1197089 - Incorrect IAccessibleText result for some text

Status: RESOLVED INCOMPLETE

(Core :: Disability Access APIs, defect)

Description

[antony]

Attachment: snipptet.html

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/44.0.2403.155 Safari/537.36

Steps to reproduce:

Using IAccessibleText's offsetAtPoint, textAtOffset can produce wrong results. For example the text under the title of items at http://www.bbc.co.uk/news on Windows Firefox. To easily see what's happening just run NVDA and move the cursor over this text screen.


Actual results:

Returns text corresponding to an aria-hidden="true" text that applies to block link - see attached example snippet.


Expected results:

Be able to get word at cursor position. 

For blind access this behaviour might be OK (but shouldn't aria-hidden elements always be hidden?), but not at all good for other utilities trying to get the word at a point on screen (e.g. dictionaries).

Comment 1

[Marco Zehe (:MarcoZ)]

Welcome to the bloody hell that is aria-hidden. This attribute has so many unanswered questions and was jammed into the specification at such rush that nobody has answered these, and many other, questions. Because all assistive technologies rely on the same accessibility APIs, be it screen readers and magnifiers as well as switch controls and more, the inconsistencies aria-hidden introduces to solve the problems for *one* assistive technologies have adverse affects on others, as you have experienced yourself with this example. The a11y API doesn't know who is asking for the text, nor should the API *have* to know that.

Sorry, getting into a rant here when I only wanted to correctly triage the bug into the right component.

Comment 2

[Marco Zehe (:MarcoZ)]

We're now hiding the tree properly in Firefox Nightly (63). Could you re-test this please? I am a bit unclear as to which steps you are actually taking in comment #0.

Comment 3

[Firefox Bug Husbandry Bot]

2019-03-06

This bug is part of a group of bugs which have had an open needinfo for at least 12 weeks.

The request for information has not been answered, and we can't move forward on the bug so we are closing it.

If the defect is still present, please reopen this bug with an updated report.