706079 - ignore aria-activedescendant referent having no ARIA role

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[alexander :surkov (:asurkov)]

Attachment: patch

That should be expensive to check aria-activedescendant parent for any element having ID. We ignore aria-activedencendant in some cases for example we do that for focusable state calculation. I think we should take this as a rule to ignore aria-activedescendant referents having no ARIA role. It doesn't make sense to manage HTML controls having no ARIA role by aria-activedescendant because they should be managed by DOM focus. And it doesn't make sense to allow aria-activedensendant pointing to elements having no semantics. So I don't see a usecase for that and thus let's save some time by ignoring such elements.

Comment 1

[David Bolter [:davidb] (NeedInfo me for attention)]

I need to better understand the purpose of this change. How did this bug come about? Is "it doesn't make sense to allow aria-activedensendant pointing to elements having no semantics" specifically talking about semantics as in ARIA semantics? html::TD has implicit semantics for example.

Comment 2

[alexander :surkov (:asurkov)]

(In reply to David Bolter [:davidb] from comment #1)
> I need to better understand the purpose of this change. How did this bug
> come about? 

cleaning up aria-activedescendant implementation

> Is "it doesn't make sense to allow aria-activedensendant
> pointing to elements having no semantics" specifically talking about
> semantics as in ARIA semantics? html::TD has implicit semantics for example.

I didn't think about html:td when I said that, so maybe 'focusable semantics'.

Comment 3

[alexander :surkov (:asurkov)]

try server build http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-cc347c2f7bf0

Comment 4

[David Bolter [:davidb] (NeedInfo me for attention)]

Comment on attachment 577572 [details] [diff] [review]
patch

Review of attachment 577572 [details] [diff] [review]:
-----------------------------------------------------------------

r=me. I hope the role requirement doesn't break anyone.

::: accessible/src/base/nsAccessibilityService.cpp
@@ -926,1 @@
>  }

This removal seems hypothetically dangerous, but I think I'm okay with it.

Comment 5

[alexander :surkov (:asurkov)]

Marco, as you asked the try server build is here http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-7476b968a388/

Comment 6

[Marco Zehe (:MarcoZ)]

All's well with the try-build. I think this is safe to land.

Comment 7

[alexander :surkov (:asurkov)]

inbound land https://hg.mozilla.org/integration/mozilla-inbound/rev/0ed3beaf8247

Comment 8

[alexander :surkov (:asurkov)]

landed https://hg.mozilla.org/mozilla-central/rev/0ed3beaf8247