GitHub Actions Check failing due to example only on Mac

Hi,

Hoping someone might have thoughts/suggestions on weird issue I'm running into. I use GitHub actions as part of my release checklist for my R package which is on CRAN. It currently runs the checks for PRs on:
macos-latest (release)
windows-latest (release)
ubuntu-latest (devel)
ubuntu-latest (release)
ubuntu-latest (oldrel-1)

I've never really run into any issues. However, when running checks on my newest version update I'm running into weird issue (PR can be found here: Release/2.0.0 by samuel-marsh · Pull Request #138 · samuel-marsh/scCustomize · GitHub). All of the checks pass except for macos-latest (release). When I go through the log it says the error comes from one of the function examples in documentation.

This seems to be weird error because this error doesn't occur on any of the other platforms (nor locally on my mac) and it is not a new function/example and has previously passed checks on macos.

The other weird thing is that this function is very simple under the hood, here is link to it: in current CRAN version: https://github.com/samuel-marsh/scCustomize/blob/2bb2b138edefee5946c523a5d1b2e418974a5d39/R/Object_Utilities.R#L623-L726. So that kinda of further confuses me as to why it's failing and only failing on single platform.

Thanks!
Sam

EDIT: I should note that the obvious solution is to just change the example not to run but hoping to avoid that.

Have you tried rerunning the failed job? The error was raised in the example but actually occurs in dplyr::slice() so there's probably something with the input data (which seems to be a dataset from another package). It may not be recommended to use datasets from other packages in examples because it makes examples somewhat fragile/prone to errors if the data changes or can't be loaded. I'd first rerun the job and if the error still occurs, I'd try with an example data created locally.

This topic was automatically closed after 45 days. New replies are no longer allowed.


If you have a query related to it or one of the replies, start a new topic and refer back with a link.