Summary: | WAI-ARIA roles not supported on image map <area> | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | chris fleizach <cfleizach> | ||||||
Component: | Accessibility | Assignee: | chris fleizach <cfleizach> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | bdakin, commit-queue | ||||||
Priority: | P2 | ||||||||
Version: | 528+ (Nightly build) | ||||||||
Hardware: | PC | ||||||||
OS: | OS X 10.5 | ||||||||
Attachments: |
|
Description
chris fleizach
2009-11-15 18:26:42 PST
Created attachment 43409 [details]
patch
Comment on attachment 43409 [details] patch > + String ariaRole = m_areaElement->getAttribute(roleAttr); It's usually a slightly better idiom to put fetched attributes into const AtomicString& local variables instead of String, because it avoids a little bit of reference count churn. > +static const ARIARoleMap& createARIARoleMap() I think it's cleaner for a create function to return an ARIARoleMap* rather than a const ARIARoleMap&. I realize you're just moving this code. > +AccessibilityRole AccessibilityObject::ariaRoleToWebCoreRole(String& value) The argument should be const String&, not String&. > + ASSERT(!value.isEmpty() && !value.isNull()); Normally we don't use && in assertions. Instead we do multiple assertions in such cases; that way we know which one failed. Also, all null strings are also empty, so the only reason to have both assertions is if you want to know from the assertion which is which, and if so, the null assertion has to come first. > - AccessibilityRole role = ariaRoleToWebCoreRole(ariaRole); > + AccessibilityRole role = AccessibilityObject::ariaRoleToWebCoreRole(ariaRole); Do you really need to change this? Since AccessibilityRenderObject is derived from AccessibilityObject I'd expect this to already be in scope. review- because of the String& thing -- everything else is too minor to matter, but that is not good Created attachment 43483 [details]
patch
made all of darin's suggestions
adding to commit queue, cuz the code is elsewhere Comment on attachment 43483 [details] patch Clearing flags on attachment: 43483 Committed r51195: <http://trac.webkit.org/changeset/51195> All reviewed patches have been landed. Closing bug. |