STR: * go to http://fr.m.wikipedia.org/wiki/Montbrison_%28Loire%29 * scroll down until "Localisation" * try to scroll again (up or down) when the first touch is on the map Expected: * it's easy Actual: * the scroll up does not work if we start by scrolling left/right. The problem is that this can be done without trying. In this case, the scrollable element can't be scrolled up or down, so I think that when the user try to scroll up/down the whole page should scroll. If we try to make it more general, the page should scroll <direction> as soon as the element touched can't be scrolled <direction>, with <direction> is either up, down, left, right. I haven't checked how it works on Android or even on Firefox Desktop with the mousewheel.
I can reproduce this. I'm not sure if it's intentional but once you've panned left/right the scrollable element seems to swallow all the panning input so you can't pan up/down. It's fairly easy to work around for the user, but it is irritating. Not sure where in the platform this bug should live.
BrowserElementPanning will lock the scrolling target if any panning takes effect. I think implementing axis lock can improve the user experience, however, we planned not to do so (see bug 824525). Maybe we need to re-evaluate about axis lock on B2G.
BTW, I tried this STR on iOS safari got the same behavior as FirefoxOS.
I come back to my proposed solution : When the user tries to scroll <direction> an element, we should scroll <direction> that element first; then if that element can't be scrolled <direction> (because it's already at the end of the scroll), then try again with the nearest scrollable parent element.
Looks like a Layers bug after all.
I can't reproduce this issue with latest gecko. I think the issue was fixed by my patch for bug 898478. Julien, could you try again with latest gecko and see if you still experience this?
The main issue is fixed but the kinetic panning does not work when the touchstart is inside the map on the URL in comment 0. :)
Julien, let's open a separate bug for the kinetic panning scenario you mention so that we don't have multiple issues in this bug.
ok, I'll dupe this one then.