[Mageia-discuss] Bug reference for updates-announce not always correct

[Michael Scherer]

Le dimanche 01 janvier 2012 à 12:16 +0100, Manuel Hiebel a écrit :
> Le dimanche 01 janvier 2012 à 13:06 +0200, Thomas Backlund a écrit :
> > 01.01.2012 10:33, Vinet Raphaël skrev:
> > > Hi,
> > >
> > >
> > > I don't know if it is really important and if it is often (I just began
> > > with these ml) but I think there is a problem somewhere for some emails
> > > sent in the updates-announce mailing list.
> > >
> > >
> > > Example (the last one I received) :
> > >
> > > [updates-announce] Package update: timezone-2011n-2.1.mga1 (core)
> > >
> > >
> > > It is written in the email :
> > >
> > > "Timezone is a package that contains data files with rules for various
> > >
> > > timezones around the world. This update addresses the following
> > >
> > > changes:
> > >
> > > - Fiji adopts DST for 2011 (effective Oct 23rd, 2011)
> > >
> > > - West Bank changes date for DST end in 2011 to Sep 30th
> > >
> > > - Fix DST for: Pridnestrovian Moldavian Republic, Ukraine,
> > >
> > > Bahia and Brazil.
> > >
> > > https://bugs.mageia.org/show_bug.cgi?id=3034"
> > >
> > >
> > > When you clicked the 3034 it is not related with timezone at all and in
> > > fact the related bug report is the
> > > https://bugs.mageia.org/show_bug.cgi?id=3045
> > >
> > >
> > > --> just a typo error ?
> > >
> > 
> > Yep.
> > I was tired and unfortunately made a typo when I pushed that update.
> > 
> > --
> > Thomas
> 
> What can the QA do to avoid such of thing ? :)
> Add the bug url in the advisory so you have only to copy-paste ?

We could validate that the bug number is indeed the good one :
- print the summary ( not ideal )
- check that the src.rpm name is correct ( ie, if the bug is filled
against glibc and that the update is for python, show a error or ask
confirmation ).

( of course, the first part would be to have the script in svn ).

If we really want to be lazy, we could have a more streamlined process,
like extracting the advisory directly from comment, etc.

-- 
Michael Scherer