My first post to this list...
I started playing around with RDNZL, Direct3D 9.0, and the evaluation version of LispWorks for Windows about a week ago. Very nice! But I've run into an interesting problem.
Direct3D uses single floats throughout its interface, but in LispWorks for Windows, all floating point types are equivalent to double-float. So I've got this rather difficult situation, where RDNZL can never find a method if it takes a single float argument! Even if it could find it, a down-conversion from double to single float would be needed, which probably shouldn't occur implicitly.
I tried explicitly calling System.Convert.ToSingle, but by the time I can get my hands on the return value, it has been turned back into a double! I haven't been able to figure out a way around this without modifying RDNZL, or writing some sort of .NET function that returned a boxed single. But I'm not sure that the latter wouldn't be unboxed, or that it would give the correct type for the method lookup in RDNZL.
Any ideas?
On Fri, 17 Feb 2006 00:49:16 -0500, "Dan Muller" s8ctxw402@sneakemail.com wrote:
My first post to this list...
Hi... :)
Direct3D uses single floats throughout its interface, but in LispWorks for Windows, all floating point types are equivalent to double-float. So I've got this rather difficult situation, where RDNZL can never find a method if it takes a single float argument! Even if it could find it, a down-conversion from double to single float would be needed, which probably shouldn't occur implicitly.
I tried explicitly calling System.Convert.ToSingle, but by the time I can get my hands on the return value, it has been turned back into a double! I haven't been able to figure out a way around this without modifying RDNZL, or writing some sort of .NET function that returned a boxed single. But I'm not sure that the latter wouldn't be unboxed, or that it would give the correct type for the method lookup in RDNZL.
Yeah, tough call. I don't see an easy general solution but for the moment I've uploaded version 0.9.3 which offers the following workaround:
http://weitz.de/rdnzl/#*coerce-double-floats-to-single*
You should be able to (temporarily) rebind this variable to T for single float arguments. Well, UNLESS your .NET method's signature contains both System.Single /and/ System.Double.
Let me know if that helps or if you have a better idea.
Cheers, Edi.
-
Dan Muller -
Edi Weitz