fix render overflow computation for input type=range
Created attachment 94089 [details] Patch
Created attachment 94092 [details] Patch
(In reply to comment #2) > Created an attachment (id=94092) [details] > Patch I moved the call to m_overflow.clear() to be outside the thumb check. It's possible for the thumb to be remove and the overflow not cleared.
Comment on attachment 94092 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=94092&action=review Test is (theoretically at least) flaky > LayoutTests/fast/forms/slider-hit-testing.html:1 > +<head> <!DOCTYPE html> plz unless this test needs quirks mode > LayoutTests/fast/forms/slider-hit-testing.html:11 > +window.setTimeout(function() { > + document.getElementById("slider").style.width = "100%"; > + changeAndCheckRangeValue(); > +}, 0); the setTimeout is a bit odd - does the test not work if you run this code in an onload handler? technically speaking, this setTimeout _could_ execute immediately after parsing the end of the <script> and before the <input> is parsed, although in practice we'll nearly always parse through the end of this document before running timeouts. > LayoutTests/fast/forms/slider-hit-testing.html:20 > + var x = slider.offsetLeft + (slider.clientWidth * .75); > + var y = slider.offsetTop + (slider.clientHeight / 2); could you add some comments indicating what you are trying to click on here and what should be happening?
Created attachment 94112 [details] Patch
Comment on attachment 94092 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=94092&action=review >> LayoutTests/fast/forms/slider-hit-testing.html:1 >> +<head> > > <!DOCTYPE html> plz unless this test needs quirks mode Done. >> LayoutTests/fast/forms/slider-hit-testing.html:11 >> +}, 0); > > the setTimeout is a bit odd - does the test not work if you run this code in an onload handler? technically speaking, this setTimeout _could_ execute immediately after parsing the end of the <script> and before the <input> is parsed, although in practice we'll nearly always parse through the end of this document before running timeouts. I didn't work in the onload handler, but I added a call to document.body.offsetLeft to force a layout. This allowed me to remove the waitUntilDone/notifyDone calls. >> LayoutTests/fast/forms/slider-hit-testing.html:20 >> + var y = slider.offsetTop + (slider.clientHeight / 2); > > could you add some comments indicating what you are trying to click on here and what should be happening? Done.
Comment on attachment 94112 [details] Patch Looks good!
Committed r87168: <http://trac.webkit.org/changeset/87168>
Revision r87168 cherry-picked into qtwebkit-2.2 with commit 49b8a1e <http://gitorious.org/webkit/qtwebkit/commit/49b8a1e>