“Maxim S. Shatskih” wrote in message
news:xxxxx@ntdev…
> Much better approach is to always maintain ReadEvent as signaled when
> there is some data in the buffer, and WriteEvent signaled when there is
> free space in the buffer.
>
> No deadlocks at all.
>
Thanks all for ‘fixing’ my code but the objective here was to invent a
scenario that would break under the new contract. The good thing about
SetEvent on an auto reset event is that it will always satisfy one wait.
Even if there are currently no waiting threads available it will not reset
until one wait is satisfied in the future, and thus also solves the problem
with waiting threads losing their wait states if they are executing APCs.
//Daniel
James Harper wrote:
I love the English language and even if the word ‘affinitize’ had never
ever been used before (it had, based on your web search) I knew exactly
what was meant by it and it does indeed seem like a great space saver…
if someone had come to you and said “you can’t use the word affinitize
here” what would you have replaced it with?
To me, the obvious replacement is “manage the affinity of”.
–
Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.