Assigning priorities to GNATS reports

[Joe Buck]

> > How about the following:
> >
> > [<branch-to-fix> regression (<branch-first>-<branch-last>)]
> 
> I don't have any objection to this (the first part is the information
> I usually want most), but it seems a bit unwieldy to me.  If, however,
> people think we can maintain that, it's fine by me!

I think it would be fine to list only <branch-to-fix> in the synopsis
(preferably in a rigidly consistent form, so we can put regular
expressions in our GNATS advanced queries).  We need the latter
info as well, but it could appear in the body of the report somewhere
(or even in the audit trail, as it will tend to change).