Open peti opened 7 years ago
Hey @peti !
No, definitely that shouldn't happen. I have seen this happening from time to time during different phases of the lifecycle of this library, and I think there is a concurrency bug lurking somewhere, but I haven't been able to squash it so far as it's very subtle and almost never manifest itself.
I guess what we need is much more visibility on which tests are running and what caused the deadlock, as in the current state the output from the test runner is not very useful.
The test suite was aborted on our Continuous Integration server because it ran for 2 hours without producing any output: http://hydra.cryp.to/build/1999502/nixlog/1/raw. I suppose that's not expected behavior?