On Fri, Sep 17, 2021 at 9:43 AM Jan Kiszka <jan.kiszka@siemens.com> wrote:

> On 17.09.21 18:25, Anton Mikanovich wrote:
> > 17.09.2021 04:48, Tim Orling wrote:
> >>
> >> Can someone please revisit debian-sid-ports and add tests for it? I
> >> hacked it to work locally (archive 2021 key and “buster” InRelease
> >> key), but I am not familiar enough to know what is hack and what was
> >> necessary. Proving sifive-fu540 is buildable should be enough.
> >>
> > I've just sent `[PATCH 0/2] Fix isar-bootstrap for sifive-fu540` on
> > maillist, but there are still more issues at least with cowsay, hello
> > and kernel building.
> > Adding sifive-fu540 to CI is also planning to be done.
> >
>
> Will be tough to find issues via CI because, due to sid, it must
> tolerate failures by nature.
>
> Agreed that sid will be failure prone. If there is a way to at least make
sure the bootstrap works and tolerate component build failures, it would be
a good thing?

Anton's fixes are simpler than mine. The hard value assignment of
DISTRO_APT_SOURCES seems to have been the key to avoid the other hacks I
did locally.