The Buckblog

Yesterday’s tip demonstrated how to refactor common RJS bits into their own method. Specifically, it gave an example that created a special render_error method which you can call whenever you need to render an error.

Today’s tip is about overriding the existing render method, to add support for your own custom conditions. Specifically, render_error is nice and all, but wouldn’t it be nice to make it look more like the standard Rails render method?

1
2
3
4
5

def create
  ...
rescue Exception => error
  render :error => error
end

It’s remarkably simple, actually. To implement the above, you’d just throw something like the following in your ApplicationController:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

def render(*args)
  if args.length == 1 && args.first.is_a?(Hash)
    case args.first[:error]
    when nil then super
    when ActiveRecord::InvalidRecord
      render(:update) do |page|
        page[:js_error_bold].replace_html(args.first[:message])
        page[:js_error_text].replace_html(
          args.first[:error].record.errors.full_messages.join("\n"))
        page[:js_error].visual_effect :blind_down, :duration => 0.2
      end
    else raise args.first[:error]
    end
  else
    super
  end
end

In other words, first we check to see if the first and only argument is a hash (since that’s all our new feature uses). Then, we test the value of the :error key, calling super whenever we want to delegate to Rails’ render method. We then compare the error raised to some list of errors that we want to handle specially, and do the appropriate work for each.

You could also make it so that it accepted the error directly, as the first argument (render(error)).

You can use this technique anywhere you find yourself rendering the same kind of thing in multiple places. Errors are a good example, and are (frankly) where I use this technique the most.