1816346 - [CTW] Deal with unintended semantics due to role="presentation" being overridden by transforms or position: fixed/sticky

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[James Teh [:Jamie]]

In bug 1806356 and bug 1814800, we forced the creation of Accessibles for elements with transforms and position: fixed/sticky. This was necessary for the cache. This is fine for generic elements like divs and spans, since clients already have to ignore those in many cases. However, it's not great for elements with semantics like tables. For example:

data:text/html,<table style="position: sticky;" role="presentation"><tr><th>test

This table will be exposed as a table in the a11y tree, even though it has role="presentation". In contrast, if you remove the styling, role="presentation" applies as expected and the table is exposed as generic Accessibles.

Note that focusability and global ARIA attributes have always completely overridden role="presentation" and preserved their semantics. However, that is specified behaviour and should remain as is.

There are a couple of ways we could fix this problem with Accessibles forced by styling:

1. Expose these forced Accessibles with generic roles. This would require us to differentiate between cases where we absolutely should override role="presentation" and cases where we just need a generic Accessible. That is, we'd need to split up MustBeAccessible or have it return different answers for these two cases.
2. Support ignored children in the core tree. This would require us to cache the ignored child count at least. It also adds another layer of complexity, given that we already have ignored children on Mac.
   • We could try to move most of Mac's ignored children logic into the core tree. However, each platform would need to be able to extend the rules for whether something is ignored or not. That would mean the core tree would have some different data on different platforms, which doesn't seem great. ON the other hand, it avoids duplication this logic.

Comment 1

[James Teh [:Jamie]]

Attachment: Bug 1816346 part 2: If we force an Accessible to be created due to transform/fixed/sticky despite it being presentational, make it a generic Accessible.

Otherwise, semantics are exposed (e.g. for a table), which completely defeats the author's intent that this be treated as presentational.

Comment 2

[James Teh [:Jamie]]

For the record, we discussed the ignored children approach. Even if we did that (and we may still do so later), we would have needed the generic Accessibles patch anyway. Otherwise, even though the table would have been ignored, we would have exposed table rows and cells.

Comment 3

[James Teh [:Jamie]]

Attachment: Bug 1816346 part 1: Fix the transform tree creation test in browser_test_simple_transform.js.

Previously, the test was retrieving an Accessible with a non-existent id and calling testAccessibleTree on it.
That meant we were calling testAccessibleTree with null, which is a no-op.
Now, we get the intended Accessible so that the test actually tests what it was supposed to test.
This is important to test a potential bug in the next patch.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/22b32e106152
part 1: Fix the transform tree creation test in browser_test_simple_transform.js. r=nlapre
https://hg.mozilla.org/integration/autoland/rev/eeb78c17513c
part 2: If we force an Accessible to be created due to transform/fixed/sticky despite it being presentational, make it a generic Accessible. r=nlapre

Comment 5

[Atila Butkovits]

Backed out for causing failures at test_table.html.

Backout link: https://hg.mozilla.org/integration/autoland/rev/376be91f6486a59d752066b306d0bbf3e8e08fb4

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&selectedTaskRun=RlkyKRhkQ4SmM2UOxfnY_A.0&resultStatus=testfailed%2Cbusted%2Cexception%2Cretry%2Cusercancel&revision=eeb78c17513c2fa161d90b61bbf90d190ea5d8b9

Failure log:

• https://treeherder.mozilla.org/logviewer?job_id=406477001&repo=autoland&lineNumber=7605
• https://treeherder.mozilla.org/logviewer?job_id=406477477&repo=autoland&lineNumber=3221

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1806356

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/daddc9682f04
part 1: Fix the transform tree creation test in browser_test_simple_transform.js. r=nlapre
https://hg.mozilla.org/integration/autoland/rev/7ce58b734414
part 2: If we force an Accessible to be created due to transform/fixed/sticky despite it being presentational, make it a generic Accessible. r=nlapre

Comment 8

[Iulian Moraru]

https://hg.mozilla.org/mozilla-central/rev/daddc9682f04
https://hg.mozilla.org/mozilla-central/rev/7ce58b734414

Comment 9

[James Teh [:Jamie]]

Comment on attachment 9318337 [details]
Bug 1816346 part 2: If we force an Accessible to be created due to transform/fixed/sticky despite it being presentational, make it a generic Accessible.

Beta/Release Uplift Approval Request

• User impact if declined: Some elements (e.g. tables) marked as presentational might be incorrectly exposed to screen reader users.
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Covered by automated tests. Only impacts accessibility exposure of elements in specific cases.
• String changes made/needed:
• Is Android affected?: Yes

Comment 10

[Donal Meehan [:dmeehan]]

Comment on attachment 9318337 [details]
Bug 1816346 part 2: If we force an Accessible to be created due to transform/fixed/sticky despite it being presentational, make it a generic Accessible.

Approved for 111.0b5

Comment 11

[Donal Meehan [:dmeehan]]

Comment on attachment 9318643 [details]
Bug 1816346 part 1: Fix the transform tree creation test in browser_test_simple_transform.js.

Approved for 111.0b5

Comment 12

[Donal Meehan [:dmeehan]]

https://hg.mozilla.org/releases/mozilla-beta/rev/5e946bd47732
https://hg.mozilla.org/releases/mozilla-beta/rev/f85e91c7be37