Stackless Python, eventual merge?

Michael Chermside mcherm at destiny.com
Tue Sep 17 16:09:33 EDT 2002


Ian Bicking <ianb at colorstudy.com> writes:
>It has been simplified, but I don't think that changes anything with
>regard to getting it into standard Python -- the same barriers still
>exist.

Martin v. Loewis wrote:
> You are making these barriers up in your mind. I don't know why you
> are doing that, though.

Okay, let's see if we can settle this... because I think it has real 
potential.

I am going to pose a hypothetical situation... please don't argue about 
whether these conditions are true now, or whether they will be true, but 
take these as assumptions:

  (1) Some version of Stackless consists of a quite small non-complicated
      patch onto C-Python.

  (2) The default behavior of this Stackless is to NOT allow stack
      switching... it must be enabled explicitly to be used.

  (3) Someone (probably NOT Christian) is willing to take on the job of
      merging it into the core and maintaining it afterward.

  (4) Stackless works on some architechures (including at least X86,
      Sparc, PPC and Cygwin). But it does NOT work on some other
      architectures. So (rather like the thread package) it is not
      as portable as Python itself.

  (5) Good reasons exist why Stackless CANNOT be implemented in Jython,
      in Python for .NET, or other Python implementations -- it's a
      C-Python-only thing.

The question is this:  Would the advantages of making Stackless part of 
standard Python outweigh the disadvantages of non-portability and 
C-Python specificity?

If the answer is "NO", (and I can see good reasons why it would be) then 
no one need bother worrying about integrating Stackless into the 
standard. Let it live a separate existance!

If the answer is "YES" (and I can see other good reasons why THAT might 
be) then integration MIGHT be possible, but there's no reason why it has 
to be, or needs to be hurried. It can wait until someone steps steps up 
to make (3) take place.

And I agree with Christian's sentiment:

> Please let me stay what I am: The one who writes Stackless
> Python, supports and maintains it. I really do not take
> responsibility for being standard or non-standard. 

... leave Christian to do what he's doing well, and let someone else 
worry about possible integration.

-- Michael Chermside







More information about the Python-list mailing list