Python 2.0 and Stackless

[Gordon McMillan]

Cameron Laird wrote:

>In article <WVf0iLAGEBj5EwXf at jessikat.fsnet.co.uk>,
>Robin Becker  <robin at jessikat.fsnet.co.uk> wrote:
...
>>so PEP for stackless please C Tismer.

>Is Christian the one to do it?  Maybe his time is better spent
>elsewhere ...  

It is, so I volunteered to author them.

>There are several questions in this area that interest me.
>What *should* a PEP say?  Is it about semantics or implementation?

There will be a technical PEP about the implementation of stackless; 
stackless enables continuations, coroutines and uthreads but by itself does 
none of these things. There will be an informational PEP about uthreads and  
coroutines which serves as justification for stackless. Continuations, like 
the raw thread module, the "new" module or metaclasses, is not for the 
faint of heart.

>Do the Stackless fans really care whether it shows up in core
>Python, apart from altruism?

Yes. Stackless needs to be part of core python, or we have a fork. The 
other stuff can all be extensions.

>Is there any serious compatibility issue with CPython?  

None at all with 1.5.2. Christian will have to bring it up to date which is 
more than just applying patches, but not a whole lot more.

>What *are* the issues for the mass of Python users?  Will they
>ever be aware of Stackless apart from its microthreadability
>and not-yet-in-existence co-routine interface?  

No. The only rational issue is that ceval.c and the manipulation of frames 
is no longer quite as straightforward. There are other issues, of course, 
but I just implicitly classified them :-).

>Gordon and Just,
>do you agree that most application developers will never want to
>  import continuation
>?

Yes.