How to calculate priority for missing tests or %check

Aleksandar Kurtakov akurtako at redhat.com
Tue Feb 25 11:14:38 UTC 2014


----- Original Message -----
> From: "Aleksandar Kurtakov" <akurtako at redhat.com>
> To: "For testing and quality assurance of Fedora releases" <test at lists.fedoraproject.org>
> Sent: Tuesday, February 25, 2014 1:13:25 PM
> Subject: Re: How to calculate priority for missing tests or %check
> 
> ----- Original Message -----
> > From: "Alexander Todorov" <atodorov at redhat.com>
> > To: "For testing and quality assurance of Fedora releases"
> > <test at lists.fedoraproject.org>
> > Sent: Tuesday, February 25, 2014 12:54:42 PM
> > Subject: How to calculate priority for missing tests or %check
> > 
> > Hi guys,
> > I've started working to identify packages with missing upstream test suites
> > or
> > not running the tests in %check. Having a list of hundreds of packages now
> > I
> > need to prioritize them somehow. This will help in later steps when spec
> > files
> > need to be fixed or somebody wants to start working with upstream on
> > writing
> > a
> > test suite.
> > 
> > My idea is to take all bugs against Fedora in the last few years (or
> > releases).
> > Packages with bigger number of bugs receive higher priority than those with
> > less. Then based on this prioritize.
> > 
> > 
> > How long should the time frame be?  I'm thinking 3 to 5 years.
> > 
> > 
> > Where low/medium/high priority boundaries are ? (could be 30%, 60%, 90% for
> > example).
> > 
> > 
> > Do you have other ideas how to calculate priority of these items?
> 
> My starting point would be from dependency POV. Packages with more
> dependencies (repoquery is your friend) can break more things if broken thus
> deserve more testing.

To clarify ordering based on the number of packages depending on the given one.

> 
> Alexander Kurtakov
> Red Hat Eclipse team
> 
> > 
> > 
> > Thanks,
> > Alex
> > --
> > test mailing list
> > test at lists.fedoraproject.org
> > To unsubscribe:
> > https://admin.fedoraproject.org/mailman/listinfo/test
> 


More information about the test mailing list