[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Stackless Python



Hi Tom,

I'd be interested working with your code when it becomes available. 
I'm doing a project in Python, and at the moment am using a very
simple implementation of channels which I threw together.  My
implementation is quick and dirty, or rather achingly slow and dirty
(depending on whether we consider development time or commstime), and
I look forward to evaluating yours.

Alan Grover

On 9/20/05, Tom Locke <tom@xxxxxxxxxxxxx> wrote:
> neil@xxxxxxxxxxxxxxxxx wrote:
>
> >All,
> >
> >Has anyone come across Stackless Python (http://www.stackless.com/)
> >before?  I've just seen some links to it, and it looks very CSP-like,
> >
> >
> Stackless does provide a reasonable basis for a CSP implementation. The
> channels do have CSP-like semantics.
>
> A while back I started building a full set of CSP primitives for
> Stackless but I switched to greenlets (a spin off from the stackless
> project). This was partly because greenlets run as a C extension to a
> standard Python build (stackless is a fork), and partly because
> greenlets have a very nice and concise API (while the stackless API is
> over-complex to my mind). Tasklet pickling is not supported by greenlets.
>
> I now have ALT, PAR, timers, regular and call channels, CREW locks, and
> the scheduler is integrated with Twisted so programming CSP based web
> servers is a cinch. All this stuff will go out on livelogix.net with a
> liberal license sometime soon, ish...
>
> Tom.
>
>