On Sun, 07 Jul 2013 20:46:12 +0200 j...@wxcvbn.org (Jérémie Courrèges-Anglas) wrote:
> Christopher Zimmermann <madro...@gmerlin.de> writes:
>
> > Hi,
>
> Hi Chris,
>
> > I updated my port on openbsd-wip with most of the changes from both
> > of you. Testing on sparc64 would be fine though. That's what keeps
> > most of my OCaml ports from being committed since I don't have
> > access to a sparc64 machine anymore. Apart from this missing test
> > runs I'm ok.
>
> sparc64 or another non-native arch, right? Any volunteers?
Maybe even someone who could grant ssh access to such a machine?
> > On Sat, 06 Jul 2013 21:27:26 +0200
> > j...@wxcvbn.org (Jérémie Courrèges-Anglas) wrote:
> >
> >> Edd Barrett <vex...@gmail.com> writes:
> >>
> >> > OK?
> >>
> >> More work is needed IMO.
> Fine with me, I was just wondering about consistency. What about
> removing them all in one go?
If you are so eager...
> >> > MODULES = lang/ocaml
> >> > BUILD_DEPENDS = devel/ocaml-lambda-term \
> >> > sysutils/findlib
> >> > RUN_DEPENDS = ${BUILD_DEPENDS}
> >>
> >> RUN_DEPENDS = ${BUILD_DEPENDS} is never right, see make
> >> show=RUN_DEPENDS.
> >
> > Are you sure? I often use this contruct. Never had any problems
> > with it.
>
> /usr/ports/mystuff/devel/utop$ make show=RUN_DEPENDS
> devel/ocaml-lambda-term sysutils/findlib lang/ocaml sysutils/findlib
> devel/gmake devel/ccache lang/ocaml /usr/ports/mystuff/devel/utop$
>
> If utop needs findlib at runtime then spelling it out clearly as
> a RUN_DEPENDS would be much nicer.
I didn't test but I would guess utop does not strictly need findlib.
Still who would want to run utop without findlib??
Anyways in wip I removed findlib altogether and turned it around.
Makes more sense this way.
I'm OK with an update as in my last wip commit (9426a9ad).
I'm also pretty sure it will work on sparc64.
Christopher
pgp9D0O3RhGJM.pgp
Description: PGP signature
