Re: [ns] Incorrect event handling at ns scheduler stop?

[Haobo Yu <haoboy@packetdesign.com>]

I think so. I've checked it in and should be fixed in tomorrow's snapshot.

Thanks.

- Haobo

On Thu, 19 Oct 2000, Thomas Kaemer wrote:

> Hello!
> 
> I have a question about the event queue. I'm afraid an event is lost
> everytime the scheduler is stopped. If I understand the source code 
> correctly, there is a scheduler bug.
> 
> scheduler.cc:
> 
> | void
> | Scheduler::run()
> | {
> |       instance_ = this;
> |       Event *p;
> |       while ((p = deque()) && !halted_) {
> |               dispatch(p);
> |       }
> | }
> 
> 
> In the while loop header the next event is taken (p = deque()) 
> and then the halted_ variable is checked. If halted_ is equal 
> to 1 the scheduler stops its work and the dequeued event is not 
> handled.
> 
> Is my conclusion correct and can I exchange both expressions 
> in the while header without other side effects?
> 
> 
> Yours sincerely
>                  Kaemer
>