On Mon, Jul 31, 2006 at 07:46:22AM -0400, Jeff Johnson wrote:
> and some rather simple code in rpm to use RPMTAG_DISTTAG as a
> tie-breaker if it exists,
That would be perfect, e.g. having rpmvercmp extend to evrd comparison
where d is the optional RPMTAG_DISTTAG.
> would seem to achieve the intended
> result much more effectively than trying to convince everyone to
> incorporate
> Release: 1.2.3.4.5.foo.bar.%{dist}
> as their packaging style.
I agree. And even better: If RPMTAG_DISTTAG could be passed to the
rpmbuild process from the outside, then no changes in macros/specfiles
would be needed at all, e.g.
rpmbuild --define 'disttag superfoo5' ...
would set a default value for Disttag: which would be used as the tie
breaker.
A non-empty Disttag would have to possibly still enter the package
name (otherwise the same package file will exist in different releases
with different content), e.g. something like
%{NAME}-%{VERSION}-%{RELEASE}%{?DISTTAG:.%{DISTTAG}}.%{ARCH}.rpm
--
Axel.Thimm at ATrpms.net
Attachment:
pgp91uOtaAX3w.pgp
Description: PGP signature