RxIdler throws IllegalStateException when buffer is used

[audriusKV]

When RxJava2 operator buffer is used like this:

RxJavaPlugins.setInitIoSchedulerHandler(Rx2Idler.create("Test scheduler"));
PublishSubject subject = PublishSubject.create();

subject
        .buffer(1, TimeUnit.SECONDS, Schedulers.io())
        .subscribe();

RxIdler throws IllegalStateException:

java.lang.IllegalStateException: Already completed
  at com.squareup.rx2.idler.DelegatingIdlingResourceScheduler$ScheduledWork.run(DelegatingIdlingResourceScheduler.java:152)
  at io.reactivex.Scheduler$Worker$PeriodicTask.run(Scheduler.java:371)
  at io.reactivex.internal.schedulers.ScheduledRunnable.run(ScheduledRunnable.java:61)
  at io.reactivex.internal.schedulers.ScheduledRunnable.call(ScheduledRunnable.java:52)
  at java.util.concurrent.FutureTask.run(FutureTask.java:266)
  at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:301)
  at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162)
  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636)
  at java.lang.Thread.run(Thread.java:764)

If this actual RxIdler issue I can provide project to reproduce this.

[JakeWharton]

I'm about to head on vacation. I'll try to repro at some point when I have downtime though.

On Thu, Jul 20, 2017 at 6:13 AM Audrius Karosevicius < notifications@github.com> wrote:

When RxJava2 operator buffer is used like this:

RxJavaPlugins.setInitIoSchedulerHandler(Rx2Idler.create("Test scheduler")); PublishSubject subject = PublishSubject.create();

subject .buffer(1, TimeUnit.SECONDS, Schedulers.io()) .subscribe();

RxIdler throws IllegalStateException:

java.lang.IllegalStateException: Already completed at com.squareup.rx2.idler.DelegatingIdlingResourceScheduler$ScheduledWork.run(DelegatingIdlingResourceScheduler.java:152) at io.reactivex.Scheduler$Worker$PeriodicTask.run(Scheduler.java:371) at io.reactivex.internal.schedulers.ScheduledRunnable.run(ScheduledRunnable.java:61) at io.reactivex.internal.schedulers.ScheduledRunnable.call(ScheduledRunnable.java:52) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:301) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636) at java.lang.Thread.run(Thread.java:764)

If this actual RxIdler issue I can provide project to reproduce this.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/square/RxIdler/issues/6, or mute the thread https://github.com/notifications/unsubscribe-auth/AAEEEfU4YtgUPCKLxu2OzntOVs28B4aTks5sPyhEgaJpZM4Od4pl .

[nhaarman]

The same goes for Observable.interval, but only if initialDelay is set to 0:

Observable.interval(0, 1, SECONDS, Schedulers.computation())
              .subscribe()

W: java.lang.IllegalStateException: Already completed
W:     at com.myapp.DelegatingIdlingResourceScheduler$ScheduledWork.run(DelegatingIdlingResourceScheduler.java:152)
W:     at io.reactivex.Scheduler$Worker$PeriodicTask.run(Scheduler.java:371)
W:     at io.reactivex.internal.schedulers.ScheduledRunnable.run(ScheduledRunnable.java:61)
W:     at io.reactivex.internal.schedulers.ScheduledRunnable.call(ScheduledRunnable.java:52)
W:     at java.util.concurrent.FutureTask.run(FutureTask.java:266)
W:     at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:301)
W:     at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162)
W:     at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636)
W:     at java.lang.Thread.run(Thread.java:764)

[JakeWharton]

The fix is trivial but the implications are very complicated so I don't want to make it blindly. Need more time to investigate.

[henorek]

@JakeWharton Is there a chance that this will be fixed soon?

[JakeWharton]

No one is working on it at this time.

On Wed, Aug 23, 2017 at 6:46 AM Jarek Jankowski notifications@github.com wrote:

@JakeWharton https://github.com/jakewharton Is there a chance that this will be fixed soon?

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/square/RxIdler/issues/6#issuecomment-324292475, or mute the thread https://github.com/notifications/unsubscribe-auth/AAEEEYVW_QgIlKHbtb9nLR4eb61-wuMvks5sbAMbgaJpZM4Od4pl .

[jonasfa]

Is there any known workaround? @JakeWharton perhaps if you explain what the issue is, somebody will work on it?

[JakeWharton]

The same state machine instance gets re-scheduled instead of the underlying Runnable being wrapped in a new one. And since we only allow forward progress in the state machine, once you execute once you cannot go backwards to being un-executed again when this happens.

On Mon, Sep 11, 2017 at 10:02 AM Jonas Alves notifications@github.com wrote:

Is there any known workaround? @JakeWharton https://github.com/jakewharton perhaps if you explain what the issue is, somebody will work on it?

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/square/RxIdler/issues/6#issuecomment-328538676, or mute the thread https://github.com/notifications/unsubscribe-auth/AAEEEXw8SXR9UZssVlgs7hbew_d1IAMNks5shT1wgaJpZM4Od4pl .

[jonasfa]

@JakeWharton thanks. You also mentioned complicated implications for a possible fix. Do you remember what was that?

[JakeWharton]

The obvious fix is to allow the state machine to go backwards, which it wasn't designed to do, so making that trivial change probably has other effects.

On Mon, Sep 11, 2017 at 10:36 AM Jonas Alves notifications@github.com wrote:

@JakeWharton https://github.com/jakewharton thanks. You also mentioned complicated implications for a possible fix. Do you remember what was that?

— You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub https://github.com/square/RxIdler/issues/6#issuecomment-328549298, or mute the thread https://github.com/notifications/unsubscribe-auth/AAEEEQ_F2o8fTH5oTX4QKOpTvA0RkOeQks5shUVvgaJpZM4Od4pl .

[Maragues]

@nhaarman I'm seeing this issue with or without initialDelay.

In the case of interval, the first tick is emitted, but it never goes beyond that. It does print the doAfterNext statement.

In the meantime, I'm using a JUnit rule that I've used in the past, which brings its own issues but at least it doesn't crash in my scenario

You can also check https://github.com/ReactiveX/RxAndroid/issues/149

[philipbjorge]

Seeing this for sample as well.

[athkalia]

Any updates on this?

[mseroczynski]

Any updates?

[niemandkun]

Any updates?

[larryng]

Made a PR that I believe addresses this issue: https://github.com/square/RxIdler/pull/14

Makes the state machine go backwards for periodic work, but AFAICT it should be safe.