Re: status update?
En réponse à Sven Luther <luther@dpt-info.u-strasbg.fr>:
> > It is a hack because it is not necessary to do this split. You
> propose
> > to make ocaml_3.06.orig.tar.gz an empty new package and people who
> > used to get the sources with apt-get source are going to be
> confused.
>
> No, i just want to upload ocaml-3.06_3.06.orig.tar.gz, i won't touch
> the
> existing package for now at least.
>
> Mmm, i understand you argument about the apt-get source, but this is a
> new one if i am not wrong. Do you really think it would be a problem ?
Well, I like the idea of ocaml_x.xx.tar.gz being the pristine upstream
tarball, unless there is no other choice.
Currently, there is no problem on the user side but only on the
infrastructure side. There are problems that are independent from
our will, RC bugs that is. Either we help fixing them, or we wait
for them to be fixed.
If you think that removing those packages from woody will fix,
then this is a good solution right now.
> BTW, what would be the apt-get source reaction if i build depended the
> ocaml package on ocaml-3.06 ?
I think this is not natural at all. It don't see the benefit for the
user. I don't see any reason on the user's side.
> > We don't have to do bugware in order to work around others'
> problems.
> > I think it would be possible to ship everything in a single package,
> > even this versioned package, instead of having multiple packages.
>
> But does the benefit of having only a single package (if they are such
> benefits) outweight the benefit of having multiple packages ?
But we don't have real problems with our packages. We are able to fix
our problem. Problems come from others currently.
> > > new
> > > name, so it can go into testing. Additionnally, all ocaml packages
> > > should depend on the -3.06-1 variation and thus will be as happy
> with
> > > the true -3.06-1 packages as they were with the virtual ones. No
> > > change
> > > on is required on either the other ocaml packages or for the user.
> The
> > > user will not see this package anyway, until i upload the ocaml
> > > 3.06-16
> > > package that is.
> > >
> > > > RC bugs are not our fault and we don't have to do bugware
> > > > on our side to solve them!
> > >
> > > Well the only two alternatives are really :
> > >
> > > o We forget about the mini-freeze and all the work it has
> implied,
> > > and
> > > compromize our chances of having 3.06 in sarge. At least this
> would
> > > imply an effort equal to the mini-freeze later on.
> >
> > Sarge is very far from being released. If it is just a matter of
> > removing packages from testing, we must ask again for the removal
> > until it is done.
>
> Ok, please go ahead, i have filled the bug report two week ago, CCed
> to
> aj, written a follow up a few days ago, posted stuff about this on
> debian-devel, and no reply whatsover. If you want, please go ahead,
> and
> post to the bug report.
I will try to catch some ftp-master on irc in order to get the work
done.
> > > o We don't move and wait for other to fix the
> postgresql/libvorbis
> > > bugs. You notice that nobody has answered the postgresql call
> for
> > > help
> > > on debian-devel, so i don't suppose this bug will be fixed
> anytime
> > > soon.
> >
> > According to the BTS, ther aren't any bugs with libvorbis.
>
> A, no, just 105 packages which depend on it, and will be broken by
> libvorbis, in testing, and more than half of them are not ready to
> enter
> testing. i have a partial list somewhere, let me check ... Ok here it
> goes.
What the fix?
--
Jérôme Marant <jerome@marant.org>
<jerome.marant@free.fr>
http://marant.org
Reply to: