cvs commit: ports/shells/zsh/files patch-Src_watch.c

[Chris Rees]

On 14 Nov 2011 11:09, "Baptiste Daroussin" <bapt at freebsd.org> wrote:
>
> On Sun, Nov 13, 2011 at 10:45:35AM +0000, Chris Rees wrote:
> > On 13 November 2011 09:56, Baptiste Daroussin <bapt at freebsd.org> wrote:
> > > bapt        2011-11-13 09:56:08 UTC
> > >
> > >  FreeBSD ports repository
> > >
> > >  Added files:
> > >    shells/zsh/files     patch-Src_watch.c
> > >  Log:
> > >  - Fix build for badly upgraded systems which have both utmp.h and
utmpx.h
> > >
> > >  PR:             ports/156263
http://www.FreeBSD.org/cgi/query-pr.cgi?pr=156263
> > >  Submitted by:   Matthew Cashdollar <matthew at cashdollar.org>
> > >  Feature safe:   yes
> > >
> >
> > Fixing it here will only mask the problem-- why are we adding a fix to
> > support a broken system?
> >
> > The proper solution is to rm utmp.h in that system, is it not?
> >
> > I suggest a different patch at [1].
> >
> > Chris
> >
> > [1] http://www.bayofrum.net/~crees/patches/zsh-utmpx-fix.diff.txt
> > [removing the patchfile too]
>
> Well no I don't think, or you would have to add the same patch in lots of
ports.
>
> In fact, I have been pondering for long what would be the best fix for
this. And
> if I need to provide a fix.
> Given that some people can use zsh for root (as I do :)) I think being
able to
> rebuild zsh a clean way before doing the make delete-old is nice, that is
why I
> did a fix that way.

I understand where you're coming from, but I would also like to point out
that make delete-old doesn't remove any libs; it's safe to run with old
ports around.

I would still be inclined to blame this on the docs; I'm working on a patch
right now.

In the meantime, make delete-old should be run before rebuilding.

Chris