On Fri, 3 Sep 2010, Jeff Johnson wrote:
On Sep 3, 2010, at 10:13 AM, Yury V. Zaytsev wrote:
On Fri, 2010-09-03 at 16:10 +0200, Dag Wieers wrote:
%{?el6:# Tag: rfx}
%{?el5:# Tag: rfx}
%{?el4:# Tag: rf}
%{?el3:# Tag: rf}
Which is better IMO than forking into 4 SPEC files and maintaining each
seperately.
I think this is a very good idea.
Hint: I am perfectly prepared to generate a patch for ANY
version of RPM that helps with multi-arch, multi-distro, multi-fork,
portability.
Just ask on <[email protected]>, describe what you want, and give
me a sample *.spec (that will be used for "make check" regression/testing),
and I WILL take care of the rest.
Jeff, I appreciate your offer, but we have to work with upstream RPM
versions. Hence the use of comments to instruct the buildsystem ;-)
It is debatable whether we want RPM to be smarter than it is regarding
this. We will have to rely on logic that will be located elsewhere in the
stack anyhow.
Also, it is debatable whether the SPEC file should be used for this
additional meta-information, but my personal preference is that a single
recipe is better than one composed of pieces living in different
locations.
Some people have resorted to turning away from the SPEC file itself in
a quest to come up with the ideal buildsystem (by using XML or generating
the SPEC file) but I prefer to be close to what is effectively used. What
you see is what you get :)
It's definitely an interesting subject to discuss, but I am more a
pragmatic rather than a theoretic ;-)
--
-- dag wieers, [email protected], http://dag.wieers.com/ --
[Any errors in spelling, tact or fact are transmission errors]
_______________________________________________
users mailing list
[email protected]
http://lists.rpmforge.net/mailman/listinfo/users
