On Oct 25, 2005, at 7:41 AM, Kai Schmidt wrote:

> I run that def() below, and gave up at about 3 Million, when my (huge)
> swap-partition startet to fill up. Very nice to look at diverse
> graphical
> (KDE) system-monitors btw.

Okay, so that seems to rule out the limited-dict-size possibility.

> I can now reproduce those memory errors with durus only having 100
> persistent
> dict() in it.

Can you send me that program (as an email attachment)?.

> Can I assume, that there is nothing really wrong in my sample code?

I didn't notice anything wrong with it.

> I have gcc 4.0.2, could that be a problem?

I don't know.

> IMHO I should not be able to
> produce a memory|pointer problem with python, should'd I?

My HO is the same as yours.

> Is there a way to run Durus with python only? No C-helper? to test ...

Delete _persistent_base.so and Durus will be pure-python.

If you really want to track this down, I think you should run python
with gdb and find out just where this happens.  If you send me the
program and I can reproduce the problem, that's what I will do.