Packaging review guidelines clarification

[Paul Howarth]

On Thu, 2006-02-16 at 16:08 +0100, Ralf Corsepius wrote:
> On Thu, 2006-02-16 at 14:08 +0000, Paul Howarth wrote:
> > On Thu, 2006-02-16 at 06:20 -0600, Mike McGrath wrote:
> > > > I think a failure to build in mock is a blocker, unless the reason is
> > > > either a deficiency in mock (in which case there should be a reference
> > > > to the bugzilla ticket for the issue raised on mock), or a dependency
> > > > not available in Core or Extras yet (which can easily be worked around
> > > > by adding a local repo containing the missing dependency to the
> > > > reviewer's mock configuration).
> > > >
> > > > Remember that the build system uses mock, so if it won't build in mock,
> > > > it won't get built for Extras at all.
> > > >
> > > > Paul.
> > > 
> > > I'm all for it, should we move 'should build on mock' to 'must build
> > > on mock' in the wiki?
> > 
> > The problem with that is that not every reviewer has the bandwidth to
> > support a mock build environment (particularly for development),
> <grin/> Set up a local one, that's what I'm doing ...

It's what I do too, but it still needs lots of bandwidth to keep the
rawhide and FE development mirrors up to date.

> >  so it's
> > probably left as a "should", but a failure being a blocker.
> Must be "must", because the buildsys uses mock, so a Review without mock
> build isn't worth the bandwidth and time it requires.

It's much more useful to test builds in mock, I agree, but much of
what's needed in a review doesn't need mock and I think just about
everything *could* be done a different way (e.g. by using things like
fedora-rmdevelrpms to check buildreqs).

Paul.