[coreboot] [PATCH] fix stage0_i586.S in v3
c-d.hailfinger.devel.2006 at gmx.net
Sat Apr 5 01:52:03 CEST 2008
On 04.04.2008 22:22, Juergen Beisert wrote:
> On Friday 04 April 2008 21:34, Myles Watson wrote:
>>> -----Original Message-----
>>> From: coreboot-bounces at coreboot.org
>>> [mailto:coreboot-bounces at coreboot.org] On Behalf Of Juergen Beisert
>>> Sent: Friday, April 04, 2008 1:29 PM
>>> To: linuxbios at linuxbios.org
>>> Cc: Ed Swierk; coreboot at coreboot.org
>>> Subject: Re: [coreboot] [PATCH] fix stage0_i586.S in v3
>>> On Friday 04 April 2008 20:43, Ed Swierk wrote:
>>>> On Fri, Apr 4, 2008 at 11:28 AM, Peter Stuge <peter at stuge.se> wrote:
>>>>> Also there is crossdev and just plain Gentoo, either just vanilla
>>>>> binutils ebuilds or even catalyst for a complete custom system
>>>> And buildroot and LFS and lots of others too. I'm not knocking any of
>>>> those, but they are more complex than what I was looking for. My
>>>> 50-line shell script (attached) builds a cross-compiling gcc (4.2.2) +
>>>> binutils (2.17) + uClibc (0.9.29) toolchain, nothing more.
>>> What about a reference toolchain everybody can build on his host and use
>>> compile coreboot? I believe we will out of luck to support every
>>> toolchain on
>>> this planet.
>> I think it is nicer to have a black list. Hopefully the build succeeds in
>> more instances than it fails. Otherwise we might pick a reference
>> toolchain that turns out to have bugs later.
> But it's easier to fix. You fix one reference toolchain and make everyone
> Black lists only gets bigger and bigger...
Turns out this was not about whitespace, but the "/" character used in
the division was interpreted as a comment character if GNU as had been
compiled for an i*86-elf target (commonly seen on MacOSX).
Stefan committed a fix in r655.
More information about the coreboot