> Here is a little program that produces some interesting results
> on at least 2 platforms. Both systems were using 6.0.1.
>
> This (silly) program works on x86-linux. It produces an illegal
> instruction SIGILL on a sparc, and there is some OS-related
> problem on FreeBSD and OpenBSD (which is my issue, I know). The
> sparc problem is the more interesting of the two:
>
> $ uname -msr
> SunOS 5.6 sun4u
>
> $ ghc -package unix bug.hs
> $ ./a.out
> zsh: illegal hardware instruction (core dumped) ./a.out
I was a bit sloppy with the implementation of getGroupEntryForName,
assuming that the system has a POSIX-compliant implementation of
getgrnam_r. This is rarely true, it seems.
I've committed a fix for Solaris (the man page says you have to define
_POSIX_PTHREAD_SEMANTICS to get the POSIX vesrion). Could you look into
the situation on OpenBSD and see if we have to do something similar
there? I'll take a look at FreeBSD later.
Cheers,
Simon
_______________________________________________
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
