LuaJIT 3.0 proposed syntax extensions

(github.com)

151 points | by phreddypharkus 8 hours ago

27 comments

  • Ardren 5 hours ago
    > For compatibility with other computer languages, the following classic Lua operators can be written in a more customary syntax:

    Why though? What does changing `and` to `&&` actually achieve? Were people confused?

    Changing the syntax seems very surface level. It's not actually fixing any problems, just making Lua no longer look like Lua. It's not going to help anyone write/learn Lua. It will make everything more complicated as there are now two ways to do everything.

    This feels like adding braces to Python because you don't like indenting your code.

    • nullpoint420 4 hours ago
      > This feels like adding braces to Python because you don't like indenting your code.

      Now this I can get behind...

    • simonask 2 hours ago
      Ruby has both kinds of operators as well, and it's fine. The thing in Ruby, though, is that the English logical operators have lower precedence than the symbolic logical operators, so you can use them in place of parentheses. Sometimes that's confusing, other times it can be used to make code very readable.

      In general, I would expect symbolic operators to be desirable in complex boolean expressions, because "loud punctuation" stands out among English words when reading the code.

    • doophus 3 hours ago
      > Why though? What does changing `and` to `&&` actually achieve? Were people confused?

      Also consider AI, that has a greater training base of JavaScript than Lua. So making Lua look more like JS, should improve output and reduce mistakes.

      • lionkor 2 hours ago
        No, don't consider AI.
      • boxed 1 hour ago
        AI has greater training on Python, which uses `and` and `or`, and it has absolutely no issue keeping that straight.
  • Heliodex 7 hours ago
    A comment <https://github.com/LuaJIT/LuaJIT/issues/1475#issuecomment-47...> has already been made on the issue regarding the ternary operator, recommending `if x then y else z` over `x ? y : z`. This is exactly how it's done with if-then-else expressions in Luau <https://luau.org/syntax/#if-then-else-expressions>, another language compatible with Lua, and makes it a ton easier to nest (especially with elseif) and I believe still easier to read than `y if x else z`.
    • noelwelsh 1 hour ago
      Exactly. I don't understand why people think the ternary operator is needed when you can just make `if` an expression instead of a statement. It there is no new syntax to learn and `if` just becomes more useful.
    • mjcohen 7 hours ago
      The ternary operator is easy to nest if you put each clause on a separate line. Then it looks just like nested if-then-else.
      • edoceo 6 hours ago
        I love the ternary operator as much as anyone. But dang if it doesn't get hard to read when there is are a few, nested even.

        Does that operator compile to faster assembly that if I make the same logic with verbose `if` logic? Is that a language specific outcome?

        • gnubison 5 hours ago

              cond1 ? res1 :
              cond2 ? res2 :
              cond3 ? res3 :
              or_else_res
          
          If they are truly nested, then that is confusing. But if you have an if-else chain, then it can be quite readable.
          • a-french-anon 1 hour ago
            Or they could just ask granddaddy for advice:

              (cond (cond1 res1)
                    (cond2 res2)
                    (cond3 res3)
                    (t     else_res))
            
            =)
          • shevy-java 5 hours ago
            I find that so much harder to read compared to if/else or case/when in ruby.

            The ? is basically an attempt to use fewer if/else, at the cost of condensed if-else like structure. I always need to look at both parts after the ? whereas in a single if or elsif I don't. case/when in ruby is even better here e. g. regex check:

               def foo(i)
                 case i
                 when /^cat/
                   handle_cats
                 when /^dog/
                   handle_dogs
            
            (I ommitted the "end"s here to just focus on the conditional logic.)
  • camgunz 2 hours ago
    One of the interesting things about Lua is because they don't really maintain compatibility between major versions, there isn't a huge ecosystem, and as a result there's less friction against making your own, slightly incompatible version. When you add on the simplicity of implementing the language, it's created a really diverse set of lua-alikes. Weird (and cool) for a language to have a diverse ecosystem of implementations, but not necessarily libraries.
  • omoikane 6 hours ago
    Lua 5.3 (2015-01-12) added the bitwise operators:

    https://www.lua.org/versions.html#5.3

    https://www.lua.org/manual/5.3/manual.html#3.4.2

    Looks like LuaJIT is catching up, but calling these "syntax extensions" is confusing. Is the intent to hold LuaJIT fixed against some earlier Lua version (I guess 5.1) and adopt newer syntax piecemeal?

    I welcome the compound assignment operators. Playdate's version of Lua also has that extension.

    • orthoxerox 1 hour ago
      LuaJIT is an involuntary fork of 5.1. It already had various extensions that conflicted with the 5.2 implementation of the same features, and Mike Pall made it clear on the mailing list he wasn't going to change how LuaJIT worked.
  • childintime 45 minutes ago
    > local gauge = direction == "up" ? count + 10 : count - 10

    local gauge = count + (direction == "up" ? 10 : -10)

    I imagine these changes make the original Lua adepts think their training wheels have come off. The language now looks like any other. That's a good thing to me, and it will help with the adoption of the JIT, but the whole language could have been syntax modernized as a result. But.. when the work is done someone else can fork it into something independent from its Lua roots.

    From that perspective the conditional operator seems defensible, where it would be feature creep otherwise, as it is generally unloved elsewhere.

  • ianm218 6 hours ago
    Tangently related but I’ve been deep in Lua recently working on a rust implementation that supports Lua 5.1-5.5 in one Rust Binary https://github.com/ianm199/omnilua.

    My ultimate goal was to support LuaJIT in Rust as well but this does not make it easier.

    • valorzard 2 hours ago
      Also, one issue I have with this repo is that, since so much of it seems to use Claude, as an actual human I struggle to read and parse any of the information.

      For example, what’s the performance like?

    • lifthrasiir 4 hours ago
      Oh wow, seriously, I always thought Lua should have been like this. The 5.1/5.2/5.3+ split was so painful.

      > My ultimate goal was to support LuaJIT in Rust as well but this does not make it easier.

      I think you could stop right before the syntax extension.

    • genxy 4 hours ago
      This is amazing! Can a program call across versions? Like could we take a Lua 5.1 codebase and upgrade only a portion of it at a time to a new Lua version?
    • valorzard 2 hours ago
      [dead]
  • pansa2 7 hours ago
    So is LuaJIT resuming active development after a decade or so of only maintenance? Great!

    A lot of these changes make sense (although some of them are a bit too TIMTOWTDI for my taste) - but perhaps LuaJIT 3 would benefit from a change of name as well? Certainly with all these changes, it would be more like a separate language than merely a JIT-compiled version of Lua.

    • orthoxerox 1 hour ago
      A bunch of them are from Luau, the Roblox fork of Lua and the dialect most young programmers know. Adding them to LuaJIT will make it easier to write for both Zoomers and AI agents, who have been exposed to a lot of Luau code.
    • 201984 7 hours ago
      >TIMTOWTDI

      What on earth is this supposed to mean?

      • Twirrim 7 hours ago
        There Is More Than One Way To Do It.

        That takes me back a bit. It's a perl-ism. I used to think it was a great design feature but I've come to strongly prefer "There should be one way to do it, and it should be obvious"

        • 201984 7 hours ago
          Interesting, thank you.
      • matheusmoreira 7 hours ago
        There is more than one way to do it.
  • ricardobeat 7 hours ago
    I see JavaScript.

    Some of these really look like QoL improvements. I'm not convinced ternary statements are an ergonomic improvement in particular. The examples given don't make a compelling case, 'visually tidy' is not the same as readable.

    • nine_k 7 hours ago
      Worse, I see C (as in ! or &&), and Perl (as in manifestly more than one way to do it).

      There are real improvements though, such as ?. and ??= that help with default-nullable everything.

      Ternary is very useful, but it I'd rather see it implemented idiomatically:

        pos += (if forward then +1 else -1)
      
      Structural pattern-matching could be fantastic, but no syntax is suggested.
    • fluoridation 1 hour ago
      Lua to me always felt very JavaScripty, just with a different syntax.
  • HexDecOctBin 2 hours ago
    It might have been better to publicly document and stabilise the LuaJIT bytecode, which would allow people to then come up with whatever syntax they wanted in their own custom frontends.
  • 3eb7988a1663 7 hours ago
    Never will I understand ternary operators. As soon as you introduce it, some chuckle heads want to use them everywhere. Worse if the syntax allows nested ternarys. I guess it keeps the language open for code golfing, but it otherwise seems like redundant syntax that at best saves a few characters.
    • demilicious 7 hours ago
      That’s why “if” should just be an expression
      • matheusmoreira 7 hours ago
        This is the best answer in my opinion. Ternary is just sugar for an expressive if. LuaJIT seems to be focusing on adding new syntax though, maintainer might not be amenable to updating existing semantics.
        • wavemode 6 hours ago
          I don't think if-expressions have to affect existing semantics. Basically, in the parser you would have two different kinds of AST nodes, one for when the `if` keyword is encountered in statement position and another for when it's encountered in expression position.

          Right now, `if` in expression position is just a syntax error ("unexpected symbol")

          • Joker_vD 4 hours ago
            Well, I believe there could be some complications with parsing related to the fact that Lua grammar doesn't really requires semicolons between the statements.

            But other than that, yeah, detecting "if" in the expression position is pretty unambiguous. No idea why most languages went with "cond-expr ? then-expr : else-expr" bracketed syntax instead.

            • _flux 3 hours ago
              Surely the most likely explanation is familiary from C?

              But e.g. ml-family languages (like OCaml, F#, Haskell) and Rust just have the *if* expression that has a non-void value. If your language accepts expressions as statements (most do?), then I think that should just be compatible out of the box.

              • Joker_vD 2 hours ago
                Yes, but why C had that syntax? Oh, right, because it didn't use if-then[-else]-end for the conditional statement, and reusing if(cond)[-else] with prohibited braces would be awkward.

                Oh, and Lua most famously does not accept expressions as statements. Which, now that I think of it, would actually evade most of the parsing complications.

      • NuclearPM 5 hours ago
        Yep. Everything should be.
    • 201984 7 hours ago
      Lua basically already has ternary operators anyway since "and" and "or" short circuit. I also don't see the need of adding additional syntax for it.

        local x = condition ? value_a : value b
        local x = condition and value_a or value_b
      • matheusmoreira 7 hours ago
        > The classic Lua idiom a and b or c has a pitfall when b is nil or false: then c is returned, even when a is truthy.

        > E.g. true and false or 42 returns 42, whereas true ? false : 42 returns the (expected) false.

    • Gualdrapo 7 hours ago
      I guess for the JS case it makes sense to be able to shave a few characters for file shrinking purposes, but generally I'm more biased to code clarity and "self-explainability"
      • NuclearPM 5 hours ago
        That’s what compression is for.
    • hiccuphippo 7 hours ago
      I find it most useful in languages that have non-mutable variables and you want to avoid a mutable variable or an extra function when the value comes from a simple condition.
  • mingodad 2 hours ago
    For others interested in alternative syntax to the Lua VM/API sometime ago I've created LJS https://github.com/mingodad/ljs and also https://github.com/mingodad/ljsjit, I've also included an utility lua2ljs program based on the Lemon parser and re2c that convert Lua scripts to LJS with line by line synchronization https://github.com/mingodad/ljs/tree/master/lua2ljs, to test it I've also translated a few non trivial projects (https://github.com/mingodad/ZeroBraneStudioLJS , https://github.com/mingodad/raptorjit-ljs, https://github.com/mingodad/snabb-ljs, https://github.com/mingodad/premake-core/tree/ljs, https://github.com/mingodad/CorsixTH-ljs).

    I'm proud of it and thankfull to the Lua/Luajit projects.

  • bawolff 7 hours ago
    += and ..= are things i find i'm constantly missing in lua.

    Personally im a fan of introducing ternaranary operator in lua. Everyone uses `x and y or z` as a ternanary which i find way more confusing than ?:

    • linzhangrun 3 hours ago
      Lua pursues "simplicity, purity, and simplicity." So... too much syntactic sugar is unlikely
  • matheusmoreira 7 hours ago
    Looks like LuaJIT is really going to fork away from Lua this time. After these changes, it won't be a compatible Lua 5.1 implementation anymore, it will be a new language.

    So shouldn't it have a new name?

    • orthoxerox 1 hour ago
      Why won't it be compatible? Any code written in Lua 5.1 will run on LuaJIT.
    • ulbu 4 hours ago
      well, it doesn’t say Lua5.1-JIT
    • a_t48 7 hours ago
      It could be opt in.
    • sourcegrift 6 hours ago
      Are there any rough estimates on popularity of lua implementations? At this point it feels lua means luajit
      • latenightcoding 6 hours ago
        not even close, because there are a lot of places where you can't run LuaJIT
        • tuvix 5 hours ago
          Where can you not run LuaJIT? Genuinely curious
          • Boxxed 4 hours ago
            Wasm and platforms like iOS and Nintendo Switch (I think).
          • extrememacaroni 5 hours ago
            anywhere that does not allow self modifying code such as app stores.
        • Dylan16807 4 hours ago
          LuaJIT is not just a JIT, it also includes high speed interpreters for x86, Arm, and more.
  • pseudony 4 hours ago
    Seems like a bad idea to actively diverge from Lua, hostile even, especially without at least a clear change of name.
  • linzhangrun 7 hours ago
    I thought luajit had completely stopped feature updates
  • spankalee 3 hours ago
    They shouldn't add the ternary operator, it keeps `?` from being usable on it's own for safe navigation and requires the ugly `?.` operator, like `a?.[b]` or `f?.()` instead of `a?[b]` or `f?()`.
  • larrry 7 hours ago
    I would love to see all of these come to LuaJIT (and love2d to support the new version too). It’s nice that Lua is simple, the syntax changes should hopefully make Lua code even simpler to read too
    • Rohansi 7 hours ago
      > It’s nice that Lua is simple, the syntax changes should hopefully make Lua code even simpler to read too

      But which Lua?

      Lua as implemented by LuaJIT is a fork of the language at this point. It's not fully compatible with PUC Lua (the reference implementation) and LuaJIT does not support features from the latest Lua version.

  • kibwen 6 hours ago
    Please don't, inscrutable bitwise operators are an accident of the past even in systems languages, let alone in a scripting language. I'm not against infix operators for bitwise operations, just please spell them out with keywords rather than giving them sigils.

    Likewise, going from `and` and `or` to `&&` and `||` would be a dispiriting regression. This is something that Zig got right.

    • astrobe_ 2 hours ago
      You can have them in C/C++ at least [1]

      [1] https://en.cppreference.com/cpp/language/operator_alternativ...

    • Dylan16807 4 hours ago
      What kind of person understands and needs bitwise operators but can't easily remember & | ~ and the arrows for shift? It's very little information.

      The part I'd call a hassle is the different kinds of right shift but you have that same hassle if you use keywords.

      I like using the and/or keywords for logical operations. Now let's make bitwise look significantly different from that.

      • Mond_ 2 hours ago
        It's not about having to remember them, it's that you shouldn't waste these short single symbols on operations that are only rarely used.

        This stuff (especially the ternary) are a step backwards. There is just no reason to waste | on a bitwise or that gets used at 1% of the frequency of the standard or. In the future you might have a better use for it (pipeline syntax, sum or union types come to mind in other languages).

        I dislike basically everything about these syntax extensions.

    • gautamcgoel 5 hours ago
      Doesn't Zig also have bitwise operators?
    • JSR_FDED 6 hours ago
      The btiwise operators library doesn’t go away
  • flumpcakes 3 hours ago
    Is LuaJIT still based on Lua5.1? I wonder why they haven't followed the language spec up to Lua5.5.
    • radiator 2 hours ago
      Because they don't like the changes, to put it mildly.
  • le-mark 7 hours ago
    I’m confused I thought Mike Pall left luajit and Laurence Tratt took over as maintainer?
    • dang 6 hours ago
      Mike Pall is to LuaJIT as PG is to Hacker News.

      Edit: meaning he can come back anytime.

  • JSR_FDED 6 hours ago
    What’s the Lua/LuaJIT story these days for bundling up all the scripts of an application into a single file? Is there a way to do the super convenient go-like thing?
    • orthoxerox 1 hour ago
      There was a Lua[JIT] fork called Idle that seems to have fallen off the face of the Earth that did exactly that: it would take a small stub program, a runtime library and all the scripts and package them into a single PE/COFF binary that would read itself when run.

      Love2D does it as well: zip -9 -r SuperGame.love . cat love.exe SuperGame.love > SuperGame.exe

      This doesn't work with ELF files, though.

    • zdragnar 5 hours ago
      There's a bunch of options from a Google search, but embedding it in a thin C program and building that with https://github.com/jart/cosmopolitan would be a pretty go-like experience, I'd think.
    • gucci-on-fleek 5 hours ago
      I personally use a hand-written C wrapper program (which is not much more than a dozen lines long), and then embed the Lua scripts using objdump. This isn't quite as easy as Go since cross-compiling C programs is often somewhat tricky, but Lua is very portable and has zero dependencies, so it's usually not too hard.
  • JSR_FDED 6 hours ago
    Cool to see this - ergonomic syntax will make it easier to recommend Lua. Hope the PUC team aligns with this.

    Also, I love this kind of pragmatism:

    > Exponentiation assignment a ^= b has been deliberately omitted to avoid a predictable pitfall: this is how xor assignment is written in most other computer languages. Also, a syntax for exponentiation assignment is rarely asked for.

    A ‘defer’ for closing files or deleting temp files at the end of a script will make life more enjoyable.

  • yxhuvud 1 hour ago
    In aggregate this looks like a godsend, but there are some examples (like foo?.:method) that looks atrocious.
  • sourcegrift 6 hours ago
    What are some pragmatic embedded scripting languages of choice these days if one has to consider:

    1) Ease of learning, ideally minimal deviant behaviour (eg i consider lua tables to be a new concept in itself)

    2) Reasonably fast. Not as much as lua jit but even half would be good enough

    3) Mature

    4) Has Rust bindings

    • NuclearPM 5 hours ago
      Lua. Lua tables are easy and awesome. My hobby language unites Lua tables with functions too.
  • shevy-java 5 hours ago
    Lua has a lot of useless syntax. For instance, the "then". I have been using ruby and python for many years. Lua is living in the old age here.

    That's just one example of so many more. I get that lua occupies a useful niche with its focus on embedded systems, but lua is not really a well-designed language in general. JavaScript has a similar problem.

    • xonre 2 hours ago
      For readability, `then` allows splitting with newlines very long conditional expressions, without having to wrap the condition in parentheses:

        if x + y + z > a
          or verylongconditionalhere ()
          or anotherverylongconditionalhere ()
        then
          ...
      
      after `if` and `elseif` the parser simply goes on until it finds `then`.
    • Dylan16807 3 hours ago
      Python spells "then" as ":"

      In Ruby you can choose between "then" and a newline.

      This is very pot calling the kettle black.

    • mjmas 2 hours ago
      English too
    • modulared 4 hours ago
      [dead]