Re: experimental does not like Haskell
On 2015-08-28 16:44, Hector Oron wrote:
> Hello Joachim,
>
> 2015-08-28 15:54 GMT+02:00 Joachim Breitner <nomeata@debian.org>:
> > Am Freitag, den 28.08.2015, 15:44 +0200 schrieb Hector Oron:
>
> > On the buildd side, if apt/aptitude does not find a solution, it is
> > already wrong. So if, as done in Johannes’ patch, in this case sbuild
> > tries again using a different solver, I cannot see how that could make
> > matters worse in any way than they are now.
>
> Adding the fallback in failure case it's already an improvement, I am
> fine with that, if none complains, let's make it happen.
In practice the fallback might not be a so good idea, given aptitude
only exits (or get killed) when it has uses all the memory, which might
takes a few hours.
We know that apt is fine when used in stable/testing/unstable, so let's
keep it. We know that apt simply doesn't work for experimental, and
that aptitude works only for simple cases. If we have a solution for
experimental which works all (or most) of the times, we should use it
by default, without fallback.
Aurelien
--
Aurelien Jarno GPG: 4096R/1DDD8C9B
aurelien@aurel32.net http://www.aurel32.net
Reply to: