Is Stackless Python DEAD?

Bruce Dodson bruce_dodson at bigfoot.com
Sat Oct 27 22:35:09 CEST 2001


Would Stackless have a better chance of making it into the core if its
initial PEP talked only about increasing performance and removing limits on
recursion?  This advantage is not as "sexy" as continuations, but is also
less controversial.  It is still something that everyone can understand, and
without any exported facilities for frame-switching, it alone does not
require changes to most existing C extensions.  It does still require
additional book-keeping code to be added to the core, much like Stackless.

Microthreads and coroutines can be sold as advantages in another PEP, one
which does treat that decoupled stack as a tree, and does permit arbitrary
frame switching.  This PEP would have to talk about the risks and changes
required to existing modules, to make them work safely in program that takes
advantage of continuations.  A "future" import might help.  Meanwhile
Stackless remains a separate branch, but hopefully becomes easier to
maintain against a CPython which already has a decoupled stack.

Bruce





More information about the Python-list mailing list