I haven't seen this myself, but in theory overflow/underflow events could bubble up to <arrowscrollbox> from some anonymous element other than <scrollbox>. This could contribute to bug 1339859 among other things. We already ignore events originating from tab elements (as of bug 658467), and we can be more explicit here by using event.originalTarget rather than event.target.

Comment on attachment 8856086 [details] Bug 1354781 - Filter scrollbox overflow and underflow events more accurately. https://reviewboard.mozilla.org/r/128030/#review130574 r=me as the patch looks reasonable, and I assume you have verified locally that this doesn't regress tab overflow. But I would be curious to know which elements were sending these overflow events if it wasn't the tabs (as you say they were already filtered).

(In reply to Florian Quèze [:florian] [:flo] from comment #2) > Comment on attachment 8856086 [details] > Bug 1354781 - Filter scrollbox overflow and underflow events more accurately. > > https://reviewboard.mozilla.org/r/128030/#review130574 > > r=me as the patch looks reasonable, and I assume you have verified locally > that this doesn't regress tab overflow. Yes. > But I would be curious to know which > elements were sending these overflow events if it wasn't the tabs (as you > say they were already filtered). Me too. Like I said it's just a theory at this point; the patch can be seen as a precautionary measure.

Pushed by dgottwald@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/c0e6cf04adb9 Filter scrollbox overflow and underflow events more accurately. r=florian