I concur. I wasn't suggesting this as an alternative to your project, just that I believe that the problems that stop many of the modules from building on Win32, run much deeper than either the compiler that is used to build them, or artificially false statistics generated by a broken, down-level, automated build process. I felt that the aims of the project as described are fuzzy and incomplete, and I requested clarification of those aims.

The bottom line from my perspective is that if any modules that are successfully transferred to the Win32 platform as a result of your project, are binary compatible with AS builds which a couple of people have said they will be, then everyone is a winner.

I still have my reservations as to how many of the incompatibilities your project will resolve. I also have doubts as to the efficacy of hacking solutions to each module individually, rather than attempting to package generic fixes to the underlying problems and feed them back into the core so that future modules pick them up automatically from there. As I said in my first post in this thread, maybe Strawberry is the right way to start, and maybe pointers to how to achieve generic fixes will come from it.

Either way, I wish you the best of luck in your endeavours, and I am sorry that my attempt to ellicit clarifications have lead to misunderstandings of the 'us & them' variety.