Hm, I found one problem with both my reimplementations, and this problem
existed in the original implementation, too: there's a race at the point
where the associated mutex is released and before the thread starts
waiting for the trigger event. IIRC, POSIX promises this will be atomic.
I'm not sure
Branko Äibej wrote:
>I've been trying to use the APR condition variables on Win32, and the
>dam' things kept deadlocking on me whatever I did. So, out of sheer
>desperation, I took a look at the implementation... well, let's say that
>I was awed by the number of bugs and race conditions I found in
I've been trying to use the APR condition variables on Win32, and the
dam' things kept deadlocking on me whatever I did. So, out of sheer
desperation, I took a look at the implementation... well, let's say that
I was awed by the number of bugs and race conditions I found in there. :-)
Anyway: I we