AFAIK, this is expected, on Windows (or some Windows versions), 
System.currentTimeMillis() only has a resolution of 15ms

http://stackoverflow.com/questions/7859019/system-currenttimemillis-is-not-accurate-on-windows-xp

Cheers,
Alex

> On 19.03.2015, at 09:11, Julian Reschke <[email protected]> wrote:
> 
> So,
> 
> as they plague the new Jenkins build and they also occur sometimes over here, 
> I spent some time with them...
> 
> 1) There was a single test for all of the clocks (simple, accurate, fast); I 
> have refactored these so that they execute separately.
> 
> 2) Most of the failures seem to be related to Clock.Fast. This is a clock 
> that relies on Clock.ACCURATE and a regular (10ms) update from that clock. At 
> least on my platform, I can see that if the updates occur, they actually 
> happen every 15ms (not 10ms).
> 
> Thoughts:
> 
> - we need to find out whether the actual code fails, or whether it's a 
> problem in the test case
> 
> - Relying on a scheduled thread to be executed in precise intervals seems 
> like asking for trouble
> 
> - As far as I can tell, the reason why we even have Clock.Fast is the 
> assumption that Clock.SIMPLE (== System.currentTimeMillis()) may be "slow". 
> Does anybody know what exactly is the problem? (Because, after all, the 
> simplest possible fix would be to get rid of this apparently fragile code)
> 
> Best regards, Julian

Reply via email to