Timeout may never trigger when execution starts after the timeout has already expired

[timothybasanov]

Reproduction:

1. Create a thread pool with one thread and infinite queue
2. Build FailsafeExecutor that has RetryPolicy and Timeout
3. Schedule three tasks (getAsync) where each is slightly faster than the timeout

Execution order (with timeout of 10s and task length 9s):

• 00s all three executions are scheduled
• 00s first task starts execution, others are in the queue
• 09s first task finishes the execution
• 09s second task starts execution
• 10s second task is interrupted due to timeout
• 10s third task is interrupted due to timeout
• 10s second task attempt 2 starts execution
• 19s second task succeeds
• Third task never starts execution.

As far as I can see if the future returned from net.jodah.failsafe.Functions#getPromiseAsync cancelled before its execution starts then supplier is never called and promise is never completed.

This is somewhat similar to #257, but does not require thread pool and its queue to overflow. This may require a special scheduler (e.g. recommended in that issue that executes timeoutFuture directly on the ScheduledExecutorPool.) With default scheduler timeouts do not start until the actual task execution starts, so there are no timeouts for all three tasks.

Reproducible on 2.4.0.

This is not a contribution

[jhalterman]

It's been a while on this issue, apologies. I tried reproducing your scenario against 2.4.0 and I couldn't exactly working as you described, though clearly with a single thread Timeouts don't fire as expected. Here's the test case I added based on your description: https://github.com/failsafe-lib/failsafe/blob/master/src/test/java/net/jodah/failsafe/issues/Issue260Test.java

This issue is now resolved in master. Let me know if the test case looks off or if I missed something.