I think I recorded this bug, but I was off playing my fiddle and didn't notice until too late. The test appeared to livelock or something like that and kept generating events until I ran out of disk space.
For unattended infra, this would be a bad problem. Similarly, if the test had deadlocked, it would have been nice for rr to time out after a while with an error message.
@rocallahan not sure how you feel about this for 1.0. My inclination is to punt.
I think I recorded this bug, but I was off playing my fiddle and didn't notice until too late. The test appeared to livelock or something like that and kept generating events until I ran out of disk space.
For unattended infra, this would be a bad problem. Similarly, if the test had deadlocked, it would have been nice for rr to time out after a while with an error message.
@rocallahan not sure how you feel about this for 1.0. My inclination is to punt.