I'll
> ask my colleague to try the newest client.
> 
> Anyway, the server sends (invalid) XML data with a single 'ü' inside.
> Eiter the client gets more tolerant or the server does an UTF-8
> translation before sending the XML data. But this XML metadata also has
> to match real filenames. I assume this is transported in two different
> ways/places, since the HTTPS URL works with the single 'ü' inside.

If you could send the results of your testing to the dev@ list, that
would be great.  Just link back to this thread so people can catch up
on the background.  That link is:

http://subversion.tigris.org/servlets/BrowseList?list=users&by=thread&from=426560

Thanks,
-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org


Re: Non UTF-8 filename problem

Posted by Johann Glaser <gl...@ict.tuwien.ac.at>.
Hi!

> If you have any way to check this problem with a latest trunk client
> and server, that would be terrific.  I realize that may be more work
> than you thought you were signing up for, just thought I'd ask... :-)

I'm sorry I can't test this issue with another server, because this is a
production machine. I'm also sure that we do have some more files with
umlauts in their names which didn't make any problems so far. But I'll
ask my colleague to try the newest client.

Anyway, the server sends (invalid) XML data with a single 'ü' inside.
Eiter the client gets more tolerant or the server does an UTF-8
translation before sending the XML data. But this XML metadata also has
to match real filenames. I assume this is transported in two different
ways/places, since the HTTPS URL works with the single 'ü' inside.

Bye
  Hansi



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Non UTF-8 filename problem

Posted by kf...@collab.net.
Johann Glaser <gl...@ict.tuwien.ac.at> writes:
> > Thanks for doing so much research to see if the bug had already been
> > reported!
> > 
> > What is the version of your svn client?  (I couldn't quite tell from
> > your report, but it seems you only said the server version...)
> 
> The SVN client also has version 1.2.3 (Debian package 1.2.3dfsg1-3).
> 
> The same problem (with a slightly different error message) also happens
> with TortoiseSVN on a colleague's windows box, but I don't know its
> version. I think the checkin was also done with a TortoiseSVN client.
> I'll investigate this in more detail tomorrow.

If you have any way to check this problem with a latest trunk client
and server, that would be terrific.  I realize that may be more work
than you thought you were signing up for, just thought I'd ask... :-)

-Karl

-- 
www.collab.net  <>  CollabNet  |  Distributed Development On Demand

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Non UTF-8 filename problem

Posted by Johann Glaser <gl...@ict.tuwien.ac.at>.
Hi!

> Thanks for doing so much research to see if the bug had already been
> reported!
> 
> What is the version of your svn client?  (I couldn't quite tell from
> your report, but it seems you only said the server version...)

The SVN client also has version 1.2.3 (Debian package 1.2.3dfsg1-3).

The same problem (with a slightly different error message) also happens
with TortoiseSVN on a colleague's windows box, but I don't know its
version. I think the checkin was also done with a TortoiseSVN client.
I'll investigate this in more detail tomorrow.

Bye
  Hansi



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Non UTF-8 filename problem

Posted by kf...@collab.net.
Johann Glaser <gl...@ict.tuwien.ac.at> writes:
> I've got a problem with a filename, which contains an UTF-8 incompatible
> character. When I checkout the repository via https (mod_dav_svn, Debian
> package libapache2-svn version 1.1.4-2, Apache 2.0.54-5, logfile says
> "SVN/1.2.3 (r15833) neon/0.24.7") or when I do an update for the
> half-checkedout repository, I get the following error message:
> 
> svn: REPORT request failed on '/svn/conas/!svn/vcc/default'
> svn: The REPORT request returned invalid XML in the response: XML parse error at line 10: Input is not proper UTF-8, indicate encoding !
> Bytes: 0xFC 0x62 0x65 0x72
> . (/svn/conas/!svn/vcc/default)
> 
> The file name is 'Fleetnet-Internet_on_the Road_Projektübersicht.pdf'
> and the 'ü' character in 'Projektübersicht' is hex 0xFC. 
> 
> When I do a checkout via a file URL, the error message says:
> 
> A  PDFs/Fleetnet-Internet_on_the Road_Projektübersicht.pdf
> svn: Malformed XML: not well-formed (invalid token) at line 3
> 
> Only when checking out with a file url, the file .svn/log is written. It
> holds the filename several times with the wrong UTF-8 character.
> Therefore svn commands which use this file (e.g. cleanup) also complain
> about wrong XML.
> 
> Fortunately the file can be accessed with a normale web browser via
> HTTPS.
> 
> Some search fo this problem revealed some bugs and discussions, but all
> issues are resolved.
>   http://svn.haxx.se/dev/archive-2004-12/0027.shtml
>   http://subversion.tigris.org/issues/show_bug.cgi?id=2157
>   http://svn.haxx.se/dev/archive-2004-12/0014.shtml
> 
> What is wrong? Why could that file be added to the repository (this was
> done on 6th December 2005)? How can I rename the file within the
> repository without using XML?

Thanks for doing so much research to see if the bug had already been
reported!

What is the version of your svn client?  (I couldn't quite tell from
your report, but it seems you only said the server version...)

-Karl

-- 
www.collab.net  <>  CollabNet  |  Distributed Development On Demand

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org