[Stackless] Thoughts on I/O Driven Scheduler
andrewfr_ice at yahoo.com
Mon Mar 28 17:19:54 CEST 2011
> Message: 1
> Date: Mon, 28 Mar 2011 10:23:37 +0800
> From: Richard Tew <richard.m.tew at gmail.com>
> To: The Stackless Python Mailing List <stackless at stackless.com>
> Subject: Re: [Stackless] Thoughts on I/O Driven Scheduler
> Message-ID: <BANLkTiknf8t1PyEC_ceiRq+HZeS-PVMCCg at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
> Greenlet is where the attention and interest is, and I do
> not find it surprising. It is a lot more approachable to use an
> extension and gain the bulk of the benefits of Stackless, than to use a >custom Python interpreter. To be honest, I see this as a positive thing >that speaks for how token and cumbersome generator coroutines are as a >solution for the same sorts of problems.
You are right about Greenlets being the centre of attention. And I think generator coroutines are a step backwards. However with stuff like virtualenv, a custom interpreter should not be so onerous.
I guess my current reality is that I will be using greenlets (or pypy-c) for prototyping new concurrency features. The stuff that interests me requires fancy features though I am still trying to decide if it is best bake them into the implementation. Analysis will answer this question.
Given the difficulty of putting new features into C based Stackless, maybe custom stackless.py modules are the way to go (I am reading Kristjan's post).
> No idea about Twisted related topics. I have little
> understanding of, and no interest in the subject.
If you are not interested in Twisted, fair enough. However I think that the general principles behind integrating Stackless with Twisted apply to most asynchronous systems.
More information about the Stackless