Re: Patch: assert in natObject.cc

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Hans" == Boehm, Hans <hans_boehm at hp dot com> writes:

Hans> Looks right to me.  I was extremely paranoid about this code,
Hans> probably with good reason.  But it seems to have enough miles on
Hans> it by now that it's appropriate to turn the assertions off.

Ok, thanks.  I'll only do this on the trunk.

Hans> Based on my experience, this doesn't have a huge performance
Hans> impact, in part because the tests probably schedule well, and
Hans> because the compare-exchange operations are typically so
Hans> expensive.

That was my experience too.

Hans> While we're at it, I attached another patch that has been in my
Hans> tree for a while.  The substantive part of the change is to
Hans> replace a % with an &.

Please check it in, thanks.


Recently I was looking at the new Thread.holdsLock API.  Seems like
we'll need a bit more support in natObject.cc to make this work.
Something else for the to-do list...

Tom