codecov[bot]
commented
5 months ago Codecov Report
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 90.90%. Comparing base (
799df38) to head (c26ad44).
Additional details and impacted files
```diff @@ Coverage Diff @@ ## master #638 +/- ## ======================================= Coverage 90.90% 90.90% ======================================= Files 128 128 Lines 8503 8503 ======================================= Hits 7730 7730 Misses 773 773 ```:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
For now all tests are commented, but maybe we should just delete them?
Tests fall into broadly 3 categories:
tests/testthat/test-append-debug.Rtests that appending 2 datasets with array variables with different aliases, but identically aliased sub variables fails. I think @aminusfu is saying that the new intentional behavior is to allow this because variables can belong to multiple arrays now.tests/testthat/test-append-subvariables.Rtests that appending 2 datasets, one with an array, and the other with unbound versions of the array's sub variables. It expects after appending for the unbound version of the variables to be only included in the arrays. However, now appends will allow 2 versions of the same subvariables to exist.tests/testthat/test-derive-debug.Rtests various things about changing the metadata of variables that are the source of derived array subvariables and making sure that the changes get to the derived version. These tests fail because they expect that the subvariable's alias will change when they are derived, so could be fixed by being flexible about the alias. However, it sounds like Mike was in favor of deleting tests that are covered elsewhere, and this seems like what he was thinking of.