Summary from yesterdays (mini) FESCo meeting

[Michael Schwendt]

On Fri, 29 Dec 2006 11:33:44 -0500, Brian Pepple wrote:

> On Fri, 2006-12-29 at 11:19 -0500, Jesse Keating wrote:
> > I'd be more in favor of a 
> > rule that just says "items checked need to be listed out in the review before 
> > building of the package will be allowed".  Vague enough as to not give 
> > reviewers a shortcut.
> 
> That sounds fine to me.  The problem I had was reviewers just putting
> 'APPROVED' in reviews, and not giving any information on what was
> actually checked.

It doesn't make sense to create detailed lists. A single "APPROVED" is
fine. I've done that multiple times myself, because everything else is too
time-consuming. Even my old-style reviews have been inconsistent and
misleading to the silent observer, because they never mentioned everything
I had checked. I can catch many packaging bugs and pitfalls with the blink
of an eye. And at the same speed it is possible to verify many things one
must not find in a spec. You don't want to slow-down the possibly
experienced reviever and force him to create detailed lists.