I can understand not wanting to put THAT code in Kendo, not sure thats the best solution, though may be the most functional one as a hack from outside.
Don't you think binding mousewheel to document causes way to much unnecessary load?
Why not bind the scroll event to the popup itself on creation of the dropdownlist or don't forget OTHER types of widgets that share the same misfortune (like multiselect and autocomplete)
Sure the solution you made is hacky, and I wouldn't want to include that in Kendo either. But there are better less hacky ways of fixing this oddity in webkit - which just happens to be a browser used by over 60% of the internet.
Really this doesn't need VOTES - it needs a proper fix. Sure it'd be best if webkit would fix it, but we all know that is not likely. Programming around browser flaws is part of our duty as front-end engineers. To turn this into a hacky solution that "meh you guys can vote on adding the fix" is not the way to handle it I'd think.
Bottom line: At least 3 of your widgets suffer from a bad UX due to an oddity (bug?) in the worlds most used browser engine.
Do you :
a) drop a hacky fix into docs and ignore it?
b) implement a real fix into the widgets themselves that is perfectly fine performance wise
c) ignore it in hopes the browser engine sees fit to change the behavior