Re: [Caml-list] Release 3.09.1
On Wed, Jan 04, 2006 at 11:05:34PM +0100, Samuel Mimram wrote:
> Sven Luther wrote:
> >On Wed, Jan 04, 2006 at 04:45:03PM -0500, Eric Cooper wrote:
> >>On Wed, Jan 04, 2006 at 10:19:01PM +0100, Ralf Treinen wrote:
> >>>On Wed, Jan 04, 2006 at 08:19:12PM +0100, Sven Luther wrote:
> >>>>On Wed, Jan 04, 2006 at 06:51:01PM +0100, Samuel Mimram wrote:
> >>>>
> >>>>>Unless someone objects to it, I'll package and upload OCaml 3.09.1 as
> >>>>>soon as 3.09.0 is in testing. Do you all agree with a full rebuild?
> >>>>
> >>>>Yes, but let's do it like this :
> >>>>
> >>>> 1) we wait until ocaml 3.09.1 is built and installed on all
> >>>> autobuilders.
> >>>>
> >>>> 2) we compose a list of all ocaml related packages.
> >>>>
> >>>> 3) we ask Steve to requeue all those ocaml related packages for a
> >>>> binNMU,
> >>>> and file RC bugs against all those that fail.
> >>>
> >>>But we still have to do a source upload for those packages which Depend
> >>>on ocaml-*-3.09.0, I guess. I'm thinking of packages that are
> >>>architecture=all, and compiled to byte-code.
> >>
> >>The packages in the SVN repository that have rules to generate
> >>debian/control based on the OCaml API version will need to have a
> >>newly-generated debian/control committed first, and then uploaded, right?
> >>So it's not just a binNMU, unless I'm missing something.
> >
> >Well, only the build-depend was generated, the rest is provided by
> >dh_gencontrol, and as we used (ocaml >= 3.09.0) for build dependencies,
> >this
> >should work without problem, provided we guarantee that there is no older
> >ocaml 3.09.0 on the autobuilders.
>
> Just for information, is there a real advantage of making binNMUs
> instead of proper new uploads?
I believe that Steve can set a massive set of hints, and we don't have to
worry about making manual uploads of all those packages, and can concentrate
on stuff that breaks.
The idea is to not have us make binNMUs, but to trigger a binNMU rebuild on
the autobuilders. Not sure if it will work out, but if we can try it this time
to test out the theory, it would be neat.
Friendly,
Sven Luther
Reply to: