[Webkit-unassigned] [Bug 139581] AX: [ATK] Use ATK_RELATION_MEMBER_OF to implement linkedUIElements.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=139581

Alejandro Piñeiro <apinheiro at igalia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |apinheiro at igalia.com

--- Comment #7 from Alejandro Piñeiro <apinheiro at igalia.com> ---
(In reply to comment #3)
> (In reply to comment #2)
> > It looks to me like ATK_RELATION_MEMBER_OF is applicable for a group of
> > radio buttons because those radio buttons are all a member of a group.
> > However, I question (quite a bit) its applicability for an internal anchor
> > connection.
> 
> Thank you for your feedback. I understand your concerns about internal
> anchor link, I had it as well. but: 
> 1. There is no other relationship ATK more appropriate to do so. 

First, as Joanmarie said, if you think that an appropiate relationship is missing, then the way to go is asking for a new ATK relation. Relationships has meaning, that different ATs use in order to know how to present information to the users. ATK_RELATION_MEMBER has a meaning, and it is not to include anchors.

Second, I really don't think that a relationship is needed here. Relations, at least in ATK, are used to relate two (or more) different objects, because is a convenient way to get one based on the other. But for the case of links/anchors, that is not needed because ATK API for links already contemplate this:

https://developer.gnome.org/atk/unstable/AtkHyperlink.html#atk-hyperlink-get-object

So you can already get the anchors from an Hypertext document without a relationship.

> 2. The linkedUIElements treats both objects together.

I know that one of the objectives of the test is also detecting things that are still not implemented on the ports, but in this case, seems that is more like that the current test infrastructure is forcing a specific accessibility implementation, that is not one of the objectives. As I said, ATK hyperlink/hypertext API already provides to ATs the way to get the individual anchors, without the need of adding a relationship. Adding the relationship sounds like adding something that ATs doesn't need, so it is just complicating the implementation (imho).

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20141217/9a25175d/attachment-0002.html>