[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Future of reasons.hs



On Tue, 26 Aug 2014 10:29:24 -0700
Joachim Breitner <nomeata@debian.org> wrote:

> Hi,
> 
> 
> Am Dienstag, den 26.08.2014, 19:12 +0200 schrieb Sven Bartscher:
> > However, solving this will be (or at least looks like) a lot of work.
> > Before I start solving this I wanted to ask if you consider reaons.hs
> > really useful. If it's not considered useful, it's not really worth the
> > effort to solve this problem. So please tell me what you're thinking.
> 
> In the current form the script is not very useful (as you say youself).
> OTOH we could use better tools that pin-point the exact problems, and
> help to ignore what are not problems.
> 
> Maybe we should not build another tool but rather improve 
> https://release.debian.org/transitions/html/haskell.html
> by telling Mehdi in what way it is not as useful for us.

Improving other tools seems much more desirable than making something
from scratch. Maybe we could at the DHG metting discuss what we would
imagine for that page.

> > The problem in detail:
> > Consider two packages A and B. A depends on B in a way that A and B can
> > only migrate together.
> > If now B has the same version in testing and in unstable, Britney won't
> > generate any excuses for it. However, it may have problems that block
> > the transition of A, but since there are no excuses, reasons.hs will
> > believe there is everything fine with B.
> > A solution to this could be, to teach resons.hs to generate it's own
> > excuses if Britney didn't do so.
> 
> Well, we don’t want to re-implement britney. (Or if we do, we might
> start from https://github.com/nomeata/sat-britney).
> 
> But I believe the information is available in
> https://release.debian.org/britney/update_output.txt. Currently, I see
> 
> trying: opendnssec/s390x
> skipped: opendnssec/s390x (0 <- 61)
>     got: 38+0: i-9:a-3:a-2:a-2:k-3:k-3:m-2:m-2:p-2:s-10
>     * s390x: libhsm-bin, opendnssec-dbg-mysql, opendnssec-dbg-sqlite3, opendnssec-enforcer-mysql, opendnssec-enforcer-sqlite3, opendnssec-signer
> 
> and I believe this means „I’m trying to migrate a binNMU of opendnssec
> to testing, but it fails by breaking these packages“
> 
> Does that help you?

A bit, but it still doesn't tell why it breaks those packages. That
information is usually provided by the excuses.

Attachment: signature.asc
Description: PGP signature


Reply to: