"portmaster -o" behaves a little strangely (IMHO)

[Conrad J. Sabatier]

On Tue, 24 Jan 2012 22:01:33 +0000
Chris Rees <crees at freebsd.org> wrote:

> On 24 January 2012 20:31, Doug Barton <dougb at freebsd.org> wrote:
> > On 01/24/2012 06:19, Conrad J. Sabatier wrote:
> >> ===>>> Gathering dependency list for lang/gcc46 from ports
> >> ===>>> Launching child to update gcc-4.4.7.20111108 to
> >> gcc-4.4.7.20120117
> >>         lang/gcc46 >> gcc-4.4.7.20111108
> >>
> >> (What is the preceding line -- with the ">>" -- and the similar one
> >> below, supposed to be indicating exactly?  If it means what I
> >> think it does, then it seems as if portmaster is misinterpreting
> >> the intention of the "-o" option)
> >
> > No, it means that the ports framework is seeing 4.4 as a dependency
> > of 4.6 for some reason. I've seen that happen in the past when
> > non-base gcc was in use, but have never had anyone track down the
> > cause authoritatively. Obviously make.conf is a suspect, but I'm
> > assuming you already looked there.
> 
> Often happens when people stick USE_something into make.conf :)
> 
> (don't do that)
> 
> Chris

Hmmm, interesting.  I do, in fact, have a USE_GCC=4.2+ line
in /etc/make.conf, conditional on ${.CURDIR} for certain ports that
just won't build with clang, which I'm using now as the default.

Doug seems to think, though, that it's something else within the ports
framework's handling of the gcc ports that's tripping up portmaster.
Not sure I even want to begin to try to track that one down.  :-)

Anyway, thanks everybody for the feedback.

-- 
Conrad J. Sabatier
conrads at cox.net