Login
You're viewing the front-end.social public feed.
  • Oct 7, 2026, 8:38 PM

    having functions would also open up the possibility of transpiling to, say, C, in a readable manner. (I could technically do this now, but it would just be a big main() function with a lot of labels and gotos.) but it would mean that the language would less directly map to assembly, which seems undesirable. i keep going back and forth (no pun intended) on this

    💬 2🔄 0⭐ 0

Replies

  • Oct 8, 2026, 5:05 AM

    @aparrish omg has nobody made a programming language, or os kernel, or architecture, named "back" yet? to take advantage of that pun

    thats right my handheld console's microcontroller runs back and forth

    would the back language be described in backus naur form

    💬 0🔄 1⭐ 2
  • Oct 8, 2026, 4:52 PM

    @aparrish Perhaps one possibility is to introduce a pseudo-op that says e.g. that a particular label "foo" refers to a function? Something akin to
    .global main
    .type main, @function
    main:
    ...
    ret
    that is generated by GCC for a target platform that uses ELF.

    💬 0🔄 0⭐ 1
  • Oct 8, 2026, 5:02 PM

    @tkchia i would consider this for sure! though i don't really like the idea of adding annotations that are only relevant for certain backends, at least not this early in the design/implementation process :/

    💬 0🔄 0⭐ 0