> There's a bug in gcc that causes it to go into a infinite loop while
> compiling xorg-server. I have a patch that I pulled out of bug 30052.
> http://cross-lfs.org/~jciccone/gcc-4.2.0-PR30052_unofficial-1.patch . I
> don't know if this has been commited yet. Hopefully I'll beable to check
> later today.
>
> Since glibc doesn't support i386 anymore. I had to pass
> CFLAGS="-march=i686 -mtune=generic -g -O2" to build 32bit and
> CFLAGS="-mtune=generic -g -O2" to build 64. This was a x86_64 multilib
> build. It's probably safe to assume that each of those flags will work
> x86 and x86_64-64 respectivly.
Why not use -pipe as well? Also, you can use -mtune=current, and it'll
do the thinking for ya. GCC ain't all bad. Just thought I'd interject.
> The reason we need to sed Coreutils, Gzip, and Tar is because glibc now
> defines futimens. Glibc's prototype for futimens isn't the same as the
> one used in the 3 packages. three seds are as follows.
>
> Coreutils:
>
> sed -i "s/futimens/gl_&/" lib/utimens.{c,h} src/{copy,touch}.c
>
> Gzip:
>
> sed -i "s/futimens/gl_&/" lib/utimens.{c,h} gzip.c
>
> Tar:
>
> sed -i "s/futimens/gl_&/" lib/utimens.{c,h} src/misc.c
>
> I'd like to see a patch makes the code compatible with glibc's
> implementation, but that's for another day.
I know I could be very useful in solving the toolchain problems, but
I don't have the internet ATM, so we're all SOL, I suppose. Ryan probably
can figure out how to get the new tools to mesh. I'd do it if I had a
computer to work from, but alas, it is not (currently) so. I'm hoping to get
the internet by the end of this month.
/me crosses his fingers
Oh yeah. Don't work yourself too hard, Joe ;)
_______________________________________________
Clfs-dev mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-dev
