[D-runtime] Why does the OSX semaphore test keep failing?

[Sean Kelly]

On May 9, 2011, at 10:06 AM, Brad Roberts wrote:

> On 5/9/2011 8:11 AM, Don Clugston wrote:
>> It's been doing it for a month or so. Every third or fourth run, it
>> fails. Is the test not lenient enough? Or is there a genuine problem?
> 
> Not a good explanation, but a bit more history:  It's failed intermittently approximately forever.  It used to be worse
> but Sean rewrote the test itself and it got a lot better.  Obviously it needs to not fail at all.  My guess is that it's
> still a test related failure, but quite possibly a bad assumption.

The test used to be timing-related, so success was unpredictable.  I rewrote it not be timing-related and don't know why it's failing now.  I'll have to look at it.