Bug#844264: release.debian.org: Please clarify "Packages must autobuild without failure"

[Santiago Vila <sanvila@unex.es>]

On Wed, Dec 21, 2016 at 11:28:15PM +0200, Niko Tyni wrote:

> A severity:important bug does not mean it's all "ok". It means it's
> still a bug, but we can release with it. Autobuilders can build the
> package given a sane number of tries, security uploads can be built
> etc. If they can't due to an insane failure rate, the bug should be
> 'serious'. The decision between those is up to the maintainer and/or
> the release team.
> 
> That's how I think things have worked for a long time (well, at least
> as far as I can remember.)

Actually, I think we have always been more strict than that.

Some people apply an informal rule saying "RC if the failure happens in buildd".

If we assume 10 release architecture autobuilders, it follows that a
package which fails to build from source 10% of the time will fail on
one of the autobuilders on average. IMO, that should be already RC,
even if one particular upload does not fail in any autobuilder.

For pure Arch:all packages, even if the package is only officially
built once, our standards of quality should be the same, so if we
decide, say, that packages must build from source more than 90% of the
time, the same rule, IMO, should apply to Arch:all packages.


So, if establishing a threshold is the only way to achieve that for
example Bug #843038 in "elki" is upgraded to serious again, so be it,
but as I said, it would be a pity if we invest our time trying to
estimate probabilities instead of actually ensuring that packages
build all the time as a shared goal with reproducible builds.

Thanks.