ES6 Support Via Babel

[kkirsche]

Hey Padrino,

I wanted to check with you guys and see if there was any interest in automatic javascript transpiler from ES6 -> ES5 via the babel-transpiler gem. I wanted to check though if this was something of interest to you before beginning to work on this type of a feature though

[dariocravero]

Hey @kkirsche that's always welcome :). ES6 is the future and we should embrace it. Is this transpiler working on-the-fly? Thanks, Darío

[kkirsche]

@dariocravero it is not an on the fly but I figure it or at least babel-core would provide the ability to extend the core functions to do it on the fly. Worst case though is I can work through the es6 module transpiler rails uses and use that as a basis to strip and refactor till we have something a bit thinner that'll work well with Padrino and it's goals

[dariocravero]

So would that work as a middleware that takes, say, .es6 files and compiles to .js? How would you deal with the module system? Would you pipe things say, through browserify?

[kkirsche]

@dariocravero

Middleware was my plan. I'd had thought about running it on .es6.js as the extension and serve it to the browser as .js or .es5.js. My logic on that is that while we could run it on all .js files, there is no value in wasting resources on transpiling something to itself. You could though run .js -> .js though you could do .es6 and convert on the fly to .js for transpiling and serving. Would depend on how Padrino wanted to address the issue of ES6. My thought is there are two real options:

• .js to .js — Simplest option for users and least likely to confuse people
• .es6.js to .js || .es6 to .js — Makes sure that the developer understands what they are writing and relies on them to make the correct name. With that said, this should come with performance benefits as you can avoid all files without the extension and not have to do any analysis.

My guess is we will need to use something like Browserify for module though I haven't looked at any of the ES6 shims enough to be able to say if there is a shim which would be lighterweight.

What are your thoughts?

[skade]

Wouldn't that introduce some kind of mini-asset-pipeline, which Padrino avoids doing currently?

#1028

Note that I know quite some users opting for Padrino because it doesn't care about how you handle JS.

[dariocravero]

@skade, yes, that would be an assets pipeline chore :).

[ujifgc]

Probably won't made it to the core library. I think it can be a nice padrino-recipe.