[Webkit-unassigned] [Bug 158634] AX: Form label text should be exposed as static text if it contains only static text
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Fri Jun 10 14:19:40 PDT 2016
https://bugs.webkit.org/show_bug.cgi?id=158634
--- Comment #4 from Doug Russell <d_russell at apple.com> ---
Comment on attachment 281045
--> https://bugs.webkit.org/attachment.cgi?id=281045
Patch
View in context: https://bugs.webkit.org/attachment.cgi?id=281045&action=review
>> Source/WebCore/accessibility/AccessibilityLabel.cpp:76
>> + staticText = true;
>
> Seems like you don't need to keep track of staticText. As soon as you get a failure can you can just return false
>
> Then you can return true at the end of it right
>
> Also, you could probably cache this value and then update when children are updated
Having bool staticText means I don't have to check if children is empty. I could switch to size == 0 at the top.
Caching seemed tricky given there seem to be a lot of cases where the funnels aren't used and m_children.append() is used directly. And this method needs to be const to be called from stringValue() so calculating lazily and caching is tricky.
>> Source/WebCore/accessibility/AccessibilityLabel.cpp:92
>> + return WebCore::containsOnlyStaticText(m_children);
>
> This WebCore:: prefix is probably unnecessary
You need it to disambiguate from AccessibilityLabel::containsOnlyStaticText() (the compiler gets huffy without it)
>> Source/WebCore/accessibility/AccessibilityLabel.h:41
>> + AccessibilityRole roleValue() const override { return LabelRole; }
>
> Add a final to these
>
> roleValue and stringValue() can probably be private
Will do on final.
I'll see about role and string.
>> Source/WebCore/accessibility/mac/WebAccessibilityObjectWrapperMac.mm:2237
>> + if (role == LabelRole && is<AccessibilityLabel>(*m_object) && downcast<AccessibilityLabel>(*m_object).containsOnlyStaticText())
>
> This check
> role == LabelRole
>
> Is probably not necessary since is<AccessibilityLabel> guarantees role = Label
It's not strictly necessary, but it seemed likely that role == Label was cheaper than calling is<> on every m_object.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160610/3f988ad1/attachment-0001.html>
More information about the webkit-unassigned
mailing list