[Stackless] PyCon 2011 Talk Acceptence and Questions on Greenlets

[Andrew Francis]

Hi Guys:

A new version of my 2010 EuroPython Talk "Prototyping Go's Select with stackless.py for Stackless Python" got accepted as a part of the extreme track. I want to learn from my EuroPython mistakes and make this an accessable, accurate, and overall great talk that makes folks want to use Stackless.

Between now and March - I would like to do the following:

0) Fix up stuff in the stackless repository.

1) Fix bugs and annoyances in the modified stackless.py (I would appreciate if someone could come up with a useful name).

2) Prototype a version that uses join patterns. I will soon be posting a series of blog posts on join patterns.

3) Protototype a version of stackless.py that tries to improve the Pike algorithm: 1) replace random section with operations/events based on a timestamp. 2) Include a "hint" for no tear down. 


4) Incorporate the changes from 1,2,3 into Stackless C Python as well as fix bugs.

On a different note, I am trying to learn more about that greenlets do under the hood. In part, I am running the greenlet code in a debugger to get a feel for what is happening but I am not quite sure what I should be looking for specifically. My question is I don't see any modifications to the eval functions in ceval.c. So is this what is meant at a more technical level by hard-switching? 

Any answers, suggestions, or overall comments would be appreciated!

Cheers,
Andrew