workspace corruption issues

[Peter Kahn <pk...@connected.com>]

I seem to lose about 4 hours to work time a week to strange corrupted
workspace issues in svn.  A merge where files have been added in the
range of changesets will result in the files placed in the workspace,
but not marked as adds, or changes will not go into the repository, or
the status command doesn't report the changes accurately.

 

In each of these cases, I suspect that someone violated the rules of
workspace etiquette, but I have no way to determine this.  This is
resulting in lower user confidence which is a problem for me.

 

Has anyone else had similar problems and if so, how did you investigate
them?  In each case, by the time I get there, all I have is a corrupt
workspace with no record of how it got that way.  Interviewing the users
doesn't seem to provide information that point to a problem.  Still,
there is a problem and I really want to determine how it got that way.

 

Anyone have any ideas (short of writing some kind of logging wrapper for
svn, which would still miss the other half of the equations, user
actions directly manipulating the files.)  

 

--

Peter Kahn

pkahn@connected.com

 

Re: workspace corruption issues

[Max Bowsher <ma...@ukf.net>]

Peter Kahn wrote:
> I seem to lose about 4 hours to work time a week to strange corrupted
> workspace issues in svn.  A merge where files have been added in the
> range of changesets will result in the files placed in the workspace,
> but not marked as adds, or changes will not go into the repository, or
> the status command doesn't report the changes accurately.
>
> In each of these cases, I suspect that someone violated the rules of
> workspace etiquette, but I have no way to determine this.  This is
> resulting in lower user confidence which is a problem for me.
>
> Has anyone else had similar problems and if so, how did you investigate
> them?  In each case, by the time I get there, all I have is a corrupt
> workspace with no record of how it got that way.  Interviewing the users
> doesn't seem to provide information that point to a problem.  Still,
> there is a problem and I really want to determine how it got that way.
>
> Anyone have any ideas (short of writing some kind of logging wrapper for
> svn, which would still miss the other half of the equations, user
> actions directly manipulating the files.)

Perhaps you could instruct your users to save a copy of their shell's 
history when they discover a problem?

Although that logging wrapper might not be a bad idea anyway....

Max.


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

RE: workspace corruption issues

[Dale Worley <dw...@pingtel.com>]

-----Original Message-----
From: Christopher Ness [mailto:chris@nesser.org]

$ fc -nlr

Should tell the story in a BASH shell.  
--------------------------

"history" works with many Unix-like shells.

Dale


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

Re: workspace corruption issues

[Christopher Ness <ch...@nesser.org>]

On Thu, 2004-30-12 at 17:00 -0500, Peter Kahn wrote:
> Anyone have any ideas (short of writing some kind of logging wrapper
> for svn, which would still miss the other half of the equations, user
> actions directly manipulating the files.)  

$ fc -nlr

Should tell the story in a BASH shell.  

If they are not using a bash shell or are using some other OS there
should be some way to record the commands the user is issuing to figure
out how they wedged their working copy.

HTH's,
Chris
-- 
Software Engineering IV,
McMaster University
PGP Public Key: http://www.nesser.org/pgp-key/
17:18:11 up 9 days, 5:23, 2 users, load average: 0.01, 0.01, 0.00 
http://www.fsf.org/philosophy/no-word-attachments.html