[Stackless] Stackless ported to 2.2.3

[Dinu Gherman]

Christian Tismer:

> I'm not aware of problems with that, so far.
> Dinu has checked Stackless for OS X several times now,
> and I'd like to ask him for help, again.
> At least I thought that before moving to Py 2.2.3, things
> worked fine for OS X.

I just compiled the very current CVS version from centera.de without
any problems. I haven't run all tests, but this one seems to work ok
(on a 600 MHz iBook running OS X 10.2.6):

[localhost:packages/stackless/src] dinu% python.exe 
Stackless/test/taskspeed.py
2.2.3 Stackless 3.0 030821 (#1, Aug 23 2003, 10:30:02)
[GCC 3.1 20020420 (prerelease)]
10000000 frame switches      took  21.93000 seconds, rate =     455996/s
10000000 frame softswitches  took   6.36000 seconds, rate =    1572327/s
10000000 cfunction calls     took   7.91000 seconds, rate =    1264222/s
10000000 cframe softswitches took   2.32000 seconds, rate =    4310344/s
10000000 function calls      took  16.77000 seconds, rate =     596302/s
10000000 generator calls     took   8.96000 seconds, rate =    1116071/s
10000000 cframe switches     took  17.36000 seconds, rate =     576036/s
10000000 cframe 100 words    took  23.00000 seconds, rate =     434782/s
The penalty per stack word is about 0.325 percent of raw switching.
Stack size of initial stub   = 108
Stack size of frame tasklet  = 220
Stack size of cframe tasklet = 204

Dinu

--
Dinu C. Gherman
......................................................................
"It was wonderful to find America, but it would have been more wonder-
ful to miss it." (Mark Twain)