[Stackless] The PyPy Stackless Module Re: Tasklet cleanup?

Andrew Francis andrewfr_ice at yahoo.com
Fri Jul 6 16:37:50 CEST 2012


Hi Christian:

Message: 1
Date: Thu, 05 Jul 2012 23:18:34 +0200
From: Christian Tismer <tismer at stackless.com>
To: The Stackless Python Mailing List <stackless at stackless.com>
Cc: Herv? Coatanhay <herve.coatanhay at gmail.com>
Subject: Re: [Stackless] Tasklet cleanup?
Message-ID: <4FF604AA.3050409 at stackless.com>
Content-Type: text/plain; charset="iso-8859-1"; Format="flowed"

>Just as a note:

>the PyPy stackless module was not really maintained in years, and I
>remember I had a more compliant version prepared, but lost interest
>after the Jit became unsuitable for the stackless version.

>Now it's time to finalize and correct this, because a Jit-enabled
>stackless is in reach, again.

>Will try to get to it, soon.


A note. The last time I looked, the version of the PyPy Stackless
module that took advantage of (single-shot) continuations. However unlike
the older version, it took out greenlet support. I rather liked the way
the code could execute over PyPy's coroutines or greenlets. 

As a sidenote. Myself I have two broad versions of stackless.py. One
is in the Stackless repository that supports select(). I have a variation where
I experiment with improving Pike's select algorithm - I think it is too 
conservative.

The second version that I haven't released since there is even less interest than
that the version with select, is my initial experiments with join patterns. Since 
in the literature, the join implementation (and channel implements) also uses STM 
and other lock-free techniques, I probably resume experimentation with CPython-atomic
as well as improve the API.

Cheers,
Andrew
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.stackless.com/pipermail/stackless/attachments/20120706/7a71d5f0/attachment.html>


More information about the Stackless mailing list