[darcs-users] tailor.py Failure getting upstream changes

[Nathan Gray]

On Thu, May 05, 2005 at 08:45:18AM -0400, Nathan Gray wrote:
> On Wed, May 04, 2005 at 07:22:34PM +0200, Lele Gaifax wrote:
> > Isn't /var/sync/capad/perl/modules/Apache the CVS repository? What
> > about manually re-executing exactly the same command tailor did?
> > What is its output?
> 
> If I run the command I think tailor is running, I get this:
> 
>   $ TZ=UTC cvs -f -d///var/sync/capad/perl/modules/ rlog -N -S -r:HEAD -d'2005-01-04 23:37:55 UTC<' Apache > /tmp/tailor6O0355cvs
> 
>   cvs [rlog aborted]: ///var/sync/capad/perl/modules/CVSROOT: No such file or directory
> 
> I think I must keep tailor pointing to my sandboxes as upstream, so that
> sync'ing back from darcs works.  So far, I haven't had any problems
> pulling down revisions with this setup, well, until last week when it
> stopped working.

Investigating further, I have found that specifying the actual upstream
CVS repository works.  So now my sync'ing continues.

> I am wondering if there might be some file that is corrupted in each
> sandbox.  Late last week, I started getting this error on one of the CVS
> modules, but not the rest.  Yesterday I tried each module independently,
> and all failed with the same error message.

This, it turns out, was a file permissions problem in the upstream
repository.

-kolibrie