On 01/11/2023 12:30 p.m., Henrik Bengtsson wrote:
I'm asking here to spare the CRAN Team a direct message, but also
because the answer is of interest to others:
Consider a package PkgA that was archived on CRAN, because it fails
checks and errors that were not corrected in time. At the moment when
package PkgA is archived, it will trigger automatic archiving of other
CRAN packages that has a hard dependency on it. Say, packages PkgB and
PkgC were archived automatically, because of their dependency on PkgA.
Question: If PkgA is at a later point revived on CRAN, will CRAN
unarchive PkgB and PkgC automatically? Or, should the maintainers of
PkgB and PkgC resubmit? If they have to resubmit, should they submit
identical versions and tarballs as before, or do they have to bump the
version?
I don't know the answer to your first question, but I'd guess "no". The
maintainers of PkgB and PkgC should have already resubmitted, demoting
PkgA to a "Suggests:" dependency or dropping it completely.
If they chose not to do that, then at the hypothetical time when PkgA is
revived, each of PkgB and PkgC would need full testing to see if they
were passing the current versions of the CRAN checks with the new
version of PkgA. That's something for their maintainers to do, not CRAN.
Regarding your last question: if it turns out the original version
still passes with the new PkgA, then I think it's still better to bump
the version number. It would need an exception to the usual "new
submissions need new version numbers" rule. Exceptions waste CRAN's time.
Duncan Murdoch
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
