Shibboleth Developers

["Scott Cantor" <>]

> We used the RPMs provided at shibboleth.internet.edu, extract from our
> documentation:

Yeah, I think my estimate of how well my patch was working was overstated,
sorry for the confusion. It wouldn't be that surprising to see that kind of
math, particularly once the up front RAM is taken out of the equation.

> We've compiled 2.8.0 but it's not yet on a machine with many logins, the
> memory usage though, has decreased significantly.

That would be surprising to me, but we'll see.

I do think that there's a problem, but I just don't think it's a (major)
leak. It's a design issue that causes the process to behave much more
reasonably on Windows than Unix, and I don't know that I can do much about
that anytime soon. I'll have to see how 2.0 behaves and then move on from
there.

Load testing of 2.0 would be more than welcome at this point.

-- Scott