Closed fweber144 closed 1 year ago
Thanks @renozao for your maintenance of doRNG thus far. Based on an email (below) from Brian Ripley, it appears that the package needs a maintainer by January 19, 2023.
Packages
ADAPTS BSL Counterfactual EMAS EPX EnsembleBase EnsembleCV EnsemblePCReg EnsemblePenReg IFAA KarsTS LongDat MERO Mediana MedianaDesigner NADIA ORdensity PRSPGx RapidoPGS Rnmr1D RobGARCHBoot Robyn SGP SparseMDC TRexSelector WebGestaltR adjustedCurves autoMrP bartMachine bestNormalize bhmbasket bigsnpr cbl deltaccd ellipticalsymmetry envi ewoc exuber fect funGp gbts gratis gsynth highMLR hwep hybridModels imp4p kernelshap latentcor lcc locStra longit markovMSM metaplus mi4p midasml missCompare missForest mixedCCA momentuHMM mplot obliqueRSF promor randomLCA redist robmixglm robust2sls scMappR scpi seqimpute simputation skpr sparrpowR speaq survidm updog varclust
require package doRNG directly or indirectly and it has now been orphaned (the 'maintainer' having failed to responds several times over months).
So the CRAN policy
"Orphaned CRAN packages should not be strict requirements (in the ‘Depends’, ‘Imports’ or ‘LinkingTo’ fields, including indirectly). They are allowed in ‘Suggests’ if used conditionally, although this is discouraged."
now applies. Please adjust your packages to comply by 2023-01-19.
If one of you wants to take over as maintainer please do so (but do not discuss this on the CRAN list -- we have quite enough to read).
I guess all the packages depending on this one will get removed from CRAN by then, right?
My intention is to bring back pkgmaker up to CRAN on time so that doRNG and other packages can still happily live together on CRAN.
Fantastic! Thanks
Wonderful, @renozao. Thank you.
I guess all the packages depending on this one will get removed from CRAN by then, right?
FWIW, if so, I think it's only hard dependencies that are affected. At least in the first round.
For example, doFuture depends on doRNG optionally via Suggests:
and I did not get a notification, doFuture is not mentioned in the above CRAN email. The current orphaned state triggers a:
Check: package dependencies
Result: NOTE
Suggests orphaned package: ‘doRNG’
e.g. https://cran.r-project.org/web/checks/check_results_doFuture.html.
@renozao Do you perhaps have any updates on this? I do like the doRNG package a lot and would love to keep using it in my package, but if the "orphaned issue" is not fixed until the 18th of January I will be forced to stop using it. Otherwise, CRAN will remove my package as well.
Hi all,
I have submitted a new version of pkgmaker
to CRAN just now.
It is all green on my side: https://github.com/renozao/pkgmaker/actions/runs/3876288108
I have submitted a new version of pkgmaker to CRAN just now.
Great!
Maybe you already know, but I think you're also gonna have to resubmit doRNG to reclaim ownership and take it out of the ORPHANED state.
Thanks for pointing this out @HenrikBengtsson . Will prepare a submission for doRNG then, to have ready when pkgmaker makes it back to CRAN.
Update: pkgmaker
is back on CRAN and I submitted an update for doRNG just now. Awaiting for checks.
Side question: @HenrikBengtsson would you know how to setup r-lib github-actions correctly for Mac OS and Windows so that package doMPI
gets installed properly? I tried a couple ways with pacman
and homebrew
without success.
I thought that you may have experience on that for future
and doFuture
.
Thanks!
That's good news
Side question: @HenrikBengtsson would you know how to setup r-lib github-actions correctly for Mac OS and Windows so that package
doMPI
gets installed properly? I tried a couple ways withpacman
andhomebrew
without success. I thought that you may have experience on that forfuture
anddoFuture
.
Sorry, I don't. I don't have access to macOS, and haven't tried on MS Windows, but I know the underlying Rmpi is a pain to install on Linux. It does not install out of the box on Linux, and honestly, I don't know why it's not configured to do so (or at least so you only have to specify which MPI type to use - now you have to specify paths too). Linux instruction can be found at https://www.c4.ucsf.edu/howto/r.html#the-rmpi-package. Maybe something similar can be done for macOS?(!?)
Thanks @HenrikBengtsson. I managed to get the macOS check to work. I used:
- name: MacOS binary dependencies
if: matrix.config.os == 'macos-latest'
run: |
brew install gcc
brew install open-mpi
brew install texlive
Installation on Linux is not too bad, since it is clearly stated which apt package to install. Only missing piece is Windows. Leaving it for later.
See: https://github.com/renozao/doRNG/actions/runs/3911749007/jobs/6685493826
Installation on Linux is not too bad, since it is clearly stated which apt package to install.
I just tried on Ubuntu 20.04, and there Rmpi does indeed install out of the box, so I guess it's just RedHat/CentOS one has to go an extra mile.
Update on this issue: a new version of doRNG is now on CRAN (version 1.8.6). I believe this solves the orphaned dependency issue and you should not have to do anything on your side. Thank you!
@HenrikBengtsson would you know of an easy way to contact all the maintainers that received the email about doRNG being orphaned?
@HenrikBengtsson would you know of an easy way to contact all the maintainers that received the email about doRNG being orphaned?
Sorry, I don't. But, I think the Tidyverse folks reach out to revdep maintainers once in a while. Maybe there's something in devtools or revdepcheck.
It seems like the CRAN package has been updated by the CRAN team to v1.8.3, with the updated
NEWS
file saying:I'm not sure whether this means that the package might get archived soon. In any case, I thought it might be better to let you know (especially because there was #16 in the past and it seems like you took some action back then). It would be great to keep doRNG on CRAN. It's a package I use a lot, thanks for developing it.