Re: How severe are FTBFS bugs caused by the source using uname?
Steve Langasek <vorlon@debian.org> writes:
> On Wed, May 19, 2004 at 12:02:10AM +0200, Goswin von Brederlow wrote:
>
>> how severe are FTBFS bugs caused by the source using uname to
>> determine the host/build architecture instead of using
>> dpkg-architecture?
>
>> The problem is that on multiarch architectures, most notably amd64
>> cpus, the uname will not resemble the host/build architecture:
>> (from a normal Debian Sid i386 system running a 64bit kernel)
>
>> % uname -a
>> Linux dual 2.6.5-amd64 #2 Sun May 9 16:34:33 UTC 2004 x86_64 GNU/Linux
>> ^^^^^^
>
> A FTBFS on an architecture that is not yet a release candidate is not an
> RC bug. Presumably, using uname does not cause substantial problems on
> release candidate archs, or this would have come up for discussion long
> ago. Since your policy citation includes no "must" requirements, I
> don't see that this is anything more than an important bug.
It is an architecture of a release candidate: _i386_.
Its just not one of the old CPUs previously used for that
architecture: _x86_64 instead of i?86.
Think of x86_64 as i786.
The problem is that the architecture and the cpu differ and uname
looks at the cpu and not the architecture. Yes, its something new
since the cpu doesn't match the "i?86-*" pattern in the configure
scripts anymore.
But I can live with important. I don't think wishlist is ok as some
maintainer seem to think. Any other thoughts?
MfG
Goswin
Reply to: