Re: [orca-list] Now working around the YouTube heading/focus bug

[Joanmarie Diggs <jdiggs igalia com>]

On 2/13/19 2:59 AM, Alex ARNAUD wrote:
> Le 12/02/2019 à 16:18, Joanmarie Diggs a écrit :
> > I just committed a fix to Orca master which should work around Firefox
> > failing to reliably update focus and caret position when navigating by
> > heading, e.g. on YouTube.
>
> Is it a known reported issue on the Firefox side ?

I haven't gotten around to it yet. It would be great if you could
reproduce it via Accerciser and then report it.

Were I trying to do so, I would first enable Firefox's caret navigation
with F7. You don't need this in terms of Orca or reproducing the problem
for users who are blind. But Orca repositions (or at least tries to) the
caret when you navigate. And when it works, it works whether the caret
is showing or not. But if you're a sighted developer, having the caret
on-screen blinking at you and moving around -- or in the case of the bug
failing to do so -- makes it possible to see the bug.

Thus the steps to reproduce are:

1. Be sure you have the caret enabled and visible (click in some text
   to be sure it's there).
2. Launch YouTube.
3. In Accerciser's pane on the left highlight the heading associated
   with one of the headings at level 3.
4. In the iPython console, type:
   t = acc.queryText()
   t.setCaretOffset(0)

Expected results: The visible caret would move to the heading.

Actual results: Sometimes the visible caret moves to the heading;
sometimes it doesn't.

(Alternative expected and actual results would be to then Tab to the
next object, which should be a link or video immediately after the
heading at level 3, but may or may not be.)

The problem, as you will see with the actual results is that "sometimes
it happens, sometimes it doesn't" is the same thing as saying "Hey
developer, figuring this one out is going to suck. You probably want to
work on something else instead." As a developer, that would be my inner
response. So, if we really want to get someone to fix this, we need to
figure out when and why it fails. And *that* is the thing I don't have
time to deal with right now. Again, if you have the time, it would be super.

In that same spirit, because the YouTube contents are different for
everyone, and change all the time, and because we have a bug that
sometimes happens and sometimes doesn't, it might be worth saving the
YouTube page locally -- and then trimming out the content that isn't
relevant to reproducing the bug. Then you could upload the test case
which would work even if YouTube changed its interface, and it would be
easier for those trying to reproduce it to find the headings at level 3
in Accerciser, etc. This is something else I don't have time to deal
with right now.

Thanks!
--joanie