Sure.
Was no bother. Just something that caught my attention when looking
around for connection management. Besides, you need so many requests
for it to throw a BufferUnderflowException, that it's very unlikely to
happen. One would probably restart your Linux server before this ever
happens.
Q
On 6/23/08, Oleg Kalnichevski <[EMAIL PROTECTED]> wrote:
> On Mon, 2008-06-23 at 20:13 +0200, Quintin Beukes wrote:
> > Hey,
> >
>
> > One can either modify the the method to either
> > (1) Implement a priority queue, which will give a perfect distribution.
> > (2) Modify the method to do this:
> > int i;
> > synchronized (this)
> > {
> > int i = this.currentWorker++ % this.workerCount;
> > if (i == 0) this.currentWorker = 1;
> > }
> > this.dispatchers[i].addChannel(entry);
> >
>
>
> I would not like to over-synchronize just to have a pretty worker count.
> and would just leave this bit of code as is. The worker count as well as
> the new channel distribution do not really need to be very precise.
>
> Can you live with that?
>
>
> Oleg
>
>
>
> > On 6/23/08, Oleg Kalnichevski <[EMAIL PROTECTED]> wrote:
> > > On Sun, 2008-06-22 at 23:04 +0200, Quintin Beukes wrote:
> > > > Hey,
> > > >
> > > > I was busy stepping the NIO code, and noticed the following:
> > > > this.dispatchers[this.currentWorker++ %
> this.workerCount].addChannel(entry);
> > > >
> > > > Is this safe? Since currentWorker will eventually wrap around to a
> > > > negative number. It will take a while, a long long while, but I don't
> > > > notice any resets anywhere, so I assume after running for a few
> > > > months, or under extremely high volume it will eventually fail.
> > > >
> > >
> > >
> > > I do not think it is unsafe given the fact we are only interested in
> > > more or less equal distribution of new connections among existing
> > > workers, but I agree it would be nice to have the count wrap around to
> > > 1.
> > >
> > > We happily take patches ;-)
> > >
> > >
> > > > PS: Of all the open source projects' source code I've worked with,
> > > > HttpComponents have been the biggest pleasure of all. The code is
> > > > very... "intelligable". Sometimes code is over "complexified", or made
> > > > so abstract that one cannot see how things link up with each other, or
> > > > just badly written. Thanks for it all!
> > > >
> > >
> > >
> > > I am glad you found source code comprehensible. I hope it is useful too.
> > >
> > > Cheers
> > >
> > >
> > > Oleg
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > For additional commands, e-mail: [EMAIL PROTECTED]
> > >
> > >
> >
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
--
Quintin Beukes
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
