On Mon, Jul 22, 2019 at 10:46:37AM -0300, Chris Lamb wrote:
> One question: Does your proposal also imply a source package being
> strictly reproducible in unstable or, alternatively, did you mean to
> imply it "not regressing" from being previously reproducible in
> unstable?
>
> This latter idea could be said to be fairer but requires the storage
> of state on either your "end" or ours and thus complicates the
> technical machinery, but more importantly IMHO it makes it somewhat
> opaque from the perspective of package maintainers...
I think britney already can track regressions, it does that for piuparts
tests at least.
intrigeri brought up another interesting angle on this:
currently (un)reproducibility bugs are are of severity normal, so it
seems a bit premature to block or even delay testing migrations because
of that. if however we treated regressions as important bug, than maybe
we should treat *all* new important bugs as delaying migrations at
least. say each new important bug would/could delay migrations by 2 days.
what do you think?
--
tschau,
Holger
-------------------------------------------------------------------------------
holger@(debian|reproducible-builds|layer-acht).org
PGP fingerprint: B8BF 5413 7B09 D35C F026 FE9D 091A B856 069A AA1C
Attachment:
signature.asc
Description: PGP signature