Aside from weird interrupt-thread issues which I am still in the
process of debugging (and which I think it's unlikely you'd hit
outside the context of a stress test), I believe that threading per se
is pretty robust on linux 2.6. I don't claim that the standard
library is thread-safe yet, but empirically it's possible to do
threadish things like internet servers.
I'm not nearly as convinced about kernel 2.4. While I realise that
the only reports we see on sbcl-devel are the failures, so there may
just possibly be people using threads happily on 2.4, I have no idea
who they are and wouldn't advise it myself.
Therefore I propose to disable threading by default if the sys_futex
system call is not available at runtime when sbcl starts. Who's
happily using threads on a 2.4 kernel and would be disadvantaged by
"please make sure that the person is your friend before you confirm"
Get latest updates about Open Source Projects, Conferences and News.