find-debuginfo.sh

[Jeff Spaleta]

On 8/8/07, Roland McGrath <roland at redhat.com> wrote:
> If I had the machine resources and an automated thing to rebuild everything
> in mock against a new rpm-build, I would have tried it.  (Sorry for using
> Fedora public development and testing, for, you know, developing and testing.)

Some maintainers do local builds in mock, before sending things off to
the build server.  If we could get the buildid stuff into local mock
environments for testing easily before it goes live in the build
system, you might find some more in-the-wild problems from people
willing to go the extras mile and be helpful, before you get the
pitchforks and torches from another niche group.

> Honestly, I don't think it's really going to be all that bad.  The fires
> will be small and quickly quenched, as with mock--at least when I'm awake.
> My plan all along was to have a few days of pain getting the regime in place.
> I intend to stay available to quickly hash out any build issues with
> build-id and find-debuginfo stuff, when the new script hits and through the
> coming mass rebuild.  (But I hadn't planned to forego sleep entirely.)

I think the public commitment to help people out will go a long way in
smoothing out the introduction period.  Would it be possible to have
the new script provide a wikipage url in the build.log if there is an
error generated by this script? And in the wiki page you can summarize
the gotchas as you find them and provide contact info for maintainers
who run into a new problem yet to be listed?

-jef