Closed brownag closed 8 months ago
The only thing I can think of is that the files in R/ are sourced in alphabetical order. But no idea why sourcing that function later would fix anything!
I suppose you could check if putting it zzzz.R also "fixes" things to test that hypothesis.
Ok, did a little testing and I feel like this has to be a bug in covr
. If you just create a utils.R file with literally any comment in it, this also fixes the issue. I'm so confused.
Ok, did a little testing and I feel like this has to be a bug in
covr
.
I suspect you are right, and I think the other issue with the evaluation environment is also likely a bug
Some strange {covr} bugs here! Thanks for taking the time to look into this strangeness!
So I don't quite understand this - it was passing all checks here? Then I merged it and it broke? What in the world is happening.
The new error is a different one actually... it also is occurring on #33
the write() function in tar_format() must not create a directory. Found directories inside the data store where there should only be files: _targets/objects/test_terra_vect
This indicates that the "shz" special handling from #21 is writing a regular shapefile or something similar
From https://github.com/njtierney/geotargets/actions/runs/8305390948/job/22732222126
This seems to get our codecov action working again, for #31, but I am not satisfied with this as a general solution. I do not understand at this time why it would matter if there was code in utils.R or not.
There seems to be some pretty odd stuff that can happen when target tests are being run inside testthat via covr.