Jonathan Ashley wrote:
> Hi,
>
> I have a file that I can't read from the repository with Tortoise. Trying
> "Open with..." from the repo-browser, or double-clicking the "Modified ..."
> line from the log window of its last changed revision produces the error
>
> GET of
> '/svn/imp/!svn/bc/62933/dev-stable/FCS/SNMP/System_Data_Manager/Proof/system_data_manager_/performance/performance_monitoring_tt.rlu':
> 200 OK (https://servername)
>
> However, I can read it fine using the command line SVN, and I can see it in
> a web browser too.
>
> Using Tortoise 1.4.5 build 10425, Windows XP client.
> Server is 1.4.3 on Windows 2003.
> The command-line client that works is "1.5.0 (dev build) compiled Dec 15
> 2006, 09:55:27". I've got a feeling that I may have built that myself, not
> sure. I don't tend to use the command-line for svn much any more.
>
> I'm not too worried about this particular file, but anything with a hint of
> repository corruption makes me nervous - I'd like to get to the bottom of
> the problem.
>
> Any help offered is much appreciated. Can anyone tell me what the problem
> is?
>
> More information:
>
> - behaviour is same with http: instead of https,
>
> - I tried to reproduce the problem by checking the previous version of the
> file into a sandbox area, then replacing it with the problem version. It
> doesn't show the problem, so it doesn't seem as though the contents of the
> file are causing the difficulty,
>
> - A similarly-named file (performance.rlu) in the same folder also works
> fine, so the path to the file seems ruled out too,
>
> - I used fsfsverify.py to check the revision file, but it doesn't seem to
> detect any error,
>
> - I suspect the file was committed from a corrupted working copy. The WC
> claimed to be up to date when I investigated, but the file was actually an
> older version that should been replaced by a copy from another part of the
> repository. I wonder if the WC could have sent a diff based on incorrect
> data that did not cleanly apply to what was on the server? Wouldn't
> explain why the command-line tools still work OK though.
>
> - If I pick "Save revision to..." off the "Modified ..." line from the log
> window of the last changed revision, and save the file, I get the same
> error from Tortoise - but the file itself seems to save correctly.
>
> - The Apache access log looks fine, or at least not obviously wrong. The
> final number in the GET, which I think is the length of the transfer, is
> slightly less than the size of the file on disk, but I've checked against
> other files and that appears to be normal.
>
> - The file has properties 'svn:mime-type : text/plain', 'svn:keywords : id'
> and 'svn:eol-style : native'.
>
You should maybe ask that question on the Subversion users mailing list.
Since this only happens with one specific file, I suspect some
corruption of the repository or some other bug in SVN.
You could also try a nightly build of TSVN and check if the issue still
is there (beware of the working copy format update).
Stefan
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.net
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tortoisesvn.tigris.org
For additional commands, e-mail: users-help@tortoisesvn.tigris.org
Received on Tue Nov 27 18:52:31 2007