OliverO2
OliverO2
Testing a new hypothesis for `GlobalTimeoutTest`: The 4 ms timeout was so short that one or both tests were interrupted before launch.
Isn't a fix of broken functionality always a breaking change? So in this case, I'd think this leads to the question: Could it make sense to use real time in...
There will be a couple of "hard real-time" tests, which would still fail occasionally with slow runners after merging this PR. I have another idea to address such cases: A...
@aSemy @sksamuel What do you think?
I agree that we have a change in behavior, but _only_ "for something that doesn't make sense". So Kotest would not be backwards bug-compatible. As I understand it, the suggested...
Maybe I have something that addresses the above concerns, yet is not too heavy for me to accommodate:
Merging this right away to make the entire CI train run again.
Wouldn't we need some way to have the CI generate updated API dumps on macOS if these had been previously created on a non-macOS host? (Emphasis mine) > When it...
If my understanding is correct: 1. The change towards macOS is only used for `apiCheck` in this PR. 2. With KLib API validation, macOS is the only target that can...
Maybe we could do an `apiDump` on macOS, but make that depend on a successful `apiCheck` on Linux. This way our dump would be always current for all operating systems...