FFI problem in 5.02.2 on solaris 2.6?

Simon Marlow simonmar at microsoft.com
Thu Mar 28 10:14:07 EST 2002


> I think it is probably due to the ancient version 2.7.2 of gcc we
> have by default on Solaris 2.6 here.  The sigset_t type is guarded
> by #ifdefs as follows:
> 
>     #if defined(__EXTENSIONS__) || (__STDC__ - 0 == 0) || \
>         defined(_POSIX_C_SOURCE) || defined(_XOPEN_SOURCE) || \
>         defined(_XPG4_2)
> 
> and I have had problems with this kind of thing before when using gcc,
> which I usually worked around by setting one of those symbols (chosen
> at random).  I guess gcc-2.7.2 predates some of the C standardisation
> efforts.
> 
> However, I just found out that we do have a newer version (2.95.2)
> of gcc installed as well, although it is not the default compiler.
> Everything seems to work okay when I use it, without any need to
> set symbols. 

That's good news.

> Sorry to bother you with the (non-)bug report!

Not at all ;)

Simon



More information about the Glasgow-haskell-bugs mailing list