495833 - Kinetic scrolling causes panning after user has stopped movement

Status: VERIFIED FIXED

(Firefox for Android Graveyard :: Panning/Zooming, defect)

Description

[Mark Finkle (:mfinkle) (use needinfo?)]

* Load a pannable webpage (http://planet.mozilla.org is long enough)
* Begin panning
* After moving your finger a short distance, stop - but do not release your finger
* With the pan stopped raise your finger

Actual result:
The page moves a small distance due to kinetic

Expected result:
The page does not move at all, since I had stopped panning before raising my finger

Comment 1

[Ben Combee (:bcombee)]

I've noticed this too.  I suggest that if there isn't significant movement for .3s at the end of the momentum buffer, we abort any kinetic movement.  This would detect cases where the user stops moving her finger intentionally.  That time could be adjusted during UI tuning.

Comment 2

[Ben Combee (:bcombee)]

Attachment: Add logic to abort kinetic movement when user stops

Comment 3

[Mark Finkle (:mfinkle) (use needinfo?)]

Comment on attachment 384652 [details] [diff] [review]
Add logic to abort kinetic movement when user stops


>+      if ((Math.abs(mb[i].sx - sX) + Math.abs(mb[i].sy - sY)) > 10) {

the "10" is my only concern. It may be too small for devices where the DPI is high.

Did you test this on a device?

Comment 4

[Mark Finkle (:mfinkle) (use needinfo?)]

perhaps use a:
const kMinimumKineticThreshold = 10;

Comment 5

[Ben Combee (:bcombee)]

I like the idea of a constant or perhaps a pref.  I've not yet tested on a device; I really need to update my n810 build this afternoon.  I figured we'd need to tune that, and maybe the .3s threshold too.

Comment 6

[Brad Lassey [:blassey] (use needinfo?)]

we're using 50 as a threshold in the input handler 
http://mxr.mozilla.org/mobile-browser/source/chrome/content/InputHandler.js#261

Comment 7

[Ben Combee (:bcombee)]

Attachment: Update #2 - use 50 pix value also used in other parts of InputHandler, make configurable

Incorporating feedback from mfinkle and blassey

Comment 8

[Ben Combee (:bcombee)]

Attachment: Update #3 - fixed latent bug with sidebars not hiding this patch exposed

Comment 9

[Ben Combee (:bcombee)]

Attachment: Update #4 - fix checkerboard issue mfinkle saw

Comment 10

[Ben Combee (:bcombee)]

DougT has similar ideas in bug 503362

Comment 11

[Ben Combee (:bcombee)]

Applied patch in bug 503362 fixed problem this patch was trying to solve.

Comment 12

[Aakash Desai [:aakashd]]

I'm still seeing a small move after I release my finger (after stopping) on finance.yahoo.com. I'm re-opening to get some traction on this bug.

Comment 13

[Ben Combee (:bcombee)]

Trying this out on my n810... I see the very minor moves after lifing the finger after a stop, but only when zoomed out.  This isn't the same behavior that was originally related to this bug -- that small move isn't from the kinetic module sending additional events.   I think this might be from phantom mouse moves sent by Maemo after the mouse up event.

Comment 14

[Ben Combee (:bcombee)]

Open a new bug for the left over moves if that's a real issue.  I didn't think it was too annoying.

Comment 15

[Ben Combee (:bcombee)]

I mean to say the original problem is still fixed.

Comment 16

[Aakash Desai [:aakashd]]

If that's the case, then bug 511997 was created for the other issue. This is verified FIXED on builds:

Mozilla/5.0 (X11; U; Linux armv6l; en-US; rv:1.9.3a1pre) Gecko/20090821
Fennec/1.0b3pre

and

Mozilla/5.0 (Macintosh; U; Intel Mac OSX 10.5; en-US; rv:1.9.2a2pre)
Gecko/20090808 Fennec/1.0b3pre