Failed to make uClibc 0.9.26

[Christian MICHON]

this is the reply from Erik Andersen to an email I sent eons ago, off list...

On Sun Mar 02, 2003 at 09:04:06AM +0100, Christian MICHON wrote:
> > The tuxscreen patches can stay in the tuxscreen repository...
> > Hopefully tinyx.mk will be much more generally applicable.
>
> Yep. I can only check on x86 on my side (at least 3 machines).
> How about Arm & co ?

I have arm, x86, powerpc, and mips boxes locally I test on.

> > It is easy enough to add 5 dependancies and wget each, one at a
> > time.
>
> That's what I thought too, but I thought you had a more general
> rules for this.

Nothing so formal...

> > I think gcc 2.95 (via the GCC_2_95_TOOLCHAIN option in the top
> > level Makefile) is as far back as I care to go.  Making and
> > maintaining a working toolchain is a lot of work....
>
> I understand. Actually, I was not asking for a support, just a
> quick-n-dirty C-only toolchain. I managed to build a real
> bootstrapped gcc-2.7.2.3 + binutils-2.9.1 with a GNU-uberbaum
> approach. It bootstrapped -O2 in ~2 minutes ;)

nice!  If you want to submit a uclibc_toolchain_2_723.mk
I'd be happy to integrate it....

> I just used it to build all of X430 in ~9 minutes. Now you
> understand maybe why I needed an old but fast compiler.
> The timing numbers are from a P4 2.4GHz.

I have just an Athlon 1600 I build everything on.  Sometimes when
I'm doing toolchain work it would be nice to have a quad P4 3 Ghz
or similar -- would certainly help.

-Erik

--
Erik B. Andersen             http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--