[boost] Re: [network] An RFC - updated

From: Don G (dongryphon_at_[hidden])
Date: 2005-04-24 16:33:45

Hi Peter,

> FD_SETSIZE only applies to select, and nobody
> uses select except in "quick and dirty" servers
> (as I did recently).

Sadly, select is by far the most portable solution. On Windows, it is
the only solution for 9x/Me. We can and should write more efficient
impls for platforms that support poll() or epoll() or kqueue, etc..

> The idea of an explicit poll() call is to allow
> the implementation to use a single-threaded
> dispatch mechanism (select/epoll/kqueue), if one
> is available. Without poll() you are forced to
> use worker thread(s) and a callback queue to
> support a single-threaded application, and this
> can be considerably less efficient than
> epoll/kqueue (or select in the < FD_SETSIZE 
> case) (in theory).

Yes, indeed. Though I may be missing something in your idea about
net::poll(). How would it avoid a queue?

> Of course, if the threaded model can be
> demonstrated to achieve comparable performance,
> there would be no need to complicate the
> interface with poll().

The "need" is performance and familiarity (IMO). If performance were
taken out of the picture, it would be hard to argue for it. I don't
know how the comparison would go which is why I was writing some
stuff to try it out and get some measurements.

Best,
Don

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com