I am working on Accessibility for WebKit. When I load a page in WebKit and invoke inspect32 tool, I could find that navigation to the accessibility parent is not proper. I have logged a bug for enhancing WebKit accessibility heuristics. (Bug 20013). As per Jon's suggestion in that bug, I am logging this bug separately to fix this issue. I will upload the patch as soon as possible. ~ Thanks.
What specifically is not proper?
Created attachment 23493 [details] This patch contains only the changes to enhance the navigation to the accessibility parent. When we load a page in WebKit and invoke Inspect32 tool, "navigation to the parent" may lead to the object that is not exposed in the accessibility tree. Ideally, "navigation to the parent" must always leads to the parent object that is not ignored and is exposed in the accessibility tree. Made changes to fix this. Attaching the patch for this. I will attach the test case for this as soon as possible. I have encountered few issues while running webkit tests (https://bugs.webkit.org/show_bug.cgi?id=200870). Once the issue is fixed, I will prepare the test cases and update the bug. ~ Thanks.
Comment on attachment 23493 [details] This patch contains only the changes to enhance the navigation to the accessibility parent. Jon is the man here...
Comment on attachment 23493 [details] This patch contains only the changes to enhance the navigation to the accessibility parent. We prefer early return, so this would be better written as: if (!parentObj) return E_FAIL; ... The change looks good, just needs a test. r- for missing test. Thanks for the patch, Sankar!
Jon, Thanks for your review comments. I am preparing few test cases for this patch. As a part of those test cases, I need to use the functions that let us walk towards the parent of an element in the test case. For this, what are the functions that I can use. I understand that support for accessibilty tests is limited on windows. It would be great, if you can tell me few examples for this. ~Thanks. Sankar. (In reply to comment #4) > (From update of attachment 23493 [details] [review]) > We prefer early return, so this would be better written as: > > if (!parentObj) > return E_FAIL; > ... > > The change looks good, just needs a test. r- for missing test. > > Thanks for the patch, Sankar! >
I browsed through sample scripts available in LayoutTests directory. I don't find any examples that uses parentElement function. Is that function supported in AccesibilityUIElement class. Thanks Sankar.
Jon, Can you provide some inputs for comment #6. Is the support available for parent navigation in Layout tests? ~ Thanks Sankar.
(In reply to comment #7) > Jon, > Can you provide some inputs for comment #6. Is the support available for parent > navigation in Layout tests? No, AccessibilityUIElement::parentElement() is not yet implemented on Windows. I'll try to add that soon!
<rdar://problem/15113186>
AccessibilityUIElement::parentElement() seems to have been implemented some time ago.