Thanks to GroovyBee for reporting the .cargs problem! Also, pretty much
all of the remaining cruft coming from the misguided "GPU in Main" cult
has been removed. Never do in an assembler what you can do with either
macros or a good preprocessor!
Removing all automatic code munging in the background by the RISC
compiler, including GPU in main idiocy. This means there are no more
built-in macros either. If you really needed this functionality, you can
put macros in your source code to do it. It doesn't belong in the guts
of the assembler.
Apparently the compiler did not care if an INCBIN happened in a BSS
section or not. Now the compiler cares and gives an error if you try to
do this terrible, terrible thing. :-D
While the assembler is in better shape vis-a-vis 64-bitness, there's
still a long way to go towards refactoring the thing to remove all
the problems it has with crufty data structures, and I'm sure there's
still bugs left in things like the .rept directive.
There's still lots to do, like refactoring a bunch of stuff that's
still basically wrong, but that's par for the course. There may be
more things that need fixing (like the .rept function, for example).
I removed a few functions that are better off being handled by the
native system libraries (malloc for amem, strdup for nstring). Also,
cleaned up files in preparation for removing pointers from the token
stream--which is preventing RMAC from working on 64-bit systems.
Corrected definitions of BYTE, WORD and LONG to appropriate uintN_t values.. (it's warning like a mofo at the moment but tested and it assembled happily enough 68K and RISC)