[Pkg-swan-devel] Bug#846026: Bug#846026: strongswan: FTBFS randomly (failing tests)

[Yves-Alexis Perez]

On Sun, 2016-12-18 at 15:16 +0100, Santiago Vila wrote:
> I disagree slightly.

Ok.
> 
> If we allow the 24864 packages in stretch to fail to build 1% of the time,
> try to calculate the probability that you get no failures if you build
> all of them.

It's not a matter of “allowing”. We do know packages FTBFS, all the time, for
various reason. When it's reliable, sure, we fix it. When it's not, we try to
fix it too. But at some point it's just not practical.
> 
> By my count you will get no failures once every 3x10^109 times,
> approximately.

And we don't really care. We usually don't build Stretch once and for all. So
if a package FTBFS, we try to rebuild it, and it's fine.

Also, that's the reason why the tests exist actually: we want them to fail if
there's a problem. Here it's usually because of entropy problem, but then just
fix the buildds, maybe?

Regards,
-- 
Yves-Alexis
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: This is a digitally signed message part
URL: <http://lists.alioth.debian.org/pipermail/pkg-swan-devel/attachments/20161218/a0c25d0d/attachment.sig>