Subj : Re: NPTL2 or something
To   : comp.programming.threads
From : Joseph Seigh
Date : Mon Jan 03 2005 08:00 am

On Mon, 03 Jan 2005 12:32:41 GMT, SenderX <xxx@xxx.com> wrote:

>> The eventcount itself has no destructor, just the usual explicit
>> destroy function.  No cancelation or exception handling.  It
>> shouldn't need it.
>
> So, waiting on an eventcount would not be a cancellation point?  
> Basically, I
> was wondering if your C++ stuff is dependant on the OS calling the
> destructors for stack-local objects upon thread cancellation. C++ and
> cancellation scared me into developing my library in assembly language  
> and
> plain C.

In Linux, and Unix if I ported to it, eventcount wait is a cancelation  
point
only because I made it one to make it consistent with Posix pthreads.   
Windows
doesn't have anything like pthread_cancel so I don't include one.  Anyway  
the
eventcount api for windows is different than the unix one, there's no point
in making them the same.

I'm staying away from C++ exceptions for now since any system that combines
synchronous and asynchronous exceptions displays a profound and fundamental
lack of common sense.

Joe Seigh

.