DistributedPhysicalActionUpstreamLong.lf times out in mac CI

[erlingrj]

We are having problems with DistributedPhysicalActionUpstreamLong.lf timing out in CI, e.g. here. I have also seen it in another unrelated PR, so it is not about the changes introduced there.

I am not sure what could have introduced this. A hint might be that @petervdonovan merged in #2249 , however it should not affect this because that was Docker-specific.

[petervdonovan]

Unless the way timeouts work in our test framework has changed, the CI run that you linked looks like a user cancellation, not a timeout.

[lhstrh]

Timeout is set to 120 mins

[erlingrj]

I reran with logging enabled, and now it seems to have timed out on a different test: https://github.com/lf-lang/reactor-c/actions/runs/8546664692/job/23459617885?pr=400

[lhstrh]

There is a history of the compiler getting stuck on macOS. We don't know why this happens and it doesn't seem to be reproducible. It was assumed that the culprit was here:

lingua-franca/core/src/main/java/org/lflang/federated/generator/FedGenerator.java

Line 339 in 205ed22

6, Math.min(Math.max(federates.size(), 1), Runtime.getRuntime().availableProcessors()));

But in hindsight, I don't think we ever actually solved the problem...