Misc fixes found when running package checks on newer releases of R

PecanProject / pecan

The Predictive Ecosystem Analyzer (PEcAn) is an integrated ecological bioinformatics toolbox.

www.pecanproject.org

Other

199 stars 231 forks source link

Misc fixes found when running package checks on newer releases of R #3237

Closed infotroph closed 4 months ago

infotroph commented 7 months ago

Description

Motivation and Context

Review Time Estimate

[ ] Immediately
[ ] Within one week
[ ] When possible

Types of changes

[ ] Bug fix (non-breaking change which fixes an issue)
[ ] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

[ ] My change requires a change to the documentation.
[ ] My name is in the list of CITATION.cff
[ ] I have updated the CHANGELOG.md.
[ ] I have updated the documentation accordingly.
[ ] I have read the CONTRIBUTING document.
[ ] I have added tests to cover my changes.
[ ] All new and existing tests passed.

infotroph commented 6 months ago

Probably need to merge #3244 first to resolve current CI failures

mdietze commented 5 months ago

@infotroph there are merge conflicts that need to be resolved in order to pull this one in

infotroph commented 5 months ago

Local debugging of the current check failures on 4.1 shows they're from dplyr refusing to load because the installed version of tidyselect was too old.

This was caused by the remotes::install_version(pkg, dependencies = TRUE, upgrade = FALSE) added to pecan.depends.R in #3244: Apparently "dependencies = TRUE" only installs dependencies that are not already installed, and does not upgrade ones with installed but unsuitable versions.

Options I see: switch to upgrade = TRUE and accept that will pull in large numbers of updates to old systems, or parse dependencies ourselves inside ensure_version to upgrade only as far as the package requires.

infotroph commented 4 months ago

All checks are passing, so I'll merge this today unless anyone speaks up before then.

Substantive changes since the last review:

Reorganized, resaved, and documented datasets in data.land, and made package data be lazy-loaded.
CI checks now ignore timestamps printed by newer versions of R CMD check
include PACKAGE argument in all .Fortran calls from RTM, fixing a test failure that only happens on R 4.2