On Fri, 16 Jan 2015 11:40:38 -0800, Ben Goetter wrote:
> On 1/16/2015 11:00 AM, Lehi Toskin wrote:
>> Good news is it's giving me a new error! Yay! "%1 is not a valid Win32
>> application.; errno=193"
>
> This is /the/ classical 32-versus-64-bit mismatch error. Most likely,
> once of the DLLs that your DLL in turn depends on is loading in the
> wrong flavor. Bitness is the most likely suspect, but you could also
> conceivably have a version compiled for a different target architecture.
>
> To decipher this and future error codes, look them up in winerror.h in
> your Windows SDK. Then go grepping through msdn.microsoft.com with the
> name of the constant. This one is ERROR_BAD_EXE_FORMAT.
>
> You probably should take Racket out of the loop here. Write a four-line
> C program that explicitly tries to load your DLL. Something like
>
> #include <windows.h>
> HANDLE h = LoadLibraryA("mydll.dll");
> if (NULL == h) printf("ERROR %lu\n", GetLastError());
> else { FreeLibrary(h); printf("Great success\n"); }
>
> Compile that fragment in the same bitness of your Racket installation on
> Windows, and use it to ensure that your DLL itself can resolve all its
> dependencies. Once your DLL can pass this test, feed it to Racket.
>
> Ben
>
> ____________________
> Racket Users list: http://lists.racket-lang.org/usersI have used that little bit of checking code on the libraries mylib depends on and they all load, but if I compile mylib against them it fails that check. As far as I can tell, everything under the sun is 64- bit. I am halfway tempted to say that my Windows VM is haunted and wash my hands of it.
signature.asc
Description: PGP signature
____________________ Racket Users list: http://lists.racket-lang.org/users
