tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: final steps for reproduceable builds.

On Mon, Mar 09, 2009 at 02:18:52PM -0400, Perry E. Metzger wrote:
> Manuel Bouyer <> writes:
> > On Mon, Mar 09, 2009 at 01:18:58PM -0400, Christos Zoulas wrote:
> >> I think that there should be a build option to restore the previous 
> >> behavior.
> >> After all, I am not even sure that gld is happy with the 0 timestamp in
> >> archives, since chris's patch did not disable the timestamp completely 
> >> before
> >> my last commit. Anyway, this build option should control all things that
> >> make builds deterministic, not just "ar" related.
> >
> > Seconded. And then we do non-deterministic by default, and deterministic
> > on the build clusters for releases.
> There is no obvious reason to do non-deterministic by default.

The fact that you can't see the value of build informations in the boot
blocks version printed, and in the kernel's uname output doesn't mean it
doesn't have value.  It's hightly valuable, especially for us debugging
user's kernel issues.

But I don't see a reason to start this thread again, it'll be as sterile as
the previous one.

Manuel Bouyer <>
     NetBSD: 26 ans d'experience feront toujours la difference

Home | Main Index | Thread Index | Old Index