(semi) stackless python

[Steven Majewski]

 The new generators in Python2.2 implement semi-coroutines, not
full coroutines:  the limitation is that they always return to
their caller -- they can't take an arbitrary continuation as
a return target.

 So you can't do everything you can do in Stackless, or at least,
you can't do it the same way. I'm not sure what the limitations
are yet, but you might be surprised with what you CAN do.
 If all the generator objects yield back to the same 'driver'
procedure, then you can do a sort of cooperative multithreading.
In that case, you're executing the generators for their side
effects -- the value returned by yield may be unimportant except
perhaps as a status code.

[ See also Tim Peters' post on the performance advantages of
  generators -- you only parse args once for many generator
  calls, and you keep and reuse the same stack frame. So I
  believe you get some of the same benefits of Stackless
  microthreads. ]

Maybe I can find the time to post an example of what I mean.
In the mean time, some previous posts on generators might give
you some ideas.

<http://groups.google.com/groups?q=generators+group:comp.lang.python+author:Majewski&hl=en&scoring=d&as_drrb=b&as_mind=12&as_minm=1&as_miny=2001&as_maxd=20&as_maxm=12&as_maxy=2001&rnum=1&selm=mailman.1008090197.6318.python-list%40python.org>

<http://groups.google.com/groups?hl=en&threadm=mailman.996870501.28562.python-list%40python.org&rnum=6&prev=/groups%3Fas_q%3Dgenerators%26as_ugroup%3Dcomp.lang.python%26as_uauthors%3DMajewski%26as_drrb%3Db%26as_mind%3D12%26as_minm%3D1%26as_miny%3D2001%26as_maxd%3D20%26as_maxm%3D12%26as_maxy%3D2001%26num%3D100%26as_scoring%3Dd%26hl%3Den>

The ungrammatical title of that last thread above:
"Nested generators is the Python equivalent of unix pipe cmds."
might also suggest the solution I'm thinking of.


-- Steve Majewski