Somehow it goes in infinite recursion, this is *not* good! Please confirm and eventually restrict test on that package, thanks!
Hmm, Diego, I wonder as there is no sci-electronics/geda-gaf in the tree. Can you please recheck?
I think you mean =sci-electonics/geda-1.6.0-r1. That package fails here too during test with your stated problem.
Yeah sorry not -gaf in there :)
After some fiddling with the problem it finds an easy solution. Add the following src_test() to the ebuild and all goes well. src_test() { cd "${S}" emake -j1 check || die "test failed" } Btw: The first tests on AMD64 are promising. The ebuild works well and seems to be made nicely. One question arises however on the definition of RDEPEND: RDEPEND="${RDEPEND} sci-electronics/electronics-menu" I am wondering if that should DEPEND or CDEPEND.
(In reply to comment #4) > After some fiddling with the problem it finds an easy solution. > > Add the following src_test() to the ebuild and all goes well. > > src_test() { > cd "${S}" > emake -j1 check || die "test failed" > } > > Btw: The first tests on AMD64 are promising. The ebuild works well and seems to > be made nicely. I'll test this solution. > One question arises however on the definition of RDEPEND: > > RDEPEND="${RDEPEND} > sci-electronics/electronics-menu" > > I am wondering if that should DEPEND or CDEPEND. > Yes, it's wrong. The right var is ${CDEPEND}
Created attachment 213498 [details, diff] patch to geda-1.6.0-r1.ebuild This patch (with Thomas's fixes :) fix this bug. Unfortunatelly Denis is away, and I haven't commit privileges yet, so would be great if someone can commit this. Thanks
After hitting this once again and almost losing the tinderbox I committed it ;) Let's hope it's all fixed by this (I'm a bit doubtful to be honest because I think it _was_ already running it that way, but still).