<input type=range> should be identified as a slider and should be controllable by platform assistive technology.
Created attachment 33976 [details] proposed patch
Comment on attachment 33976 [details] proposed patch Seems this should be in some sort of method on HTMLInputElement instead: 8 // Fire change event manually, as RenderSlider::setValueForPosition does. 109 input->dispatchFormControlChangeEvent(); Null check needed? ntRect AccessibilitySliderThumb::elementRect() const 139 { 140 IntRect intRect = static_cast<RenderSlider*>(m_parentSlider->renderer())->thumbRect(); 141 FloatQuad floatQuad = m_parentSlider->renderer()->localToAbsoluteQuad(FloatRect(intRect)); 142 Eventually we should write a helper to do this: 296 thumbRect.setWidth(thumb->style()->width().calcMinValue(contentWidth())); 297 thumbRect.setHeight(thumb->style()->height().calcMinValue(contentHeight())); I feel like we do that often all over the code. :) Same is true with the centering algorithms you use in the next line. :) (Not for this patch, just noting that eventually we'll need more Rect/Point math code.) I find AccessibilityUIElement.h is a confusing name for this DRT-only class. It's sad that everyone replicates the js-testing harness in their tests. :( You know about fast/js/resources/, TEMPLATE.html, and make-js-tests right? In general this looks fine though.
http://trac.webkit.org/changeset/46720
I was slightly surprised to see no response to my comments. ;)
Oh, I took the final line in your comment literally, sorry. For completeness: I added the NULL check, and I did not know about TEMPLATE.html or make-js-tests (but do now, thanks!).
(In reply to comment #5) > Oh, I took the final line in your comment literally, sorry. It was definitely intended literarily! So you took it as intended. :) > For completeness: I added the NULL check, and I did not know about > TEMPLATE.html or make-js-tests (but do now, thanks!). Excellent. Mostly I wanted to hear that you had added the null check when landing. Thanks!