[CentralOH] Python on Ubuntu Question

[Steven Huwig]

As I recall, there are many reasons Stackless isn't now in Python:

- it would break existing C library bindings, and make embedding and extending CPython much more difficult.

- when it was active, it targeted x86 only, but SPARC and PowerPC were significant players among many Python users.

- It would be too difficult to support its semantics in Jython

- It would be much more work for the maintainers.

It was basically a pronouncement from Guido that it won't go in. Frankly I think it was the right choice -- PyPy is a much better long-term platform for that kind of work than CPython is.

Thanks,
Steve

On Nov 30, 2011, at 1:46 PM, Bryan Harris wrote:

> If you want to get higher performance out of a python program by using
> more than one core, you basically have to use stackless or processing.
> Threading can never use more than one core.
> 
> Stackless sounds like a nice solution and frankly I don't know why
> stackless hasn't just merged into the main python tree.  As it is,
> stackless is a PITA to get set up and the performance isn't that
> different for most things.  It would be nice to take away recursion
> limits and such.
> 
> Does anybody know what the disadvantages of stackless are?  Why hasn't
> default python interpreter gone stackless?  Is there a performance hit
> in some situations?
> _______________________________________________
> CentralOH mailing list
> CentralOH at python.org
> http://mail.python.org/mailman/listinfo/centraloh