Closed olivroy closed 1 month ago
Thanks @olivroy ! Can you rerun?
You can accept workflow run to see if everything works! I will also run locally
@gmyenni I can confirm that everything passes cleanly on my end locally. Can't reproduce the failures encountered on CI.
@gmyenni I just noticed that you weren't testing on macOS before. I will revert that change. I still don't know what is causing the R-devel problem, but it is unrelated to this PR.
For the R-devel checks to pass, I suspect we need 049f67c9c963c28228dd8b20508860788444c2ed and 761423b3df17fc70c510bf7e3cfbf650a77b0f29 from #302 to be added
I will merge these changes and we solve the na.strings = c("", "NA"), in a new PR
@henrykironde there may be one other action required from you regarding the test-coverage action. codecov now requires a token to be set. You can set a repo or organization token by following the instructions here
~Tests won't pass because of #298~
Fixes #295
Otherwise, use cli for messaging
use testthat 3rd edition (use snapshot test for message validation + other lints)
Update gh actions with
usethis::use_github_action()
Update pkgdown to use bootstrap 5 (adds a search bar)
FIx a tidyselect deprecation warning
Make format_*() functions redundant (no longer used in the package)
Happy to answer any questions you may have!
cli has the advantage of detecting automatically if it should do ansi output and fallbacks otherwise. This means that if you are using RStudio (or a terminal that supports links), test output is now clickable (as you can see, I was able to get rid of setup.R and teardown.R
R CMD CHECK passes cleanly locally ~except for the test failure~