Re: library installation/Mac OS X

[Nathaniel Irons <ndi-l-pari@bumppo.net>]

On 11/6/01 at 12:31 PM, Karim BELABAS <Karim.Belabas@math.u-psud.fr>
wrote:

> OTOH, I would have expected it to work; doesn't Darwin support
> dlopen? Does 'man dlopen' mention anything exotic that would
> prevent both
> 
>   /usr/bin/gcc  -O3 config/has_dlopen.c
> 
>   /usr/bin/gcc  -O3 config/has_dlopen.c -ldl
> 
> from compiling ? (that's $CC $CFLAGS ..., expanded from your log)

dlopen is in place, but both commands fail with:

    config/has_dlopen.c:2: header file 'dlfcn.h' not found

I have a man page for dlfcn, shared with dlopen, dlclose, etc, but
locate turns up no dlfcn.h.

> BTW, could you send me your processor and clock speed so that I can
> update the MACHINES file ? (no entry for Darwin yet)

G3, 350.  I'll send you a bench profile from a different machine later,
off-list.

On 11/6/01 at 09:31 AM, Gerhard Niklasch
<nikl@mathematik.tu-muenchen.de> wrote:

> (NB this is documented.  And make bench should in fact have
> issued the explanation:
>   NOTE: actually, the only problem noted is that install() does
>   not work in your configuration, so you will not be able to use
>   this feature. Otherwise you can safely ignore the above warning.
> )

FWIW, that notice wasn't issued on either of the OS X systems on which I
installed pari.

Thanks for the insight.  Please let me know at this address if there's
anything I can do to continue to test under darwin.  I'll remain
subscribed for a week or so after this discussion has apparently run its
course.

  -nat