[Webkit-unassigned] [Bug 68142] New: AccessibilityRenderObject::previousSibling is not consistent with nextSibling

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=68142

           Summary: AccessibilityRenderObject::previousSibling is not
                    consistent with nextSibling
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: Unspecified
        OS/Version: Unspecified
            Status: UNCONFIRMED
          Severity: Normal
          Priority: P2
         Component: Accessibility
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: dmazzoni at google.com
                CC: cfleizach at apple.com, aboxhall at chromium.org,
                    davidbarr at chromium.org


I replaced the last line of AccessibilityRenderObject::nextSibling with a check that the previous sibling of the new "next" sibling is equal to this, as follows:

    AccessibilityObject* axNextSibling = axObjectCache()->getOrCreate(nextSibling);
    ASSERT(axNextSibling->previousSibling() == this);
    return axNextSibling;

This resulted in assertion failures in 7 different tests:

Regressions: Unexpected DumpRenderTree crashes : (7)
  accessibility/adjacent-continuations-cause-assertion-failure.html = CRASH
  accessibility/aria-hidden-updates-alldescendants.html = CRASH
  accessibility/div-within-anchors-causes-crash.html = CRASH
  accessibility/duplicate-child-nodes.html = CRASH
  accessibility/image-link-inline-cont.html = CRASH
  accessibility/inline-continuations.html = CRASH
  accessibility/removed-anonymous-block-child-causes-crash.html = CRASH

If previousSibling and nextSibling agreed, this assertion should not fail.

The severity of this is quite low, since most ports do not appear to use previousSibling at all - but it's surely worth fixing.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.