Ralf Hemmecke wrote:
>
> > Note that GMP test is not perfect: we absolutely need
> > _shared_ GMP,
>
> Why is that absolutely necessary? Because of SBCL? Or because you set up
> the Makefile in such a way that it only builds gmp_wrap.so? Is there
> really no way to make it working with libgmp.a?
>
AFAIK at runtime sbcl can only load shared libraries. Theoretically
only gmp_wrap.so needs to be shared library, but on most platforms
it requires shared GMP. To use static library we would need
sbcl sources and patch its build process to link to GMP at
build time ...
> > but AFAIU the test will pass if only static GMP is installed.
>
> That's true as far as what I've tested now.
>
> So if libgmp.a alone is useless, then I'd have to refine the test a bit.
>
> I guess, I'd have to search LD_LIBRARY_PATH for the libgmp.so (after the
> test, that I have already). Or is there any other variable I'd have to
> check. I guess that might be platform dependent, right?
>
What we really need is ability to compile a file using GMP header,
and link a shared library to GMP. This means that we need correct
compiler flags to force generation of shared library instead of
an executable. As I wrote, I found it hard to invent a test that
is significantly simpler than gmp_wrap.c...
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/fricas-devel?hl=en.
