On Sunday 14 December 2008 16:39:17 Jeremy Olexa wrote: > On Mon, Dec 8, 2008 at 10:33 AM, Jeremy Olexa <darks...@gentoo.org> wrote: > > I am seeking a positive code review on the following change to > > flag-o-matic.eclass, diff is below (reasons are below that): > > Er, cancel that. The proposed patch isn't robust enough to catch > something like "append-ldflags -Wl,--bad-flag"
sounds like we need a test-flag-LD() func that can go the extra mile
> As such, we will leave
> it local to Gentoo Prefix until we can come up with a better idea.
stick it behind userland_GNU ? i dont mind extending append-ldflags in such a
way, but the "pro" you listed originally (protect users from typos in ebuilds)
isnt a pro in my book -- it should fail.
i.e. something like:
append-ldflags() {
use userland_GNU || set -- $(test-flag-LD "$@")
......
}
-mike
signature.asc
Description: This is a digitally signed message part.
