Open drahnr opened 3 years ago
Unfortunately, I can't reproduce this locally (Ubuntu 20.10).
Can you try running sccache --stop-server
? The test suite should do it automatically, but maybe this helps...
Which SCCACHE_*
env vars are present in your environment? Can you reproduce this using some older commits, e.g. before std::future::Future
migration PR was merged?
I think the reason here is, that the local sccache server running via systemd restarts itself and hence blocks the port. I am not sure if any of those tests make much sense without containerization to avoid false negatives like this.
To solve that, it would be good to assign a random new port for each test, no?
rev
a70346110554f9f37601b9dcc1e19a3e71ede889
fails consistently for me with one test case: