Re: Soooooooo...
On Thu, Jul 19, 2001 at 07:48:33PM -0400, utsl@quic.net wrote:
> ii binutils 2.11.90.0.7-2 The GNU assembler, linker and binary utiliti
> ii cpp 2.95.3-7 The GNU C preprocessor.
> ii cpp-2.95 2.95.4-0.01042 The GNU C preprocessor.
> ii gcc 2.95.3-7 The GNU C compiler.
> ii gcc-2.95 2.95.4-0.01042 The GNU C compiler.
>
> * gcc: was a complete PITA until I figured out that I didn't have the archtable
> setup right in dpkg, and configure was doing bogus things with the arch. I
> also disabled everything but gcc itself. (No gcj, g77, gpc, etc.) I think I
> had to hack the Makefile for some hurd/linux logic that didn't handle
> freebsd. There were a couple packages like that.
Is this using the stock Debian package for gcc? What have you defined the
architecture as?
> I keep running into problems due to not having glibc, so I'm taking a look at
> porting it next, rather than trying to port all of Debian off of it. Hopefully
> that'll take care of some of the really icky problems. (i.e. sysvinit,
> debianutils, etc.) It looks like I need to port ld.so, too, and probably have
> to hack the FreeBSD kernel to use it, because of the dpkg-shlibdep problem.
> (That's actually easy. It's a one-line patch to switch the loader path,
> which'll have to happen anyway to become FHS compliant.)
That would be pretty cool. I'd be interested to see the patches you end up
with. A complete GNU system on top of a BSD kernel (rather than a mainly
GNU with a small amount of BSD one) would be something to aim for, I think
(not having a GNU libc reduces the value of the Debian GNU/BSD a little,
IMHO :) ) but I'm currently looking at just using the NetBSD one to avoid
having to rewrite huge chunks of code...
> Anyway, I'm working on FreeBSD, making decent progress, and having a lot of
> fun. If this helps people working on NetBSD, that's cool too.
Sure. I'm more interested in NetBSD due to the increased portability (I'm
likely to get hold of an Alpha in the near future, but it's Turbo Channel
and so won't run Linux. Having Debian on it anyway would be nice :) ), but
I think the issues involved are going to be pretty much identical in
both. If we need to contribute any patches to the package maintainers, I'd
expect them to be beneficial to both kernels.
--
Matthew Garrett | mjg59@srcf.ucam.org
Reply to: