fbsd06 at mlists.homeunix.com
Sat Dec 22 18:57:01 PST 2007
On Sat, 22 Dec 2007 18:47:52 -0600
Paul Schmehl <pauls at utdallas.edu> wrote:
> --On December 23, 2007 1:19:21 AM +0100 Peter Schuller
> <peter.schuller at infidyne.com> wrote:
> > In particular, given a re-build (e.g. upgraded) port X, all ports
> > depending on X will also be re-built regardless of whether that is
> > required according to the dependency relation. This is handled in
> > such a way that it is not dependent on the entire procedure
> > completing in one session, as you are with portupgrade (meaning
> > it's restartable, as mentioned above).
> I don't understand this statement. I have killed portupgrade on
> numerous occasions, both locally and remotely, and have never had a
> problem restarting later.
portupgrade -fr perl
is pretty hard to restart efficiently.
> > In practice, I find this is the most useful upgrading method. I have
> > never been able to use portupgrade for more than a week or two on a
> > real machine without running into issues (stale dependencies,
> > failed builds due to weak dependency information, etc).
> I *really* don't understand this. I can count on one hand the number
> of times that I've run into dependency problems with portupgrade, and
> all of those were addressed in /usr/port/UPDATING or by simply
> deinstalling and reinstalling the port in question.
It was really intended to handle major upgrades where multiple UPDATING
instructions run together. And back in the days when Gnome upgrades
involved wrapping portupgrade in a shell script run in single-user
mode with a 50:50 chance of success, portmanager just took it in its
stride. I think it is a useful approach because it trades a lot of cpu
cycle for me not having to sober-up and think about things - and that
always a win. Unfortunately, it's gone without developer support for
too long now and I'm getting a bit wary about it.
More information about the freebsd-questions