rgdal archived

[rsbivand]

@dankelley rgdal was archived yesterday, so please do move to submit an updated version of argoFloats as soon as possible. See https://github.com/r-spatial/evolution https://r-spatial.github.io/evolution/ for details.

[dankelley]

Thanks, I'll submit the github version to CRAN today or tomorrow. (This has been fixed in the github version for quite a while.)

[dankelley]

Step 1 is to ensure that the github version no longer uses the retired packages. I did that by noting that the following (on branch "develop") produced no output.

git grep -i maptools
git grep -i rgeos
git grep -i rgdal

[dankelley]

Step 2: run check_package.R (shown after the dots). Here are the links for some remote checks. I will tick them off when I get the results. If there are problematic notes (as opposed to problems with the setup of the remote machines, which is quite common for rhub), I'll make notes and fix them.

Also, if I see problems (via email) from the devtools checks, I will note them.

• [x] windows release R version 4.3.1 (2023-06-16 ucrt)
• [x] windows old-release R version 4.2.3 (2023-03-15 ucrt)
• [x] windows devel R Under development (unstable) (2023-10-16 r85339 ucrt)

If previous experience serves well, I ought to get results on all the tests by the end of today.

• [x] https://builder.r-hub.io/status/argoFloats_1.0.7.tar.gz-550faebef26b4006a5f70b97459fa283
• [x] https://builder.r-hub.io/status/argoFloats_1.0.7.tar.gz-0ea9a8e51c6c4b55a15ab0593e6cfb95
• [x] https://builder.r-hub.io/status/argoFloats_1.0.7.tar.gz-dd98563c2f364698b87dafa4d36c350a
• [x] https://builder.r-hub.io/status/argoFloats_1.0.7.tar.gz-9985f91aab1e41e8b0332ade3851e826 PREPERROR (a common result on rhub, which seems to tell little about the tested package

check_package.R

rhub <- !FALSE # rhub is VERY slow, and seems never to reveal problem not revealed otherwise

#install.packages("codemetar")
requireNamespace(c("codemetar", "devtools", "urlchecker", "rhub", "revdepcheck"))
# codemeta changes a timestamp, so requiring a commit after every call. That is
# senseless, so I only run the false part of the following conditional in the
# run-up to a release.
if (FALSE) {
    codemetar::write_codemeta()
} else {
    message("run 'codemetar::write_codemeta()' and then git push")
}
t <- devtools::spell_check()
stopifnot(t == "No spelling errors found.")
urlchecker::url_check()
devtools::check_win_release()
devtools::check_win_devel()
devtools::check_win_oldrelease()
if (rhub) {
    # rhub was broken in 2022 June/July but seen to work again on Aug 16
    rhub::check_for_cran(email="Dan.Kelley@Dal.Ca", show_status=FALSE)
    rhub::check(platform="debian-clang-devel", show_status=FALSE)
    #> rhub::platforms()
    #debian-clang-devel:
    #    Debian Linux, R-devel, clang, ISO-8859-15 locale
    #> rhub::check_rhub()
    # remotes::install_github("r-lib/revdepcheck")
}
revdepcheck::revdep_reset()
revdepcheck::revdep_check(num_workers=4)
message("run following if desired: pkgdown::build_site()")

[dankelley]

I've just submitted to CRAN, but I plan to leave the present issue open until I see that it has been accepted. That might take until tomorrow, I think.

[dankelley]

I'm closing this now, since all the CRAN-supplied binaries are up-to-date, and all but two of the remote checks are using the new version (without the warning). I don't see how the remaining two checks will fail, but in case they do, I plan to keep issue #598 open until I see the "all clear".

@rsbivand thanks for pointing this out. And, of course, many thanks for making the process of transition so very smooth for developers and users alike. Your patience and your generosity are an inspiration to a very large community. And that's to say nothing of your brilliant insights and skills. The world could use more like you!

[rsbivand]

@dankelley Thanks! Edzer and I did talk through the process not just as a technical exercise, but also as "performative" maintenance: 1) as maintainers get older with no team or lab need to continue maintenance, thinking about "clearing one's desk" shouldn't be put off, and 2) showing how publishing open source software has consequences, like responsibility for downstream users. We'll try to do a talk maybe at useR! 2024 or similar, I'm quite happy that of 300+ initially affected CRAN and Bioconductor package with strong dependencies on retiring r-spatial packages, less than 80 got archived for not updating in time, and some of these are still trickling in. All with their own reverse dependencies seem to be settled now, only one with my non-maintainer update. So it has gone much better than I thought it might.