RFR: 8178698: javax/sound/midi/Sequencer/MetaCallback.java failed with timeout

[Phil Race]

On Tue, 15 Nov 2022 07:31:46 GMT, Alexander Zuev <kizune at openjdk.org> wrote:

> On m1 mac sometimes (in very rare instance) it takes a long time to get the sequencer so increasing timeout helps with that.
> Also on m1 mac sequencer got notified about the first MetaMessage twice. The reason is unclear but watching for that fixes the problem. Also on Linux sometimes process hangs indefinitely without finishing hold by the com.sun.media.sound.MediaDispatcher. Closing the sequencer at the end of the test helps with that.

test/jdk/javax/sound/midi/Sequencer/MetaCallback.java line 43:

> 41:  * @summary MIDI MetaMessage callback inconsistent
> 42:  * @key intermittent sound
> 43:  * @run main/othervm/timeout=300 MetaCallback

"On m1 mac sometimes (in very rare instance) it takes a long time to get the sequencer so increasing timeout helps with that."

Hmm. 5 minutes is a really long time for something that should take less than 1 second (getting the sequencer).
Is that because something else (perhaps another process) is holding a needed resource ?

Are any other tests running ? Or were you running this test in a loop ? Perhaps it is something like what you say happened on Linux where the test wouldn't exit until the sequencer was closed ? Suppose some other process managed to exist without properly releasing some system resource .. and the next app that needed it has to wait until the O/S cleaned up?
Is that possible ?
Our CI test harness automatically multiplies test timeout by 2  so you've now in effect got a VERY long 10 minute timeout on this test .. way too long I think.

test/jdk/javax/sound/midi/Sequencer/MetaCallback.java line 69:

> 67:     // On M1 Mac sometimes system notifies listener about the same message twice
> 68:     static List<MetaMessage> received = Collections.synchronizedList(new ArrayList<>());
> 69: 

Is this our bug ?

test/jdk/javax/sound/midi/Sequencer/MetaCallback.java line 112:

> 110:         sequencer.stop();
> 111:         sequencer.close();
> 112:     }

I can believe this was a problem.
IIRC on x64 mac I had some program block until I closed the sequencer.

Can you check other jtreg sound tests for similar potential problems

-------------

PR: https://git.openjdk.org/jdk/pull/11157