Package Review Stats for the week ending January 18th, 2009

[Christoph Wickert]

Am Donnerstag, den 29.01.2009, 01:15 +0100 schrieb Kevin Kofler:
> Christoph Wickert wrote:
> > >       * Recently I updated some of the Xfce 4.6 packages. One of them
> > >         was approved without _any_ docs.
> > https://bugzilla.redhat.com/show_bug.cgi?id=477732
> > also all the desktop files were installed and listed in %files twice and
> > if the reviewer had tested the package he would have noticed that. Site
> > note: The reviewer has been made a sponsor 2 weeks later.
> 
> s/he/she/ ;-)

Thanks, didn't know that.

> Not that it really matters, I'm just being pedantic. ;-)

Me too, epically on reviews. ;)

> >> >       * A font package was approved although it contained another font
> > 
> > https://bugzilla.redhat.com/show_bug.cgi?id=481501
> 
> In this case the reviewer clearly said what needs to be fixed and the
> version which got imported was fixed, so it wasn't that bad. Maybe it would
> have made more sense to wait for a fixed version, but are there actually
> any issues with what was imported?

None that I know of. The package owner told me a different story on IRC
and I was under the impression that the reviewer did not spot the
problem.
For me it is ok to say "fix this before import", at least if the
packager is experienced enough. 

> > Do RH employes have sponsors too? A lot of the bad reviews are done by
> > RH people and a lot of bad specs come from RH folks.
> 
> In the case of #459535, neither the original reviewer (before you rejected
> his review and rereviewed it) nor the people who commented before were RH
> employees. The reviewer in #481501 wasn't from Red Hat either. You can't
> just blame it all on RH 

I didn't do that, sorry if it sounded like this. I just say "a lot of
reviews". AFAICT RH people are known to do a lot of 3-word-revievs
("looks fine. Approved")

> > Somebody pointed me 
> > to:
> > https://bugzilla.redhat.com/show_bug.cgi?id=433678
> > and I had a quick glance over it before Andreas added his comments:
> >       * no list of tests that have been run
> 
> Bad.
> 
> >       * SourceURL is missing
> >       * I can't even find the source because URL is wrong
> 
> Very bad.
> 
> >       * without the source you cannot check the License tag, md5, etc
> 
> You could check the License by extracting the source from the SRPM. But it's
> pretty likely that wasn't done here.

Yes, the package is GPLv2+ but GPLv2.

> >       * docs not marked %doc
> 
> RPM marks files in some directories as %doc automatically, AFAIK %{_docdir}
> is one of those.

Correct, but IMO there should always be a %doc line, at least for the
usual stuff like AUTHORS, COPYING, ChangeLog etc.

>         Kevin Kofler

Regards,
Christoph