Python Interpreter Reentrancy?

Courageous jkraska1 at san.rr.com
Sun Apr 22 13:02:37 EDT 2001


>>I'm asking for a reentrant Python interpreter -- one in which a global
>>interpreter lock isn't necessary. Clearly that's not around yet.
>
>Nope, although Stackless comes pretty close.  Quite frankly, I find the
>GIL rather useful in a lot of ways, and I hope that if/when the GIL
>removal happens it's strictly optional.

I could go with that.

I presume when you say "Stackless comes really
close," what you really mean is that the frame objects, being freed from
linkage to the c-stack, might allow the interpreter to eventually be
made fully reentrant?

I ask this because the threads (and underlying thread capability) in
Stackless are all cooperative, and just wanted to be sure on what
we're talking about.

C//




More information about the Python-list mailing list