To show a button for a list attribute of an input element, introduce a pseudo CSS selector, new appearance value, and a ControlPart value.
Created attachment 33697 [details] Proposed patch * See https://bugs.webkit.org/attachment.cgi?id=33600 for the actual appearance. * I'll add tests with another patch.
dhyatt was working on <datalist> support, he should be CCed on all bugs about it
Comment on attachment 33697 [details] Proposed patch No { }: if (part == ListButtonPart) { 337 [buttonCell setBezelStyle:NSRoundedDisclosureBezelStyle]; 338 } el Tabs: 47 case ListButtonPart: 448 { 449 LengthSize result = sizeFromFont(font, LengthSize(zoomedSize.width(), Length()), zoomFactor, buttonSizes()); 450 result.setWidth(result.height()); 451 return result; 452 } Where are the tests? Parts of this at least should be guarded by some sort of DATALIST define. Otherwise this looks OK.
Created attachment 38715 [details] Proposed patch (rev.2) - update for the latest WebKit source - add a test Please refer to bug#27247 for the proposed UI.
Add more Forms people to CC, and Eric who reviewed the previous patch.
Comment on attachment 38715 [details] Proposed patch (rev.2) Should we be getting these from AppKit somehow? 317 static const IntSize sizes[3] = { IntSize(21, 21), IntSize(19, 18), IntSize(17, 16) };
(In reply to comment #6) > Should we be getting these from AppKit somehow? > 317 static const IntSize sizes[3] = { IntSize(21, 21), IntSize(19, 18), > IntSize(17, 16) }; We use similar fixed-size arrays for other controls; radioSizes(), checkboxSizes(), buttonSizes(). They were introduced by hyatt 4 years ago. http://trac.webkit.org/changeset/10091 I don't know the reason why fixed values were used in them. performance? simpler code?
Comment on attachment 38715 [details] Proposed patch (rev.2) This looks right to me (and no one has complained in the 3 weeks this has been up for review, so it can't be *that* awful). The only thing it's missing is the mac pixel results. I'll approve this now, but it would be better if you could post a patch with the mac pixel results. Otherwise someone will need to make a second commit to add this missing mac pixel results.
Comment on attachment 38715 [details] Proposed patch (rev.2) Adding cq+ since it looks like this is still waiting for commit.
Comment on attachment 38715 [details] Proposed patch (rev.2) Clearing flags on attachment: 38715 Committed r49103: <http://trac.webkit.org/changeset/49103>
All reviewed patches have been landed. Closing bug.