Right, with the syntax I chose there ends up being, IIRC, 6 different forms (<op>, <op> <arg1>, etc.), which can't be nested. Some of those forms share patterns which could be described as a nested grammar, but I just abstracted them in the parser code as functions (i.e. functions to parse patterns that show up in multiple forms).

I did use a lexer. I wrote an assembler before without using a lexer; using the lexer isn't that much simpler IMO, but it makes it more robust and makes it much easier to systematically catch syntax errors and produce meaningful errors.