Michael L. VanLoon:

> I think it's been blown a little out of proportion.  I have had
> uptimes of greater than a month on my system, and I run some rather
> large processes fairly often.  If you let your system stay up for
> awhile, you will notice that the swap usage rises quickly at first,
> but then tapers off.  It doesn't just keep consuming until it dies
> (unless you don't have enough swap).  It reaches a plateu, then stays
> there.  It's not a "leak" -- it will not increase without bound.  It's
> just over-usage.

Well, if this is how it works, then the problem isn't really that
serious.  I haven't had very long uptimes recently because of the
problems I was having with the MLink program.  Uptime is about 21
days now and swap usage has grown from <6 MB to 13 MB.. but maybe
it's now growing any more.

> This is my experience, anyway.  If someone who has actually worked
> with the code has enough experience to tell me I'm wrong, and why,
> I'll accept that.  Until then, I'll believe this is not a "leak".

Well, I hope someone will tell us whether you're right or wrong. ;)
I hope you're right. ;-)

  -jm