Re: [orca-list] Clickable table rows on the web

[Joanmarie Diggs <jdiggs igalia com>]

Hi Peter.

The tl;dr version is: We're already doing the checks you described and
have been since the start. Table row was the odd one.

Table cell, list item, and the other roles in that list are not
automatically considered focus mode widgets. They are considered as
candidate focus mode widgets. Then an additional check is done, namely
to be sure we shouldn't treat them as text block elements. List item and
table cell are also in the list of candidate text block elements. The
checks for whether or not a table cell and a list item are text block
elements include:

* Is it editable? If it's editable, it cannot be a text block element.
* Does it lack both state focusable and state focused? If it lacks both,
  it cannot be a text block element.

Again, those checks have been there forever.

The problem with your particular case is that table row was not included
as a text block element. It shouldn't be included as a text block
element. So I removed it last night.

I hope this clarifies your understanding. Thanks again for your report!
--joanie

On 4/24/19 1:03 AM, Peter Vágner wrote:
> Hello,
>
> And thank you for a prompt change.
>
> Table cell, list item are also considered focusable widgets.
>
> I know adding more checks affects performance but can these other
> objects become abused like table rows on the web?
>
> Would check for focusable or focused be appropriate when looking for
> focusable widgets in order to try to hack this independently from the
> object role? Or is that proven not very reliable as well?
> I mean checking for focusable or focused state as an addition to the
> object role.
>
> Greetings
>
> Peter
>
>
> Dňa April 23, 2019 10:20:52 PM UTC používateľ Joanmarie Diggs
> <jdiggs igalia com> napísal:
>
>     Hi Peter.
>
>     I just removed table row from the list of focus mode widgets and your
>     rows now show up as clickables. We have logic which also checks if an
>     object is a grid descendant, and that should catch most table rows we
>     want to treat as focus-mode widgets.
>
>     I wouldn't be surprised if my change just broke something. By the same
>     token, I'm doing a major overhaul of Orca's web support at the moment,
>     so might as well include this change now so it's included and gets lots
>     of testing. You are, of course, testing with master and looking for
>     regressions, right? Right?? <grins>
>
>     Thanks!
>     --joanie
>
>
> -- 
> Odoslané z môjho Android zariadenia prostredníctvom K-9 Mail. Prosím,
> ospravedlňte moju stručnosť.