portupgrade -> portmaster Rosetta Stone?

[RW]

On Sat, 3 Mar 2012 07:15:23 -0500
Robert Huff wrote:

> 
> Doug Barton writes:
> 
> >  On 3/2/2012 11:06 PM, Conrad J. Sabatier wrote:
> >  > Doug, is there a way to emulate portupgrade's "-k" (keep going)
> >  > option, to have the remaining list of ports to be built still
> >  > continue processing even if one port's build fails?
> >  
> >  You haven't missed it, the answer is no. It's part of that
> >  "portmaster can't read minds" problem that if something fails, I
> >  have no way of knowing if the rest of the updates should stop as
> >  a result.
> 
> 	But ... isn't this a case where you don't have to read minds?
> It seems (to me) the user would be saying "I understand the risk,
> and accept responsibility for dealing with the consequences.".  At
> that point, whether thet're right or wrong is not your problem ....

I think it worth pointing out that portupgrade has two levels of "keep
going". By default portupgrade (and portmanager) will continue with
any port that doesn't depend on a failed port. This is more or less
what would have happened if it terminated on the first error, but you
had been luckier with the arbitrary part of the build order.

The minus -k flag allows it to build the rest as well. When you
finally fix the failed ports you end up with ports built out of
dependency order. 

In my experience it's the default "keep going" that's most useful
because it means that portupgrade usually completes most of what it
needs to do even if it fails on the first port.