You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Andreas Schultz <an...@gmail.com> on 2005/06/24 14:25:55 UTC

svnversion weirdness with svn:externals

Hi,

After upgrading a svn server from 1.1.4 to 1.2, doing a svnversion in
a working copy that has svn:externals, reports as mixed version based
on the versions of the working copy and the external, e.g. 352:422M.

This very same problem that had been discussed and solved almost two
years ago here:
http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=48985

Any ideas how to get the old behavior back?

Regards
Andreas

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


Re: svnversion weirdness with svn:externals

Posted by Philip Martin <ph...@codematters.co.uk>.
Andreas Schultz <an...@gmail.com> writes:

> After upgrading a svn server from 1.1.4 to 1.2, doing a svnversion in
> a working copy that has svn:externals, reports as mixed version based
> on the versions of the working copy and the external, e.g. 352:422M.
>
> This very same problem that had been discussed and solved almost two
> years ago here:
> http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=48985

Ah yes, that got broken when svn_wc_status got revised to
svn_wc_status2.  A classic case of a fix without a regression test
failing to remain fixed.

> Any ideas how to get the old behavior back?

Tweak the svn_client_status2 call in svnversion/main.c and change the
last FALSE to TRUE.  I'll work up a more complete fix that takes
advantage of svn_client_status2.

-- 
Philip Martin

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

Re: svnversion weirdness with svn:externals

Posted by "C. Michael Pilato" <cm...@collab.net>.
kfogel@collab.net writes:

> Current head of trunk (r15159) still has the buggy behavior:
> 
>    $ svn up
>    
>    Fetching external item into 'svntest'
>    External at revision 15159.
>    
>    At revision 1428.
>    $ svn st -q
>    
>    Performing status on external item at 'svntest'
>    $ svnversion
>    1428:15159
>    $ 
> 
> I agree that this is a bug, or at least a suboptimal behavior.  An
> acceptable solution would be to just ignore svn:externals entirely.

Done in revision r15161.

> An even better solution might be to print "E" at the end to indicate
> that externals are present in the working copy, thus:
> 
>    $ svnversion
>    1428E

This had that kinda rank smell associated with a floodgate, so I left
it closed.

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

Re: svnversion weirdness with svn:externals

Posted by kf...@collab.net.
Andreas Schultz <an...@gmail.com> writes:
> After upgrading a svn server from 1.1.4 to 1.2, doing a svnversion in
> a working copy that has svn:externals, reports as mixed version based
> on the versions of the working copy and the external, e.g. 352:422M.
> 
> This very same problem that had been discussed and solved almost two
> years ago here:
> http://subversion.tigris.org/servlets/ReadMsg?listName=dev&msgNo=48985
> 
> Any ideas how to get the old behavior back?

In other words, you want the pre-0.27 behavior?  It looks like the
sequence has been:

   * Up to and including 0.27, svnversion ignored svn:externals.

   * From 0.28, until sometime before 0.32.1, it payed attention to
     svn:externals.

   * Sometime between 0.32.1 and 1.1.4, it started ignoring
     svn:externals again.

   * Sometime between 1.1.4 and 1.2.0, it started paying attention to
     svn:externals again.

That's pretty amazing :-).

The "M" on the end of your example, "352:422M", is extraneous, by the
way, right?  The real issue here is the presence of the ":422", and
whether the external happens to be locally modified ("M") or not is
irrelevant.

Current head of trunk (r15159) still has the buggy behavior:

   $ svn up
   
   Fetching external item into 'svntest'
   External at revision 15159.
   
   At revision 1428.
   $ svn st -q
   
   Performing status on external item at 'svntest'
   $ svnversion
   1428:15159
   $ 

I agree that this is a bug, or at least a suboptimal behavior.  An
acceptable solution would be to just ignore svn:externals entirely.
An even better solution might be to print "E" at the end to indicate
that externals are present in the working copy, thus:

   $ svnversion
   1428E
   $ 

Would you mind opening an issue for this bug, and pointing to this
mail thread from the issue?  If you don't have time, that's fine, just
let me know either way.  (And of course, if you have time to make a
patch to fix the bug, that's always great!)

Thanks for the report,
-Karl

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