Agents

Deep in the darkest depths lurks an ancient horror, when the time is right it will rise forth and leave you screaming for mercy and begging for forgiveness…

OK, I have a penchant for being over dramatic but in this post I am going to reveal some little known caveats in a well known and much revelled area of F#, agents aka the MailboxProcessor. Gasp!

First let me give you a demonstration:

This will be the last post on rebuilding the MailboxProcessor using TDF, here’s a quick discussion of the missing pieces…

First, lets start with the simple ones, these don’t really require much discussion.

DefaultTimeout

let mutable defaultTimeout = Timeout.Infinite

member x.DefaultTimeout
   with get() = defaultTimeout
   and set(value) = defaultTimeout <- value

This simply provides a mutable property using Timeout.Infinite as a default setting.

CurrentQueueLength

member x.CurrentQueueLength() = incomingMessages.Count 

Another simple one, this methods uses into the underlying BufferBlock to extract its current queue length using its Count property.

Right, no messing about this time, straight to the code.

Construction

This is pretty straight forward and I don’t want to detract from the important bits of this post, the only thing of note is the cancellationToken which is initialized to a default value using the defaultArg function if the optional parameter cancellationToken is not supplied. The TDF construct that we to use to perform most of the hard work is incomingMessages which is a BufferBlock<'Msg>.

This is going to be a new series on using TPL Dataflow with F#. First a little bit of history and background.

TPL Dataflows heritage and background

TPL Dataflow or (TDF) has been around for quite a while, it first surfaced more than a year ago as the successor to the Concurrency and Coordination Runtime (CCR) and with coming release of .Net 4.5 it will be part of the System.Threading.Tasks.Dataflow namespace. Elements of the now halted project Axum are also present within the design of TDF.

One of the areas that I am very interested in is agents and I have been doing quite a lot of work in this area lately.

Agents can be used for a multitude of different purposes ranging from: isolated message passing, object caching, finite state machines, web crawling, and even reactive user interfaces.  One of the ideas that I have been looking into lately is agent based scheduling.

Everyone knows F# agents are cool right?  Well here’s yet another example of how versatile they can be…

There was a series of posts last April by Stephen Toub from the pfxteam at Microsoft.  I was reading through some of the posts again the other day and thought some of the ideas presented there would make interesting projects in F# to demonstrate the flexibility and succinctness of the language.  I thought the ObjectPool example would make an interesting project in F# using agents aka MailboxProcessors. An ObjectPool is basically a pool of objects that have been pre-created so that you can grab one and use it, and then place it back in the pool when you’re finished.  They are useful in situations where the cost of creating object from scratch is very high or you want to cut down on allocations in the garbage collector.