MemoryError in 2.2.0 in odd case: Can I reduce memory requirements?

[Doug Lee]

On Wed, Jan 12, 2011 at 10:01:13AM -0600, Matthew D. Fuller wrote:
> On Wed, Jan 12, 2011 at 09:52:44AM -0500 I heard the voice of
> Doug Lee, and lo! it spake thus:
> > 
> > I'm trying to manage local changes for a team using Bazaar.  I share
> > local commits through a common branch on our Internet server, which
> > (sheepish grin) happens to run FreeBSD 4.11 - very old, but
> > difficult to upgrade at this point.  The server runs Python 2.4 and
> > uses the slow Python code in place of things like CElementTree, in
> > case that matters (but for this issue, I doubt it does).
> 
> Well, you'd have to multi-step it to upgrade, but it's not impossible
> in the general case.  You should (I think) be able to get as far as
> py2.6 with the 4_EOL ports tree.

I did a cvsup from that tag today and only see through python25.

> > I can't seem to extend the datasize limit past 512 meg even as root,
> > so I assume this is set in stone for some reason.
> 
> This is an in-kernel limit; i386 uses 512 meg as a default data size
> limit.  You can tune this with the MAXDSIZ option in the kernel
> config.  Actually, I think it's a loader tunable in some later
> versions, but I doubt for 4.  But if you've got the RAM, you can bump
> it to a gig or so without worries.
> 
> 
> > I believe if bzr hit a stacksize limit, it would get a SIGSEGV,
> > which I'm not seeing.
> 
> Yes, it's almost certainly running out of heap space.

So which thing should I try bumping to increase heap space?  I should
know this, but I seem not to...


-- 
Doug Lee                 dgl at dlee.org                http://www.dlee.org
SSB BART Group           doug.lee at ssbbartgroup.com   http://www.ssbbartgroup.com
Reality is not what we see.  Reality is what is, whether we see it or not.
(03/01/09)