[rspec-users] Not seeing an RJS change

Zach Dennis zach.dennis at gmail.com
Mon Sep 29 19:33:36 EDT 2008

On Sun, Sep 28, 2008 at 11:59 PM, Nick Hoffman <nick at deadorange.com> wrote:
> On 2008-09-28, at 17:20, Nick Hoffman wrote:
>> 431         it 'should hide the map filter errors <div>' do
>> 432           do_xhr @xhr_params
>> 433           response.should have_rjs
>> 434 #          response.should have_rjs(:hide, 'map-filter-errors')
>> 435         end
> I just discovered that I need to pass :chained_replace_html to #have_rjs ,
> like so:
> response.should have_rjs(:chained_replace_html, :hide, 'map-filter-errors')
> However, after googling around and looking through the source for
> #assert_select_rjs , I have no idea what :chained_replace_html does, nor why
> :hide is insufficient. Would someone mind enlightening me?

The problem has to do with the javascript that is produced by Rails.
You can do things in RJS in one of two ways.

Way #1:

Way #2
  page.hide :some_id

Based on which way you go Rails will generate Element.hide('some_id')
or $('some_id').hide() IIRC. assert_rjs uses regular expressions to
match against the generated javascript. The fugliness that are these
regular expressions require that two different patterns exist to match
against the generated javascript correctly.

Having chained_replace_html is a hack to separate the regular
expressions based on if you used Way #1 or Way #2.

I have stopped relying on using RJS selectors to test against
generated JavaScript. I have also stopped relying on RJS for the most
part, and instead am doing a lot of UJS and JSON (thx Mark VanHolstyn
for enlightening me on this). I'm not saying that you shouldn't do
this, but that I've hit the problem you're facing, and I don't think
using regular expressions from the bowels of Rails for asserting
against generated javascript is the way to go,

Zach Dennis

More information about the rspec-users mailing list