[webkit-reviews] review requested: [Bug 30997] [Gtk] Implement AtkDocument's attribute support : [Attachment 42301] Layout Test - Take 1

[bugzilla-daemon at webkit.org]

Joanmarie Diggs <joanmarie.diggs at gmail.com> has asked  for review:
Bug 30997: [Gtk] Implement AtkDocument's attribute support
https://bugs.webkit.org/show_bug.cgi?id=30997

Attachment 42301: Layout Test - Take 1
https://bugs.webkit.org/attachment.cgi?id=42301&action=review

------- Additional Comments from Joanmarie Diggs <joanmarie.diggs at gmail.com>
Given that this is my very first attempt at writing a Layout Test, I'm sure
it's not right. :-) But it's something concrete from which you can
direct/enlighten me.... I also have a few questions:

1. In bug 30964 comment #8, Jan says:

> Layout tests. We're not testing ATK APIs but testing the accessible
> characteristics of AX objects. AccessibilityController exposes some of this
> attributes that we can test in JavaScript. It's still missing a few
> implementations for Gtk though, but shouldn't be difficult to implement
those.

This to me sounds like I should be doing something with
AccessibilityController. :-) However, I don't grok why. Aren't these just more
attributes of an HTML element? Granted it's a very specific HTML element which
is likely (albeit not necessarily) the root document.... Therefore, for now, I
added the support to AccessibilityUIElement and just did an Atk role test.

2. Is there some magic foo that occurs which causes these tests to pass on
other platforms (in particular, win and mac)?

3. I assume that the way to test that the right thing is being exposed is to do
Atk calls. Right?

4. The patch which implements the attribute functionality includes docType. In
the case of local files, no doctype is returned. I haven't taken the time to
ascertain if that is intended functionality or a bug. Regardless, what's the
best way to add a layout test for doctype? (since I assume local files are
preferred/expected).

Please be kind in your review. ;-) ;-)