Thanks for explanation, and patch.
This time it is failing on hypot
unhandled SIMPLE-ERROR: The foreign symbol "hypot" is undefined.
2008/10/28 Nikodemus Siivola <nikodemus@...>:
> On Tue, Oct 28, 2008 at 1:32 PM, Marko Kocić <marko.kocic@...> wrote:
>> This time the problem is acos. Maybe defining these operations in lisp
>> is not the way to go.
> As long as there are no active build maintainers for Win32, defining
> these operations in lisp is exactly the way to go: debugging "why
> doesn't the symbol FOO end up in the runtime and what magic do we need
> to add to take care of it" at distance is not something anyone does
> for fun.
> Attached patch adds rest of the trig and hyperbolic functions not
> implemented directly on iron as lisp-side definitions. (No, I didn't
> check that the branch cuts are what they are supposed to be -- that
> can be done once the Win32 build works again...)
>> The question is why it is not recognized during build.
>> I remember that build worked before I switched to gcc-4.3.2. Maybe
>> that is the cause that should be fixed.
> The problem is that GCC doesn't link the definition to the runtime, so
> we don't get the static address we want. I'd rather maintain the
> lisp-side definitions than fight with GCC.
> -- Nikodemus