Handling Errors in IcedCoffeeScript

---

malgorithms 22 hours ago | link

I think the great thing about IcedCoffeeScript + the ESC library is how it fits into more complicated flow logic, while allowing easy refactoring. Max doesn't really get into the otherwise impossible examples in his post.

Even if you're firing off RPC's awaiting in the middle of a loop or switch statement, you can move logic around just by shifting individual lines. Consider how simple this looks:

    for user_id in user_ids
      await load_user user_id, esc defer user
      user.whatever()
      # etc., with user

It takes a lot of getting used to but it is really, really nice to be able to surround error-prone code in what amounts to a try/catch, with a single catch that can (like Java) branch based on error type.

For a second it feels like sync programming.

As cool as ICS is, I strongly prefer the yield syntax to the await/defer syntax. It's a single keyword and it's much more obvious where the value is coming from. Plus, it's going to be everywhere soon.

At keybase, do you guys intend to continue with ICS or eventually move to ES6?

    errTo = require 'errto'

    readDirectoryAsync = (catalog, cb) ->
        fs.readdir catalog, errTo cb, (filenames) -> # Notice no 'err' argument.
            # Error check is done automatically in errTo and cb(err) is called on error.
            # Subsequent code is called ONLY when no error given.
            callback null, filenames

    # Express sample.
    app.get '/', (req, res, next) ->
        readDirectoryAsync __dirname, errTo next, (filenames) -> # Use Express error handling by calling next(err)
            res.send filenames

    function *() {
      try {
        let body = yield request("https://x.io/");
        return yield request("https://x.io/?q=" + body.hash);
      } catch (e) {
        // do what you like with the error
      }
    }