On Mon, Dec 29, 2008 at 11:40:55PM +0100, Ralf Wildenhues wrote: > > I think the patch is mostly ok, but some details matter: you are not > adjusting _LT_ENABLE_LOCK, which probably means $LD is not getting an > appropriate '-m $emul' flag set, right?
This only appears to be an issue for biarch arches like x86_64, powerpc or s390, am I right? Note that this port is i386 only for now. > Secondly, at a glance I don't see support for your system's canonical > name in config.guess/config.sub. Have you already sent a patch to the > config-patc...@gnu.org address for that? I'm afraid it was rejected. The config maintainer has changed his policy and is now very strict about assigning new triplets. He requires a fully featured system with a user community, etc. And I suppose you understand that libtool support is essential before we can consider archieving this goal. > This matters because while matching *-gnu might be the right thing for > all systems with glibc, the ordering in the file may matter in that we > shouldn't by accident have some other case branch that matches it > earlier (this is mostly so we don't regress accidentally in the future). We're using 'i386-pc-kopensolaris-gnu' as triplet. I don't think it's likely that we match anything else by accident. As for non-glibc systems matching *-gnu, I verified (by inspecting config.guess) that the only possibilities are: - Linux-based systems (not a problem since you're matching them in the same checks anyway). - The old non-glibc Debian GNU/NetBSD port. I believe it is now defunct (I can confirm if you like), but even if it's not, it'd be possible to keep it happy by matching it before the generic *-gnu test. Thanks! -- Robert Millan The DRM opt-in fallacy: "Your data belongs to us. We will decide when (and how) you may access your data; but nobody's threatening your freedom: we still allow you to remove your data and not access it at all."