padpadpadpad
Windows CI R CMD CHECK failing
@padpadpadpad I figured it was best to pull together what we know about the R CMD CHECK failing on windows for rTPC v1.1.0.
As far as I can see the error comes when quickfit_tpc_multi() can't load rTPC, but it's unclear as to why.
My suspicion is we could possibly try adding the line everywhere(library(rTPC)) after the daemons(3) call to see if that gets it to push through. This shouldn't be the problem though, and is more useful when testing dev code (i.e. by running everywhere(devtools::load_all()) after the mirai daemons call).
That said I also noticed that the call to in_parallel() does not include quickfit_tpc_multi = rTPC::quickfit_tpc_multi as an arg. This is required by in_parallel (see the purrr documentation for details), so I am not sure if we're falling through the gaps into some strange undefined behaviour land, leading to unpredictable behaviour at the CI stage.
I think that my second point here will probably fix it, but because I can't easily reproduce the exact error locally, it's a bit hard to tell.
A final thought is we should probably encourage users to go for the oneliner if (!daemons_set()) daemons(3) to stop them having problems where an error has meant that the daemons(0) code later did not run.
