🚀 go-pugleaf

Ok. The original functions will be used with only TCriticalSection (single
line). The reason for all the code is that it is needed for detecting the
deadlock. I want to raise an exception so that MadExcept can show the stack!
Hope you understand now. Sorry for the confusion.

I cannot create the mutexes at startup because there are countless
criticalsections. It must work within these two functions with named
mutexes.

Best Regards,

SZ

"Bob Gonder" <notbg@notmindspring.invalid>, iletisinde sunu yazdi, 
news:aev6339dmkoh5ih93feqeofs77v59f6iic@4ax.com...
> SubZero wrote:
>
>>which does not work for no obvious reason. It simply raises the exception
>>when there is no reason!
>
> What exception would that be?
>
> Why would you want to use _both_ mutex and CriticalSection ?
>
> I understand you want a timeout, so you use mutex, but that should be
> enough.
>
> I think it would be better to have a global handle for the mutex so
> you aren't constantly creating and opening the mutex (which you don't
> use CloseHandle on, so are leaking resources).
>
> Your lockCriticalSection() calls CreateMutex() with true, which is not
> correct for this. Think what happens if 2 threads (or applications)
> try to lock at the same time.  "Create with ownership" should fail for
> one of them.
>
> Furthermore:
> <help>
> The thread that owns a mutex can specify the same mutex in repeated
> wait function calls without blocking its execution. Typically, you
> would not wait repeatedly for the same mutex, but this mechanism
> prevents a thread from deadlocking itself while waiting for a mutex
> that it already owns. However, to release its ownership, the thread
> must call ReleaseMutex once for each time that the mutex satisfied a
> wait.
> </help>
> What this means is your CreateMutex(true) satisfies one "ownership",
> and the wait() satisfies a second, but you only release once.
>
> IMO you should CreateMutex(false) in application startup, with
> CloseHandle() at shutdown.
> Then the Lock() and and Release() would just use the global mutex
> handle.
> BTW, since your code should not reach releaseLock() without having
> aquired a lock, releaseLock() should not wait(), but simply
> ReleaseMutex().
>
>
>