Re: debootstrap/1.0.102 appears to break debuerreotype autopkgtest
On Mon, 18 Jun 2018 at 02:43, Paul Gevers <elbrus@debian.org> wrote:
> It would be nice if you responded to this idea of Ansgar, as the
> regression in debuerreotype is currently delaying the migration of
> debootstrap. Is the new behavior of debootstrap really bad for
> debuerreotype or is it just resulting in a different image which is fine
> otherwise for your package?
>
> In the former case, you should file an RC bug in that case to block
> debootstrap from migrating until debuerreotype is ready.
>
> In the latter case, you can unblock the migration of debootstrap by
> adapting debuerreotype and/or its autopkgtest (and please let me know).
> Or, e.g. because you are lacking time, you can tell me that you are fine
> with the current regression and I can instruct the migration software to
> ignore it (less preferred).
>
> Just to be absolutely clear, if nothing happens, debootstrap will
> migrate in 10 days to testing and the new behavior will be the baseline
> for autopkgtesting.
To be frank, I'm a little surprised that debuerreotype is believed to
be the blocker here -- I replied via mail to Ansgar (which I now see
you got dropped from CC on -- my apologies, that mail is at [1]), and
uploaded debuerrotype 0.7 [2], which implements the behavior he's
described (explicitly using "--merged-usr" / "--no-merged-usr" instead
of relying on the implicit default).
[1]: https://lists.debian.org/debian-boot/2018/06/msg00178.html
[2]: https://lists.debian.org/debian-boot/2018/06/msg00179.html
Then, I realized he was actually describing what the behavior in
_debootstrap_ should be, and I sent an initial debootstrap patch to
the list [3] fixing the implementation bug (which Hideki [4] turned
into a merge request at [5]).
[3]: https://lists.debian.org/debian-boot/2018/06/msg00182.html
[4]: https://lists.debian.org/debian-boot/2018/06/msg00190.html
[5]: https://salsa.debian.org/installer-team/debootstrap/merge_requests/12
IMO, that merge request needs some review and an upload and then this
bug will be fixed properly in debootstrap too. I think the only
points in question are whether defaulting the variable to the empty
string (and adjusting that empty default based on suite later where we
actually know it) is the desired approach and what to do with
EXTRACT_DEB_TAR_OPTIONS / adding "-k" (from bug 838388) [6].
[6]: https://lists.debian.org/debian-boot/2018/06/msg00186.html
Cheers!
♥,
- Tianon
4096R / B42F 6819 007F 00F8 8E36 4FD4 036A 9C25 BF35 7DD4
Reply to: