I looked into why projpred started to fail package checks after my update to doFuture 1.0.0. From my troubleshooting, I do not believe this is related to the doFuture update. I also do not see how the actually updates in doFuture would be able to affect your back this way. Instead, I think this is related to recent updates to R itself that apply to the R 4.3.0 and R 4.4.0 branches.
Since I already investigated time in this, and followed the progression thus far, I figured sharing my notes could help you track this down on your end.
Right now, 'r-devel-linux-x86_64-debian-clang' does not fail. It runs R Under development (unstable) (2023-03-22 r84023). It fails on 'r-patched-linux-x86_64', which runs R version 4.3.0 alpha (2023-03-24 r84048), and on all other R-devel which run even newer revisions.
So, my best guess it that this due to some update in R between r84023 and r84048.
The following files have changed between r84023 and r84048;
Hello,
I looked into why projpred started to fail package checks after my update to doFuture 1.0.0. From my troubleshooting, I do not believe this is related to the doFuture update. I also do not see how the actually updates in doFuture would be able to affect your back this way. Instead, I think this is related to recent updates to R itself that apply to the R 4.3.0 and R 4.4.0 branches.
Since I already investigated time in this, and followed the progression thus far, I figured sharing my notes could help you track this down on your end.
Looking at https://cran.r-project.org/web/checks/check_results_projpred.html on 2023-03-27, it fails on R-devel (=R 4.4.0) and R-patched (=R 4.3.0), but not elsewhere. The error on 'r-release-windows-x86_64' is an installation error, so not a check error.
Right now, 'r-devel-linux-x86_64-debian-clang' does not fail. It runs R Under development (unstable) (2023-03-22 r84023). It fails on 'r-patched-linux-x86_64', which runs R version 4.3.0 alpha (2023-03-24 r84048), and on all other R-devel which run even newer revisions.
So, my best guess it that this due to some update in R between r84023 and r84048.
The following files have changed between r84023 and r84048;
If my hypothesis is correct, then the ERRORs should only show up on check servers running R (>= 4.3.0), but not on those running R (<= 4.2.3).