From: "Mauricio Fernández" <batsman.geo@...>
Date: 2003-02-08T18:50:38+09:00
Subject: Re: drb and signals

On Sat, Feb 08, 2003 at 02:57:35AM +0900, ahoward wrote:
> On Sat, 8 Feb 2003, Eric Hodel wrote:
> 
> > ah-ha!  You really want #synchronize, it does the waiting for you, and
> > its not a busy-wait.
> >
> > class DistributedObject
> >   def method
> >     sem.synchronize { yield }
> >   end
> > end
> 
> not in my case actually :  i have an Array of PGconns, iff i cannot obtain the
> lock on _any_ of them, i must go to sleep on a ConditionVarible and am woken
> up appropriately (when a connection becomes free).  eg.  there is a Mutex for
> each connection AND a Mutex on the array of connections.  the threads which
> release the lock on a particular connection also notify the condition
> variable.  in that way, i DO use the non-busy-wait feature of sem.synchronize
> _but_ this is wrapped around the entire array of connections.  now, if i used
> sem.synchronize on conns[0], for example, conn[1] would never be used!  eg. i
> cannot _automatically_ go to sleep just because a _particular_ conn is not
> available.  i only should go to sleep in _all_ connections are not available.
> does that make any sense?

Might not be completely pertinent, but have you seen 
http://www.rubygarden.org/ruby?ObjectPoolingAndThreading ?

I think it can be made to work the way you want and it doesn't use
busy-wait.

-- 
 _           _                             
| |__   __ _| |_ ___ _ __ ___   __ _ _ __  
| '_ \ / _` | __/ __| '_ ` _ \ / _` | '_ \ 
| |_) | (_| | |_\__ \ | | | | | (_| | | | |
|_.__/ \__,_|\__|___/_| |_| |_|\__,_|_| |_|
	Running Debian GNU/Linux Sid (unstable)
batsman dot geo at yahoo dot com

And Bruce is effectively building BruceIX
	-- Alan Cox