------- Additional Comments From dlstevens at us dot ibm dot com 2006-03-14 19:59 -------
"The user" would be me, and I didn't get incorrect results from a man page.
The problem, as I said in the first two entries, was that I was never able
to get a NULL return from malloc(). I always either got "success" (in some
cases without actually getting memory) or a segmentation violation. If I
recall, a malloc() that exceeded the soft limit returned the same pointer as
an already-allocated and not freed prior malloc(), which would be wrong.
It certainly is possible that I had a misconfiguration on my system, and I lost
the context beyond what I wrote here more than a year ago. So, closing the bug
is not unreasonable, but I'd be happier if you had a test that manipulates
the soft limits (correctly, if what I did was wrong) and results in a
successful allocation or NULL return from malloc(), always. No segmentation
faults, no garbage returns, etc.
In other words, if you can demonstrate a case where setting a soft limit
results in malloc() returning NULL (ever), then I think your test will be
farther than I ever got, and I'd be happy with having the bug closed. :-)