Hi,

I am about to write some x86_64 code again, but I am wondering how much fasmg (x86_64) assembler code is friendly to a C preprocessor. As far I as understand it, fasmg tokens should be ok with C preprocessor tokens.

Am I wrong? Or is there some obvious pitfalls experienced fasmg coders see which I don't.

This assembly code will very probably share some definitions with some C code.

Additionnaly, I may try to write the code in a way I could assemble it with binutils GAS *and* fasmg. Is this reasonable? I may do some "sed"-ing on the source file in order to "include" some "assembler abstraction layer" (trying to keep that minimal).

One pitfall that I know of is that quoted string tokens are interpreted differently by C vs fasmg. For example "\"" is a valid token for C, but error for fasmg, which requires """" syntax instead (legacy of assemblers like TASM).

I've been considering adding a switch (similar to RETAINCOMMENTS/ISOLATELINES directives) to support C-style quoted strings, but that would require a substantial refactoring of the tokenizer and its data structures, so I postponed it indefinitely.

Since sed is on the table then pre-parse the source.

Allright, I may give it a try, keeping in mind a sed pre-preprocessing may be required.