Open jmalloc opened 4 years ago
Ok, I've found at least one of these lines. I doubt it's the only one; I'll start a list.
Reopened, updated list.
Ok, something very strange is going on. The builds finish and Codecov reports a coverage failure because some coverage has dropped. But then they change the status to passing?
Many of the tests have been refactored, and this issue does not appear to be happening any longer.
I can't win with this one! This line is flapping now:
There's some time-sensitive test somewhere (I suspect in the
pipeline/queue
package) that's taking a different code-branch some of the time and causing coverage to flap about.This manifests as failed codecov reports for PRs that have 100% diff coverage, like #187.