On Saturday 23 July 2005 22:32, Ben Collins-Sussman wrote:
> On Jul 23, 2005, at 4:33 PM, kfogel@collab.net wrote:
> > Branko Čibej <brane@xbc.nu> writes:
> >> (I still don't understand the usefulness of sharing working copies,
> >> though...)
> >
> > In addition to the Extreme Programming teams mentioned by Morten
> > Kvistgaard, I have also found it helpful when co-maintaining a live
> > website working copy with others. So there are real-world uses for
> > sharing working copies.
>
> I should remind folks that we actually have a precedent of putting
> real effort into fixing bugs that hamper working-copy-sharing. So
> unless we're ready to make a reverse policy shift, we "officially"
> endorse the sharing of working copies already.
I think there is a difference between sharing working copies, and sharing
logins. I'm not sure how much work we should put behind people sharing
logins. In this particular case, 'svn --username' would cache the username
and password the first time the developer made a commit. From that point, he
doesn't have to enter that information again. When they switch, the new
developer would have to put --username on the command line again. Besides,
you still have to remember to change environment variable. So I'm not sure
this is really helping anything. And ideally, you wouldn't want to store
your authentication tokens in someone else's auth cache. With both
solutions, you still have to remember to pass --no-auth-cache, or set the
store-auth-creds=no in the config file (which limits usability when working
by yourself). I'd consider the caching of a person's credentials in someone
else's auth cache a bigger concern. *shrug*
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Jul 24 13:10:15 2005