Re: Bug#1054657: transition: r-bioc-biocgenerics
Hi Graham,
Am Wed, Nov 01, 2023 at 09:02:10AM -0100 schrieb Graham Inggs:
> > Sorry, my question was probably confusing. I was not talking about the
> > new packages. I was talking about the 170 r-bioc-* packages. If I
> > upload these to experimental, will it be necessary to upload these to
> > unstable again or can these be moved to unstable in one rush. Also
> > interesting in this connection: Will the tracker display the levels
> > of packages uploaded to experimental?
>
> I still don't think this has ever been possible.
Thanks for the clarification. In the last transition someone stated
this might be possible.
> > I'm comfortable with doing source-only uploads of packages that have
> > passed NEW. I'm not comfortable with uploading 170 packages twice -
> > once to experimmental and once again to unstable. Given that all
> > this work has mainly ended up on my shoulders I would prefer to
> > upload directly to unstable and simply bear with the waiting time
> > in NEW.
>
> Why do you want to upload 170 packages to experimental? We are only
> asking that the NEW packages involved be uploaded to experimental and
> clear NEW review before we start the transition.
The point is that we simply have no better means to know what new
packages might be required. We simply learn about new requirements by
building the new version of the r-bioc-* packages == in the process of
the transition. Thus someone suggested to do the transition completely
inside experimental. If there is no chance to move packages from
experimental to unstable without a fresh upload (which I doubted myself
- thanks for confirming) its in fact no option to build everything
in experimental first.
> > But you just mentioned the tracker in your previous mail[1].
>
> The tracker is visible [0], but is still in the 'Some planned
> transitions' section along with many others.
Ahhh, OK.
> > I admit I personally see the bigger drawback on spending my time twice
> > on 170 packages than waiting for new packages. Please explain (again)
> > the real drawback of a transition that was delayed due to waiting for
> > new packages in total by estimated (not verified) by about two weeks.
>
> >From what I saw in the last transition, r-bioc-biocgenerics was
> uploaded on 2023-07-17, and the last package (I think) to clear NEW
> was r-bioc-pfamanalyzer, which was accepted on 2023-08-15, almost one
> month after the transition started.
Maybe it has lasted for one month. I admit I personally see no drawback
in this time frame but possibly I'm to less involved of the work of the
release team to understand this. If it is a real problem I might
imagine to remove those (few, mostly not more than 5-10) leaf packages
that really need those new dependencies from testing so the majority of
the packages can migrate? I honestly try to understand this issue since
for the moment our only strategy to upgrade to new BioConductor is to
simply build the tree of dependencies and see what is happening.
> Again, we are not asking for the entire transition to happen in
> experimental. We are only asking for the NEW packages, so that NEW
> processing happens before the transition, and not during.
Understood this item now - but I'm lacking any clue how to find out
what new packages are needed.
Kind regards and thanks a lot for your patience
Andreas.
> [0] https://release.debian.org/transitions/
--
http://fam-tille.de
Reply to: