You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Aur Saraf <so...@gmail.com> on 2008/03/19 16:29:22 UTC

Probable bug in svn client

Hello,

I think I have stumbled upon a bug in the SVN client.

The bug is described in this chat log of #svn:

http://colabti.org/irclogger/irclogger_log/svn?date=2008-03-19,Wed&sel=396#l711

Basically:

repo structure:
trunk
trunk/dir

working copy:
$ ls -a .
.svn dir
$ ls -a dir
.svn
$ svn status
? dir
 # that is, dir is itself a checkout of trunk/dir, unrelated to the
checkout of trunk into . that happened before dir was added to the
repo
$ svn add dir
 bork bork bork dir already in repo
$ svn revert dir
$ mv dir dir.moved
$ svn update
 bork bork bork dir already scheduled for addition
 # but it isn't! I reverted it! and there isn't even a directory ./dir anymore!
$ svn status
? dir.moved

This is not reproduction of the bug on a shell, just writing from the
top of my head. I might have missed something. If you think this is
really a bug, I can send a long transcript of my shell session, as
well as the content of ./.svn (or did I forget to back those up before
I did the workaround that worked? damn, think I did...)

Oh, of course:

[son@moaddib ~]$ uname -a
Linux moaddib 2.6.24-ARCH #1 SMP PREEMPT Wed Mar 5 12:07:52 UTC 2008
i686 Intel(R) Core(TM)2 CPU T5600 @ 1.83GHz GenuineIntel GNU/Linux
[son@moaddib ~]$ svn --version
svn, version 1.4.5 (r25188)
   compiled Sep 29 2007, 17:48:30

 -- Aur

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

Re: Probable bug in svn client

Posted by David Glasser <gl...@davidglasser.net>.
On Mon, Mar 31, 2008 at 3:21 PM, Aur Saraf <so...@gmail.com> wrote:
> On Mon, Mar 31, 2008 at 11:10 PM, David Glasser
>  <gl...@davidglasser.net> wrote:
>  > If you're interested in more investigation of this bug, a full
>  >  reproduction recipe would be helpful.
>  >
>  >  --dave
>
>  Does it look like a bug at first sight? Developers would be far better
>  at that kind of guess than newbie users, and it would probably take me
>  hours to research a recipe, so I'd want such a first opinion before I
>  try to reproduce it...

"First sight" for me includes searching the source code for the exact
text of error messages, among other things.

(My first instinct is "this is yet another bug caused by our wc
design, and I'd rather put the effort into centralizing wc metadata
than fixing yet another bug in the current implementation".)

--dave


--
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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

Re: Probable bug in svn client

Posted by Aur Saraf <so...@gmail.com>.
On Mon, Mar 31, 2008 at 11:10 PM, David Glasser
<gl...@davidglasser.net> wrote:
> If you're interested in more investigation of this bug, a full
>  reproduction recipe would be helpful.
>
>  --dave

Does it look like a bug at first sight? Developers would be far better
at that kind of guess than newbie users, and it would probably take me
hours to research a recipe, so I'd want such a first opinion before I
try to reproduce it...

 -- Aur

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

Re: Probable bug in svn client

Posted by David Glasser <gl...@davidglasser.net>.
If you're interested in more investigation of this bug, a full
reproduction recipe would be helpful.

--dave

On Wed, Mar 19, 2008 at 9:29 AM, Aur Saraf <so...@gmail.com> wrote:
> Hello,
>
>  I think I have stumbled upon a bug in the SVN client.
>
>  The bug is described in this chat log of #svn:
>
>  http://colabti.org/irclogger/irclogger_log/svn?date=2008-03-19,Wed&sel=396#l711
>
>  Basically:
>
>  repo structure:
>  trunk
>  trunk/dir
>
>  working copy:
>  $ ls -a .
>  .svn dir
>  $ ls -a dir
>  .svn
>  $ svn status
>  ? dir
>   # that is, dir is itself a checkout of trunk/dir, unrelated to the
>  checkout of trunk into . that happened before dir was added to the
>  repo
>  $ svn add dir
>   bork bork bork dir already in repo
>  $ svn revert dir
>  $ mv dir dir.moved
>  $ svn update
>   bork bork bork dir already scheduled for addition
>   # but it isn't! I reverted it! and there isn't even a directory ./dir anymore!
>  $ svn status
>  ? dir.moved
>
>  This is not reproduction of the bug on a shell, just writing from the
>  top of my head. I might have missed something. If you think this is
>  really a bug, I can send a long transcript of my shell session, as
>  well as the content of ./.svn (or did I forget to back those up before
>  I did the workaround that worked? damn, think I did...)
>
>  Oh, of course:
>
>  [son@moaddib ~]$ uname -a
>  Linux moaddib 2.6.24-ARCH #1 SMP PREEMPT Wed Mar 5 12:07:52 UTC 2008
>  i686 Intel(R) Core(TM)2 CPU T5600 @ 1.83GHz GenuineIntel GNU/Linux
>  [son@moaddib ~]$ svn --version
>  svn, version 1.4.5 (r25188)
>    compiled Sep 29 2007, 17:48:30
>
>   -- Aur
>
>  ---------------------------------------------------------------------
>  To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
>  For additional commands, e-mail: dev-help@subversion.tigris.org
>
>



-- 
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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