Langlang (the language language) is a small programming language meant to make it easy to write parsers for new languages. It compiles to JavaScript, and some day maybe other languages if I get around to implementing them.
See the getting started page for details on the language.
Running the compiler on a langlang file will generate a library that can be included by your code:
$ python langlang.py myfile.ll
Writing output to ./myfile.js
$ node
> var parser = require('./myfile.js')
> parser.add('1 + 2')
{ left: '1', right: '2', _type: 'Add' }
If you want a stand-alone "binary" for testing purposes or whatever, you can specify an parser that will take its input from stdin and print the output as JSON:
$ python langlang.py myfile.ll --stdin add
Writing output to ./myfile.js
$ echo "1 + 2" | node myfile.js
{
"left": "1",
"right": "2",
"_type": "Add"
}
There are, like, a billion parser generators. Why does this exist?
I needed something that allowed for fine-graned controls on the error messages and recovery patterns, because most parser error messages suck and we as a society can do better.
If this "langlang" is so great, why isn't it written in itself?
One day it will target Python and this will happen. One day...
This doesn't have the "fine grained" features you said earlier in the FAQ. What gives?
It's a work in progress. Even the syntax changes a couple times a week.
How can I contribute?
My emails is in my bio. If anyone's interested in contributing, send me a message and I'll add contribution directions. I haven't yet because they take time and I'm guessing nobody will want them.
I want changes, but I don't want to contribute.
File an issue and I'll see what I can do.