Re: [boost] condition Win32 based on events

From: Roland Schwarz (roland.schwarz_at_[hidden])
Date: 2005-09-12 10:53:05

Alexander Terekhov schrieb:

>Your locking is way too strict (with two waiters on CV, you're blocking 
>the second signal across context switch).
>
I already was suspecting something along this lines.
Thank you for having it pointed out.

> Your pthread_cond_destroy()
>doesn't conform to pthreads (you need to lock the gate sema to 
>synchronize destruction with respect to exiting waiters).
>
Yes I know this. But the code sketchy at best yet. Do you think there 
would be
a principal problem? I won't try this until I am confident the main idea is
worth extending.

> Uhmm, and 
>what about timed waits and/or cancellation?
>  
>
I planned to extend to use WaitForMultipleObjectEx with timeout != 
INFINITE and
APC's to cancel it. Reasonably?

Regards,
Roland