414976 - Support RELATION_NODE_CHILD_OF for ARIA tree views

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[Aaron Leventhal]

Each tree item should use this relation to point to the parent tree item.

Comment 1

[Aaron Leventhal]

Attachment: Needs testing. Support various kinds of ARIA tree views with this relation

Comment 2

[Aaron Leventhal]

Test builds:
https://build.mozilla.org/tryserver-builds/2008-01-30_15:[e-mail]-AriaTreeItemChildOf/

Comment 3

[Aaron Leventhal]

Attachment: Works!

Comment 4

[Aaron Leventhal]

Attachment: Testcase: tree directly setting level,setsize and posinset

Comment 5

[Aaron Leventhal]

Testcase #2 is a tree using role="group" to organize the levels:
http://www.mozilla.org/access/dhtml/tree 

Testcase #3 is the bookmarks tree in the side panel

All 3 now support node_child_of

This bug will allow screen readers on Windows to announce the ancestry of a given tree item.

Comment 6

[alexander :surkov (:asurkov)]

Comment on attachment 300574 [details] [diff] [review]
Works!


>+
>+  // Possibly a tree arranged by using role="group" to organize levels
>+  // In this case the parent of the tree item will be a group and the
>+  // previous sibling of that should be the tree item parent.
>+  // Or, if the parent is something other than a tree we will return that.

could you upload testcases for these?

>+  nsCOMPtr<nsIAccessible> prevAccessible;
>+  parentAccessible->GetPreviousSibling(getter_AddRefs(prevAccessible));
>+  if (!prevAccessible)
>+    return;
>+  prevAccessible->GetFinalRole(&role);
>+  if (role != nsIAccessibleRole::ROLE_OUTLINEITEM) {
>+    nsCOMPtr<nsIAccessible> tempAccessible = prevAccessible;
>+    tempAccessible->GetPreviousSibling(getter_AddRefs(prevAccessible));
>+    if (!prevAccessible)
>+      return;
>+    prevAccessible->GetFinalRole(&role);
>+  }

I didn't get why you try to get previous + previous siblings?

> 
>   //currentlly only for ATK. and in the future, we'll sync MSAA and ATK same. 
>   //that's why ATK specific code shows here

please remove this too.

Comment 7

[Aaron Leventhal]

For the group structure testcase, see comment 5 :b

> I didn't get why you try to get previous + previous siblings?
Sometimes there's an extra text accessible in between. Maybe we should look into why, I suppose.

Comment 8

[alexander :surkov (:asurkov)]

(In reply to comment #7)
> > I didn't get why you try to get previous + previous siblings?
> Sometimes there's an extra text accessible in between. Maybe we should look
> into why, I suppose.
> 

Are there more strong rules? Otherwise what if someone will want to put one more accessible?

Comment 9

[Aaron Leventhal]

Attachment: Fix Surkov's nits

Comment 10

[alexander :surkov (:asurkov)]

Comment on attachment 300633 [details] [diff] [review]
Fix Surkov's nits

looks ok

Comment 11

[alexander :surkov (:asurkov)]

checked in

Comment 12

[alexander :surkov (:asurkov)]

in-testsuite?

Comment 13

[alexander :surkov (:asurkov)]

mochitests have been added in bug 502164