Allow locating disabled buttons or fields

[wapcaplet]

Commit a8f970d appears to have removed the ability to locate a disabled form field or button. For example, given this HTML:

<input type="submit" value="Save" disabled="disabled" />

Then an attempt to merely find the button fails:

page.has_xpath?(XPath::HTML.button("Save"))
=> false

I don't think the button xpath should be contingent upon the button being enabled. The button exists, and is visible to the user. I understand that for the purposes of interactivity (clicking/activating it), we may want the button to be "nonexistent". Perhaps it would make sense to include a second parameter to the #button method?

def button(locator, must_be_enabled=false)
  if must_be_enabled
    ...
  else
    ...
  end
end

Then one could do:

XPath::HTML.button("Save", true) # nil if the button exists, but is not enabled
XPath::HTML.button("Save", false) # returns enabled or disabled button

Or maybe separate methods specifically for locating enabled buttons/fields:

def button(locator) ...
def enabled_button(locator) ...

(I think I prefer the second solution; positional boolean parameters make for poor readability in client code).

Or is there another way to locate disabled buttons/fields that I'm overlooking?

I'd be happy to implement either of the above solutions and submit a pull request, but thought it best to discuss beforehand.

[jamesarosen]

I strongly agree. I'd like to test that certain fields and buttons start off as disabled, then become enabled as the user changes other data.

I had even written the following Capybara extensions before I realized that XPath was doing the filtering automatically (and without an override):

Capybara::Node::Element.class_eval do
  def enabled?
    !disabled?
  end

  def disabled?
    self[:disabled] == 'disabled'
  end
end

Capybara::Selector.class_eval do
  all[:button].filter(:enabled)  { |node, value| value ^ node.disabled? }
  all[:button].filter(:disabled) { |node, value| value ^ node.enabled? }
end

None of the other XPath selectors take multiple arguments. Thus, I recommend either

1. multiple methods: field (disabled or enabled), enabled_field, and disabled_field
2. the locator argument can be a Hash: field(:text => 'foo', :disabled => true) or field(:id => 'bar', :disabled => false)

[caseyprovost]

I agree with jamesarosen, the ability to find a disabled element is crucial to testing the UI of our application. There are many reasons why someone might want to find disabled inputs for testing.

Some examples:

• an input starts of disabled, but after some js validation becomes enabled
• an input is disabled, but when a permissions is granted it becomes enabled

I am a fan of the following tweak to the field finder api:
field(:id => 'bar', :disabled => false)

[Darkvater]

I agree. Just ran into this issue. Even if a field is disabled you might want to assert its value. Not being able to select the field at all is not optimal to say the least.

[MarceloCajueiro]

I agree with you guys.

I'm having the same issue here. I'm updating the capybara to 2.0 and now I can't find disabled fields.

So @jnicklas what do you think? Do we can do a patch?

[MarceloCajueiro]

I did a temporary monkey patch (shame on me) to we use until we found a solution:

module XPath
  module HTML

  protected

    def locate_field(xpath, locator)
      locate_field = xpath[attr(:id).equals(locator) | attr(:name).equals(locator) | attr(:placeholder).equals(locator) | attr(:id).equals(anywhere(:label)[string.n.contains(locator)].attr(:for))]
      locate_field += descendant(:label)[string.n.contains(locator)].descendant(xpath)
      # locate_field[~attr(:disabled)]
    end
  end
end

[jnicklas]

We no longer filter away disabled elements. Capybara 2.1 implements this as a filter instead of hardcoding it in XPath itself.

[jamesarosen]

Thanks, @jnicklas :)

[MarceloCajueiro]

👏