Tim Peierls

In general, Failsafe users should __not__ expect timeouts to stop the execution of tasks. If you really want to ensure that tasks stop when they time out, you must carefully...

I think you're asking too much of Timeout. Timeout can't prevent a task (in general) from taking up an arbitrary amount of time, and tasks that take a lot of...

After code is missing method name, should be: ```java policy.handleIf((result, failure) -> failure.isServerError()); ```

`VarHandle` is Java 9+, and I would very much like to maintain Java 8 compatibility.

This seems like a very big ask. In general, Failsafe doesn't have access to the innards of the scheduler being used. Even if it did, you wouldn't want to saddle...

Clarifying: https://github.com/OpenFeign/feign

Amplifying what @whiskeysierra wrote: Making classes extensible is extremely hard to get right, and it's often at odds with the goal of a builder-based API. It can add maintenance costs...

I don't understand what "sync CompletionStage execution" is. Also, why is it important to implement sync retries "without spawning another thread"? If you're handed a CompletionStage, there's nothing you can...

Is there something undesirable about using sleep to implement the delay? It does tie up a thread, but there should only ever be one sleep happening at a time per...

And here's a way to do it with a scheduler, again adapting the toy example: ```java private final Scheduler scheduler; public long nextPrime(long n) throws InterruptedException, ExecutionException { return nextPrimeAsync(n).get();...