[MLton] Re: cygwoes

Matthew Fluet fluet@cs.cornell.edu
Thu, 13 Oct 2005 11:28:13 -0400 (EDT)


> I compiled a slightly modified standalone version of that program
> (below) under MinGW32 gcc and Cygwin gcc

Thanks a lot.

> and got totally different results for MinGW32 gcc which lead me to 
> wonder whether there is something like a C typecast mismatch problem in 
> calculating the starting point and/or passing the result back to 
> VirtualFree or outputting the results.

Indeed, the behavior is definitely curious.  I'm becoming a little more 
convinced that the Virtual{Alloc,Free} that one sees through Cygwin are 
not precisely the Win32 native functions; exactly how Cygwin is wrapping 
them is what is puzzling.

> May I also suggest that you check out the VM Validator tools avalable
> free of the web.

O.k. I may try that over the weekend.