Bug#710696: severity of 710696 is serious, because it's a FTBFS bug

[Santiago Vila]

On Thu, 29 Sep 2016, Michael Biebl wrote:

> Sigh, you really like to argue [...]

No, I don't really like to argue.

I was trying to be nice by explaining things with detail instead of
moving the discussion to -devel, the technical committee or the
release managers.

If you just "don't like to argue", no problem, I will ask release
managers to override your decision and everybody will be happy.


As I have already explained, we don't wait for bugs to happen in
buildd.debian.org before making them of serious severity.

Examples here:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?include=subject%3AFTBFS;submitter=lamby%40debian.org

or here:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?include=subject%3AFTBFS;submitter=lucas%40debian.org

or even here:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?include=subject%3AFTBFS;submitter=sanvila%40debian.org

Just count how many of those bugs have a reference to
buildd.debian.org and how many of them have a reference to our own
autobuilders (including the ones from reproducible builds).

If any of us had to wait for FTBFS bugs to happen in
buildd.debian.org, we would probably not file any bugs at all.

(Not to mention nobody would file bugs about Arch:all packages,
as many maintainers still do not upload in source-only form).

So what you propose contradicts common practice *completely*.

Instead, we file as "serious" any FTBFS which happens in a policy
compliant autobuilder, i.e. any autobuilder which builds packages in
the standard way (i.e. dpkg-buildpackage + fakeroot, not the weird
example you tried to put as an example that "whatever happens in
buildd.debian.org determines RC-ness").


Since you finally forwarded this upstream, I infer that you finally
managed to reproduce this, at least once.

Therefore, I ask again:

Can we already make this bug serious (as it should be) or do you
really need to check that my autobuilder is not misconfigured?


> instead of actually addressing the issue

The issue will be properly "addressed" when everybody can autobuild
this package without failures, as explained by release policy. Hiding
this problem by downgrading the severity is not "addressing the issue",
and it's also against the definition of sarge-ignore.

Thanks.