[Stackless] bad version in CVS
Christian Tismer
tismer at tismer.com
Tue Sep 17 11:41:55 CEST 2002
Robin Becker wrote:
...
> well you could always
>
> cvs admin -oREV xxx.py
>
> to delete the offending revision
Thanks! (I didn't know that command).
I will try to find out what's going on, a bit.
My change touched the most hairy code section:
the C stack switching.
My goal was to virtualize the C stack handling,
with the primary intent to be able to override
these methods from stacklessmodule and to be
able to do C stack compression.
The reason for that is that C stacks seems to be
annoyingly huge under gcc. Windows is performing
great, 2-3 times less stack space.
The final reason for this all: I'm going to
San Francisco this weekend for three weeks,
in order to incorporate the new Stackless into
the Ironport software. They used the old version
to quite some incredible extent, with bazillions
of switchable frames. I cannot afford to have
a memory footprint of 2-3 times of the old version.
A Frame object costs something around 80-90 words,
mainly due to the huge waste in the block stack
(20 slots with 3 words).
So, finally some frame compression is on my to-do
list as well :-)
My fear is now that this recent change shows a weekness
of the whole system. The primary change is that the
C stack is no longer restored by a memmove, but through
a call. I assume that my stack reference has been
correct just by chance. I need to find out if it is too
high or too low, and if it can be determined in a
platform independant manner or not...
wish me luck - chris
--
Christian Tismer :^) <mailto:tismer at tismer.com>
Mission Impossible 5oftware : Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a : *Starship* http://starship.python.net/
14109 Berlin : PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34 home +49 30 802 86 56 pager +49 173 24 18 776
PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04
whom do you want to sponsor today? http://www.stackless.com/
_______________________________________________
Stackless mailing list
Stackless at www.tismer.com
http://www.tismer.com/mailman/listinfo/stackless
More information about the Stackless
mailing list