On Fri, 2005-05-27 at 12:33 +0100, dan@... wrote:
> The interrupt-thread bug has a low priority for fixing (at least for
> me) because it only shows up in intense use of interrupt-thread, and,
> well, your code would have to be pretty strange to do that a lot.
> interrupt-thread is for debugging and introspection, not for use as
> a general programming construct
Are you sure the bug has anything to do with interrupts though? I
thought noone understood what caused this bug and consequently it's
quite possible that it can be triggered by things other than interrupts.
If you know differently, please tell me so I can stop wasting time on
it. I always thought the bug had more to do with making allocations
atomic and using interrupts was just a conveniant way to test it.
> When you say that your bug "appears to be the same bug", do you mean that
> you're abusing interrupt-thread in the same way, or that you see similar
> symptoms with /different/ usage patterns?
Very similar symptoms but different usage patterns. We don't use
interrupts. I have two different systems which suffer the same problem.
The only thing I can think of that they share besides threads is use of
sb-bsd-sockets (indirectly through kmrcl's socket interface) for serving
> If you are not using interrupt-thread in anger and can isolate a better
> test case that you know is actually due to /your/ bug and not just something
> with similar symptoms, that would probably be much more help to
> developers than these threats of "we may never use lisp again and it will
> be all _your_ fault".
Isolating the bug is one of the problems I'm having. That's why I keep
asking about known problems so I can use them as hints on where to look.
So far that interrupt test is the best lead I've got.
I'm not making any such foolish threats, nor am I looking for anyone to
blame. I'm just asking for help, Dan. Really, I wonder what drives you
to read such nonsense into what I write.
Robert Marlow <bobstopper@...>