[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-hotspot-runtime-dev
Subject:    Thread.sleep() returns too early
From:       David.Holmes () Sun ! COM (David Holmes - Sun Microsystems)
Date:       2008-10-13 12:51:56
Message-ID: 48F3446C.5030005 () sun ! com
[Download RAW message or body]

Hi Artem,

There are some issues with Thread.sleep on Solaris but they involve a 
number of factors. I'm presently investigating an issue where the system 
time changes and then a signal is sent to the process and is apparently 
handled by the sleeping thread which then returns immediately from sleep.

Is your problem readily reproducible? (The test behaved somewhat 
erratically when I tried it, but there was no timeout message.)

Cheers,
David Holmes

On 13/10/08 09:59 PM, Artem Ananiev wrote:
> Hi, HotSpot team,
> 
> I have faced a problem with Thread.sleep(), which looks very weird from 
> my perspective. The test is attached. It first forks a new thread and 
> then performs some actions on the main and EDT threads. The first 
> (forked) thread is a kind of timeout/hang/deadlock prevention stuff and 
> contains a single sleep(600000) call, but I found it may return from it 
> at any arbitrary moment without any exceptions, causing the test to fail 
> with "timeout".
> 
> This only happens on Solaris. Any ideas?
> 
> Thanks,
> 
> Artem
> 


[prev in list] [next in list] [prev in thread] [next in thread]