[MLton] Debian package (was: Finished (?) MLton.Child)

[Wesley W. Terpstra]

On Thu, Dec 02, 2004 at 05:41:08PM -0600, Henry Cejtin wrote:
>         ulimit -s x to set max stack size in K-bytes

That's one solution. Actually, I was surprised the kernel (2.6.8) didn't
kill the process itself. I might try that if I end up needing swap after
all. Besides, laptop disks are dirt slow; things seem much snappier now.

> I wonder if it really is a RAM problem.  All of my real machines have ECC,
> but for my laptop that isn't really an option. 

The machine with ECC had this problem?
Then I don't know.

What you describe only happens to me on my (non-ECC) laptop.

OTOH, are you setting ulimits like above?
MLton needs a lot of memory and it might be you aren't giving it enough.

> Did you ever try compiling a linux kernel?  A segfault there used to be
> the standard way to test RAM.

Yes, I've used my system reliably for half a year.
It passed every RAM test I threw at it (including many kernel compiler over
the months) and Windows based system testers. Only MLton found the problem.
Remember that these days gcc doesn't use nearly as much memory as MLton ...

-- 
Wesley W. Terpstra