Thread: Re: pgFoundry WAS: On pgweb project

Re: pgFoundry WAS: On pgweb project

From
"Dave Page"
Date:

> -----Original Message-----
> From: Josh Berkus [mailto:josh@agliodbs.com]
> Sent: 10 March 2004 16:44
> To: Dave Page
> Cc: pgsql-www@postgresql.org
> Subject: Re: pgFoundry WAS: On pgweb project
>
> Dave,
>
> > Even for the mailing lists and all the subscribers?
>
> Even so.   Both GBorg and GForge use mailman, which makes
> moving stuff over a
> snap.

Yes, but all the list addresses will presumably change, and there are
bound to be users that don't notice for a while.

> The projects who will have difficultly are those who use
> pgsql-xxx@postgresql.org mailing lists off the main Majordomo
> server (such as ODBC), but then they had the same problem under GBorg.

No problem there as they won't change anyway.

> >I assume also that
> > users will have to coordinate with their projects to sort out fresh
> >CVS  checkouts at an appropriate time?
>
> Nope.   CVS should also be automatic.   Advantage of our
> virtual machines ....
> Right, Marc?

I don't see what VMs have to do with it. If you move the CVS to a new
host, then everyone using that repository will have to checkout a new
tree. It's only convenient to do that when all changes have been
committed.

>
> Then they wouldn't be much of a test, would they?   <grin>

pgAdmin would because it's a complex one but with no current users to
inconvenience.

The other two are simple ones that make easy tests, also with little
inconvenience should they go wrong. I assume there will be some sort of
redirect added from Gborg to the new project's home?

Anyway, makes little difference to me - just thought I'd offer some
simple test fodder...

Regards, Dave

Re: pgFoundry WAS: On pgweb project

From
"Marc G. Fournier"
Date:
On Wed, 10 Mar 2004, Dave Page wrote:

> Yes, but all the list addresses will presumably change, and there are
> bound to be users that don't notice for a while.

Actually, that's an easy matter of adding a forwarding alias ...

> > Nope.   CVS should also be automatic.   Advantage of our
> > virtual machines ....
> > Right, Marc?
>
> I don't see what VMs have to do with it. If you move the CVS to a new
> host, then everyone using that repository will have to checkout a new
> tree. It's only convenient to do that when all changes have been
> committed.

Oops, Josh had misunderstood your original, and I added to it ... yes, you
are correct in that there should be some co-ordination to ensure that all
commits are in before moving CVS ...

----
Marc G. Fournier           Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org           Yahoo!: yscrappy              ICQ: 7615664