>>73305118I don't know that I really like the jump-if-conditional design. It's commonplace in assembly (at least Z80, ARM and x86), but part of the fun of AWA5's conditionals was finding interesting ways to abuse the skip-one-on-conditional nature of the language. As long as you have labels, you can replicate the jump-if-conditional behaviour anyway, but the other option is more interesting in certain cases.
I don't think free text labels are as bad as you say. Clobbering function names is an existing issue in any language, so the only real issue is local labels like in for loops. x86 and z80 (maybe ARM, but I can't remember) have local labels which are prepended with a . These labels are basically syntactic sugar for the assembler and don't exist in the executable's symbol table, so they aren't exported or seen when linking libraries. While this could be added to an AWA preprocessor, you'd have to also find a way to have these defined in awatalk code. For that matter, how would free-text labels be defined in awatalk? A different awatism followed by an AWASCII encoded string and an end byte?
I don't understand what you meant by this:
>There are various ways to fix this, the easiest one is to allow textual labels for all control flows (there is enough space to duplicate all of them in 5 bits, eventually) and make it similar to actual assembly.Any chance you could reword it or expand on what you mean?
Regarding straying from AWA5. I don't necessarily mind, but I personally don't want this to just become "assembly but with a weird memory model". Also, any chance you could provide some insight into how you plan to do graphical operations?