Initialization

[koresar]

• Kill Babel, write in ES5 instead
• Use jshint instead of eslint
• Implement compose and core with perf optimizations
• Implement check-compose and use it in compose
• Drop node <=0.12 support

[danielkcz]

Huh, jshint ? Well I see we are moving backwards, good luck then :)

[koresar]

Unfortunately, recently I don't have enough free time to fight babel, transpilation, browser features support, performance and the dozens of other bugs we had in the past.

We'll move to >=ES6 and eslint same day all browsers support ES6. :) Currently, none of them does http://caniuse.com/#search=ES6

[koresar]

Less is more. KISS etc

[danielkcz]

Well do you realize that this is never ending battle? Once all browsers implement ES6, there will be ES7 & ES8 with absolutely same story.

Currently support is missing only for ES6 modules and it wont be there for a very long time. You can easily write without ES6 modules and in that case you already have full support. Unless you are expecting IE 11 to add support too...

As I said, good luck, I wont be helping with prehistoric stuff, that's not interesting enough for me.

[koresar]

Hi.

Sorry coming back that late. Life is busy recently. Thanks for the feedback.

There is a reason lodash v4 is written in ES5, and lodash v3 is in ES3. It's ease of development, deployment, bundle size, and performance. What I want to achieve - is less time waste on new cool stuff.

Do you have any concrete proposals?

Btw, do you find the jest good choice?

[danielkcz]

I personally don't use Lodash anymore. There isn't that many functions that I would really need. I think that stamps can get rid of that dependency too easily. And btw there is also lodash for ES6 in case you missed it.

You can just use the Bublé and get transpilation without hassle really. It's a matter of 5 minutes, especially since it has been already done on stampit repo. I don't understand why would throw that experience away and went on figuring out some prehistoric path from scratch.

Not even talking about getting rid of you ESLint which means throwing away a lot of helpful rules that improve code quality.

Have you ever really tried to compare performance of pure ES5 code and transpiled ES6 one? Of course the native ES6 code is slower currently, but it will only get better.

I haven't used Jest yet, so I don't have an opinion there except that I don't like much describe / it pattern that much anymore.