stackless python

Courageous jkraska at san.rr.com
Wed Dec 19 23:51:06 CET 2001


>OTOH, the chances of seeing stackless python or something like it in
>the core are currently in the vanishing-to-nil range.  We went round
>with this one on the newsgroup a few weeks back -- google is your
>friend here -- and the conclusion was there was noone willing and able
>to do the enormous amount of work required.

A big rub is that it might very well be impossible to get stackless
to work properly in Jython, and even if it weren't, it's a tough
mark over there as well.

>CT has done the first 90%
>of the work, but there's still the other 90% to go, and even if the
>work was done, it's far from certain that Guido would accept the
>changes anyway.

Makes the kernel tougher to understand. Programmers with an under-
standing of stack maniuplation tricks, implementation of cooperative
multithreading frameworks, and forth will have no problems. But this
group is a rare breed.

If you really want to have a bad day, try implementation something
like continuations in C++. Then, when something goes wrong, try to
debug it. Har, har, har. *cringe*

C//




More information about the Python-list mailing list