Not sure if it's a VS thing since I get different compile errors with either the MS or Intel C compiler under the same environment (Visual Studio for x64 console). Let me start with a more basic question: I'm using the package available on SourceForge (1.2.1, 2008-10-29 14:36). If it can be confirmed this package builds happily via VS 2008 x64 console, can I get a look at the build log and/or configuration? I should be able to duplicate this with identical tools (I hope). If it makes a difference I will try to grab the latest & greatest source instead.

   ~Mike C.

On Thu, Jan 29, 2009 at 8:50 AM, Matthieu Brucher <> wrote:
> numpy\core\src\umathmodule.c.src(64): error: expected an identifier
>   static float frexpf(float x, int * i)
>                ^
> numpy\core\src\umathmodule.c.src(64): error: expected a ")"
>   static float frexpf(float x, int * i)
>                ^
> numpy\core\src\umathmodule.c.src(65): error: expected a ";"
>   {
>   ^

Perhaps a macro that is replaced? Visual Studio defines min, max as
macros. Maybe it is the same for frexpf? Could you check in the
preprocessed C++ file ? /E to generate it Intel Compiler, I think.

Information System Engineer, Ph.D.
Blogs: and
Numpy-discussion mailing list