MetaES – JavaScript in JavaScript interpreter | Hacker News Reader

MetaES – JavaScript in JavaScript interpreter

metaes.org

51 pointsbartqk31 comments

Previously: https://github.com/mozilla/narcissus

Yep. But it went nowhere, as far I know.

IIRC I believe tailspin was based partly off the Narcissus codebase. I think MetaES is better because it gives you a debugger whereas Narcissus isn't pausable (last I checked). Additionally MetaES is written in plain ES5, which I think is another plus over Narcissus in terms of general usability. Narcissus' focus was mainly on prototyping language features and I get the impression that compatibility was never a concern.

Awesome work! I particularly like the code completion.

Would this be quick enough to integrate into an editor like atom.io? For variable and function completion

That's the plan, although it may take few weeks or months to create something solid :)

I think there's definitely value in this kind of JavaScript-in-JavaScript type of dev utility. The main advantage is cross-browser dev utilities - which brings us closer to a cross-browser dev environment.

Browsers have historically opted for hiding runtime management (debugging, etc) behind myriad APIs, most of which can't be accessed from the running program itself. I would love to see a cross-browser way to hook into these sorts of things without extensions or at least with standard APIs.

Overall this would be much better for us as coders. And this would open the way for all sorts of cool tools. Consider V8's live code reloading functionality. But imagine if it worked cross-browser and could be called from the program itself. Your webserver could send a signal over websocket, and you could perform hot-swapped deployments of your webpage, and all connected browsers would be updated. The browsers could be customers', or they could be inside that mobile device that's a pain to debug.

It would also be much easier to prototype new dev tools and ways of coding. For example, programming against Chrome's extensions APIs (as great as they are) is somewhat limiting - last I checked you lack direct access to the embedded CodeMirror editor for example. There are a lot of security-focused restrictions. So people workaround this. They rewrite editors and file managers from scratch so they can get the desired flexibility (lighttable, brackets, atom, etc).

Personally I love to see projects like MetaES because I feel like they are inching us closer to something that could be the next step.

Indeed MetaES may be a way to pave a path to new kind of development. Bret Victor's work, LightTable, Swift Playground and others proved that small improvements are possible. I believe though, that next bigger step requires new fundamentals and metacircular interpretation plus smart IDE and partial evaluation may be very interesting.

I don't know honestly what's going to pop up along the way, but with MetaES we can get smart completion and better debugging experience in popular editors for sure, it just requires more work and care.

hey Bartqk, I'm working on something I think you might be interested in, but I can't find your email in your profile. Would love to chat. My email is cammarata.nick@gmail.com

Any hopes of a Vim/NeoVim Plugin?

certainly, after there's a better API for completions and evaluations created. Will be easy to write plugins then.

and thanks! ;)

Reminds me of https://github.com/int3/metajs

I like this kind of stuff.

A particularly interesting take on this is tailspin, although the repo appears have been taken down: https://github.com/wthimbleby/tailspin That one offered a reversible JavaScript debugger written in JavaScript, which I thought that was very impressive from a technical standpoint.

It is very similar. But MetaES tries to be production ready and reliable. As far I rememember, tailspin wasn't metacircular, but just an interpreter that interprets some kind of bytecode created from JavaScript source.

Yeah, it seemed designed for educational examples, and worked within in a VM. But I couldn't find a way to easily interop with the external JavaScript environment. It was well-tested though and passed the vast majority of test262.

MetaJS has some open issues - some related to the fact that it's intended to let you step through code, which I think makes certain interop scenarios difficult (let's say for example native code calls MetaJS code and expects a return value - it's not possible to set a MetaJS breakpoint there because the native runtime isn't paused).

I would be curious to know how MetaES fares under something like test262. I think a metacircular interpreter that does well in an extensive suite like that could indeed be a proper, "safe alternative to eval" (as suggested by another commenter).

MetaES was tested using test262. Passed most of the cases.

https://github.com/metaes/metaes - "MetaES was tested using test262 test suite. Runnable test suite is currently not added to the repository."

You're right about breakpoints while code is called by native functions. That's something that should be discussed more.

Ah, sorry I missed that! Very nice!

Great job.

Do you know this interpreter ? (in D)

https://github.com/higgsjs/Higgs

It's one of the most advanced I found (outside browsers).

If you want to get rid of the "switch/case", you can look at threaded code interpreters.

http://www.complang.tuwien.ac.at/forth/threaded-code.html

You can find useful ressources in Forth about it.

I write a stack-based language in javascript. Each statement is a closure around the arguments of the expression stored in a flat array. The instruction pointer is an index in that array. It is partially compiled.

I also use function object constructor for a kind of FFI.

There is also good potential with asm.js to reach near-native performances.

Thanks. Didn't heard of Higgs, but looks interesting. You could use JITting in MetaES as well, you have all information about execution during runtime, so why not? That was one of my ideas for demo, but this one is quite heavy ;)

Where's your language's website?

https://github.com/vmorgulys/crabble/tree/master/js/object/m...

It is far less advanced than yours. It's inspired by Forth (a mixture of interpret and compile "word"s).

I'm doing that right now:

  let print foreign console.log
  print "hello"

Every "word" is a function receiving a stack and returning a stack. Whitespace can be replace by a "|".

Each line is evaluated from right to left. Some words are used to change the state of the compiler and the compiler is a state machine (words are events).

The game is to find a minimal set of words to express high-level features like lambdas (from a word "jump", "until" or "instruction-pointer" for example).

It's cool to program :-)

This reminds me of a javascript weirdness I found. When passed as `this`, primitives (string, number, boolean) get wrapped in their classes.

You can get the 'Number' function with:

    Number = (function(){return this;}).call(1).constructor

nice trick. You can, but you can't reference `Number` directly, that's the point. `Number` is a part of standard library, so being in JS it's possible use hack that you've shown. But you won't find a reference to let's say `console`, because it's not a part of ECMAScript.

or just `1..constructor` -> function Number().

> or just `1..constructor` -> function Number().

Ah ha. much simpler :)

Nice one - during my studies I was implementing the The Spineless Tagless G-machine in Haskell - @bartqk - maybe you can find this interesting / inspiring: http://ow.ly/OBXQ6. Or in general search for any paper by Simon L. Peyton Jones. (like this one: http://ow.ly/OBY0K)

This could actually pretty damn cool as a method of using ES6 features and supporting browsers that don't yet offer it. Of course at a serious performance penalty I'm sure.

If you don't need to dynamically evaluate the code, compiling ES.next code will be more performant, e.g. using Babel.

Yeah, it would be slower. But you could use some kind of JITting while the interpreter goes and create appropriate ES5 code. That's possible next step.

Similar project: https://benvie.github.io/continuum/

Yeah, much older project. Looks like it intereprets some kind of IR code, it's not the case in MetaES, it just executes AST. I don't know either if it allows interoperate with surrounding interpreter. MetaES interops native code on purpose.

This is pretty cool. Maybe can be useful as a safe alternative to eval()?

Could be :) If you accept that it's slower. Benchmarks would show the difference.

Reply on news.ycombinator.com