OSDev.org

konacake wrote:I was trying to continue the discussion, I was saying OP is blindly afraid of MinGW of "what it might do" to your binary because you feel like it would be easier to make a whole new compiler/linker than just find out how an existing one works, an existing one pretty much everyone here uses at that (GCC, not specifically MinGW), all because you're so petrified of Linux

I am not petrified of Linux, I just don't like it.

mathematician wrote: I am not petrified of Linux, I just don't like it.

That didn't exactly address my points. The point is you don't know how MinGW works, so you automatically believe it "adds in gunk". You should learn how to use something before saying it doesn't work or worse, criticize it. If you like Watcom, fine, but don't say something doesn't work because you can't work it.

konacake wrote:

mathematician wrote: I am not petrified of Linux, I just don't like it.

That didn't exactly address my points. The point is you don't know how MinGW works, so you automatically believe it "adds in gunk". You should learn how to use something before saying it doesn't work or worse, criticize it. If you like Watcom, fine, but don't say something doesn't work because you can't work it.

You shouldn't have to fight tools, and I have no appetite for doing so.

The Microsoft (free!) and Intel compilers run natively under Windows and can be used for osdeving. They output PE files, but it's not a big deal because it's no harder than dealing with ELF files. If it's really an issue, you can write a program that converts a PE file to your own binary file format, and that would be much simpler than writing an entire compiler yourself (which would also involve writing a linker that parses and produces binary files - so you're not avoiding that task).

mathematician wrote:You shouldn't have to fight tools, and I have no appetite for doing so.

Then you're not going to have any luck as an operating system developer.

konacake wrote:I'm sure you're happy to know Windows NT was actually Microsoft's attempts at cloning Unix, and was designed by some Unix wizards.

Actually, the primary model of NT was VMS, not Unix, though there were many Unix influences stemming from a half-hearted effort at POSIX compliance. But frankly. I'd rather not get into the arguments on this point. In the meanwhile, let me see what I can find to say about macrology.