[R-pkg-devel] Rolling Back an Archived Package

[Lluís Revilla]

Dear Eric,

While changing the version number of an old package could solve this, it is
a work-around that won't help much:
Users will be able to install the binaries of an old package without the
new fixes and improvements of the later at the cost of another submission,
which might delay the submission which fixes the current issues.

Recently with Henrik Bengtsson we prepared a fallback for these cases:
Packages archived from CRAN are built for 35 days more to be able to
provide binaries to users: https://www.cranhaven.org/dashboard-live.html
This time is what it takes for half the archived packages to be back to
CRAN.
You could point/use that repository while you fix your package up to CRAN
checks.

Is the package archived Colossus?
Maybe some users might be able to help, as this looks like a compiler
specific issue on Fedora.
I am not sure if rhub can help you check your package in this environment,
but that is something you could explore too.

Best wishes,

Lluís


On Mon, 23 Sept 2024 at 22:53, Ben Bolker <bbolker using gmail.com> wrote:

>  I can imagine a (perhaps overly complicated) scenario:
>
> * version 1 passes tests OK when it is submitted to CRAN
> * version 2 passes tests OK when it is submitted to CRAN
> * CRAN introduces new tests (compiler versions, etc. etc.) that break
> version 2 but not version 1.
>
> Now version 1 would still pass but version 2 would fail, so bumping
> the version number of version 1 to 3 and resubmitting could (??) work.
>
> On Mon, Sep 23, 2024 at 4:24 PM Duncan Murdoch <murdoch.duncan using gmail.com>
> wrote:
> >
> > On 2024-09-23 4:13 p.m., Ben Bolker wrote:
> > > Hmmm. Assuming the previous version of the package (without the
> > > extensions/updates) avoids all of the testing errors, I would say that
> > > RAW ("rules as written") the only constraint I can see on submitting a
> > > "rollback" version of the package would be the CRAN request for
> > > updates “no more than every 1–2 months".
> >
> > I think the original package wouldn't have been archived unless it was
> > failing tests and the author didn't address the failures.
> >
> > Duncan Murdoch
> >
> > > https://cran.r-project.org/web/packages/policies.html   But an e-mail
> > > to the CRAN maintainers (if none reply on-list) seems appropriate.
> > >
> > > On Mon, Sep 23, 2024 at 3:49 PM Eric Giunta <egiunta using ksu.edu> wrote:
> > >>
> > >> Hello all,
> > >>
> > >> Recently I submitted a large update to the package I maintain and was
> unable to resolve the testing errors prior to it being archived. I've been
> unable to reproduce the errors, so I expect to have to setup my own
> fedora_clang virtual machine to debug my package. Ideally I'd want a
> previous version to be more easily available while I figure out my issues.
> Is it against CRAN policy to resubmit an earlier version of an archived
> package, assuming I fully explain what led to the archival and
> re-submission in the submission comment? I'm sorry if this has already been
> asked, I couldn't find an answer online and wanted to check before
> resubmitting anything.
> > >>
> > >> Thank you for your time,
> > >> Eric
> > >> ______________________________________________
> > >> R-package-devel using r-project.org mailing list
> > >> https://stat.ethz.ch/mailman/listinfo/r-package-devel
> > >
> > > ______________________________________________
> > > R-package-devel using r-project.org mailing list
> > > https://stat.ethz.ch/mailman/listinfo/r-package-devel
> >
>
> ______________________________________________
> R-package-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-package-devel
>

	[[alternative HTML version deleted]]