--- In iolanguage%40yahoogroups.com">iolanguage
yahoogroups.com, Steve Dekorte <steve...> wrote:
>
>
> On 27 Mar 2007, at 08:15 pm, jeback2 wrote:
> > 222 longjmp(next->env, 1);
> > Program received signal SIGSEGV, Segmentation fault.
> > 0x00002aaaab17d35e in __longjmp () from /lib64/libc.so.6
> >
>; > I haven't the foggiest as to the inner workings of Io. If anyone has
> > any suggestions as to how to proceed with trying to resolve this and
> > is willing to play the pupeteer, I would be happy to play the puppet.
>
> It looks like the instruction and stack pointers aren't being set in
> the jmpbuf the way that amd64 wants them. Has ucontext been fixed on
> linux/amd64 (the bug is in linux, not Io) yet? If so, we could switch
> to using it.
>
> - Steve
>
I'm not familiar with the ucontext issue; I did a quick search and
came up with this:
http://sourceware.org/bugzilla/show_bug.cgi?id=1468
A bit more searching and I learned that a coroutine package for Perl
(Coro) by Marc Lehmann can be configured to use ucontext or setjmp. I
tested the ucontext version and it passed all of the included tests,
but I have no idea if it tested the functions that would be relevant
for Io.
Do you know of any programs that would serve as a good test for the
status of ucontext on amd64/linux?
Regards,
Jan Erik
.
