This revision changes the logic for creation and updating of accessibles
corresponding to img elements such that alt="" (without click listeners or any
other aria attribute that forces an accessible) will effectively remove that
img's accessible from the accessibility tree. Next, this revision removes the
concept of eNoNameOnPurpose from accessible name handling, since it's now
unnecessary now that alt="" means "don't create an accessible" in most cases.
This revision also adds tests to verify the functionality and updates existing tests.
Differential Revision: https://phabricator.services.mozilla.com/D160706
DocAccessible::GetAccessible() won't return the document if a root element
like body is passed. We need to check for that in NotifyOfPossibleBoundsChange
and use the document accessible in that case.
We can't modify GetAccessible() because that is not that method is used
extensively and that is not the expected behavior.
Differential Revision: https://phabricator.services.mozilla.com/D160677
For some reason the default monospace font in Linux is not actually
monospaced and a linebreak is not guaranteed.
Also disable spellcheck. It causes a reflow after caret move with certain fonts.
Differential Revision: https://phabricator.services.mozilla.com/D160301
Layout can trigger several text attribute changed events. We need to
wait until we get the correct one before continuing the test.
Differential Revision: https://phabricator.services.mozilla.com/D160300
We use profiler categories to separate out the profiler markers for major
components, such as DOM or Layout, to quickly understand what is happening in
each of them: this will also be helpful for accessibility. One might be
concerned that we're adding too many categories but I think it's fine: the
accessibility category shouldn't appear if accessibility services aren't in use.
Furthermore, if accessibility code is running when it's unexpected, having the
accessibility category appear will only highlight that fact.
Differential Revision: https://phabricator.services.mozilla.com/D160846
Adding datetime-local atom to the HTMLMarkupMap to map `<input type=datetime-local>` to `DATE_EDITOR` role and providing test coverage for this new accessible.
Differential Revision: https://phabricator.services.mozilla.com/D160871
Layout can trigger several text attribute changed events. We need to
wait until we get the correct one before continuing the test.
Differential Revision: https://phabricator.services.mozilla.com/D160300
DONTBUILD because this is just whitespace cleanup.
Note: mccr8 mentioned in D160577 that this is MSCOM IDL and not XPIDL, for what
it's worth. In any case, there doesn't seem to be any functional significance
to the tab character that I'm removing here, since it's in a declaration
pattern that's repeated in this file, and only this one instance has stray tab
before the closing-paren.
Depends on D160577
Differential Revision: https://phabricator.services.mozilla.com/D160580
This prevents a11y from getting instantiated shortly after clipboard paste, in
order to prevent hangs with the Windows 11 suggested actions feature.
When combined with the previous patch, the behavior is the following:
* For users with a11y already-enabled:
* No hang (due to clipboard flush).
* Quick actions menu is positioned at selection offset.
* For users with a11y disabled (most):
* No hang (due to no a11y instantiation + clipboard flush).
* Quick actions menu is positioned at pointer (cursor) offset.
Co-Authored-By: Emilio Cobos Álvarez <emilio@crisal.io>
Differential Revision: https://phabricator.services.mozilla.com/D160652
Depends on D160646
For spinbutton, the spec says there should be "no value" for all of these.
Importantly, this prevents the incorrect value of "50" being reported for ARIA spinbuttons with no value.
Differential Revision: https://phabricator.services.mozilla.com/D159924
HTMLComboboxListAccessible::SelectedOption can return null, so we need
to prepare for that. Same with HTMLComboboxListAccessible::List when
the combo box's child is removed.
Differential Revision: https://phabricator.services.mozilla.com/D159961
These don't work on emulated flexbox. We only have a couple of uses.
See D159726 for the diagnostic patch I used to catch these.
Differential Revision: https://phabricator.services.mozilla.com/D159727
Previously, we were inconsistent in what we treated as a table row in various places.
This caused breakage after bug 1726124, since nsAccUtils::TableFor() started returning a table when an ARIA row had an invalid parent, but Filters::GetRow wouldn't find such a row
This was causing broken expectations (and consequently, an extremely long loop) in TableCellAccessible::PrevColHeader.
To fix this, avoid creating an ARIARowAccessible (and thus ARIAGridCellAccessibles) if the row's parent isn't valid.
This way, there aren't conflicting expectations because the TableCellAccessible code can never run.
Furthermore, clients don't get a broken table cell interface on invalid cells, which was a problem even before bug 1726124.
Differential Revision: https://phabricator.services.mozilla.com/D159679
These don't work on emulated flexbox. We only have a couple of uses.
See D159726 for the diagnostic patch I used to catch these.
Differential Revision: https://phabricator.services.mozilla.com/D159727
ATK expects row, col, but RemoteAccessible previously expected col, row.
The ATK code was passing row, col, thus querying the wrong cell.
To make things more confusing, PDocAccessible specified row, col, even though DocAccessibleChild and RemoteAccessible specified col, row.
Fix this by making the order of the arguments consistent with ATK everywhere; i.e. row, col.
Differential Revision: https://phabricator.services.mozilla.com/D159239
