Closed havocp closed 9 years ago
This was caused by a short-circuit in the cancel case; when canceling, we removed a task from the queue and synthesized start/finish events for that task. I am removing that short-circuit and just always letting executions hit the ServerEngine, even if they are already canceled. This keeps events in order and reduces special cases.
From https://s3.amazonaws.com/archive.travis-ci.org/jobs/36180244/log.txt (which may get deleted), we start infiniteLoop then compile but we try to run infiniteLoop after compile, in CanCancelTasks integration test. Here is a piece of the log.