Issue metadata
Sign in to add a comment
|
[scroll anchoring] feedback loop on forums.lenovo.com
Reported by
yfulgaon...@etouch.net,
Oct 27 2016
|
||||||||||||||||||||||
Issue descriptionChrome Version : 56.0.2902.0 (Official Build) 22203b32bdae9236d23cc89e70772721315f2d1f-refs/heads/master@{#427892} 64-bit OS : Mac(10.10.5, 10.11.4, 10.11.5) Test URL : https://forums.lenovo.com/t5/Moto-G4-Moto-G4-Plus-Moto-G4/Android-7-0-Nougat-Update/td-p/3455513/page/6 What steps will reproduce the problem? 1. Launch chrome and navigate to above URL. 2. Click inside the ‘Search’ box seen at the top RHS of page. 3. Now gently scroll down the page using mouse wheel and observe the page. Actual : Unable to scroll down the page using mouse wheel and page flickers. Expected : User should be able to scroll down the page using mouse wheel and page should not flicker. This is a regression issue broken in ‘M-56’, below is the Manual Regression range and will soon update bisect info. Good build : 56.0.2895.0 Bad build : 56.0.2896.0 Note : 1. This is Mac specific issue and the same is working fine on Windows & Linux OS. 2. Issue is not seen in Safari, Firefox browser.
,
Oct 27 2016
This is another sticky-header content-shift feedback loop. We have three div's, all siblings: (A) .lia-quilt-row-static-header [position: fixed] (B) .lia-quilt-row-top-header (C) .lia-quilt-row-inner-header In the initial state, (A) is display: none. In the sticky state, (A) is display: block, and (B) and (C) are display: none. The "position" values never change. None are ancestors of the anchor node. The page flickers at the transition, but only gets "stuck" if you scroll very slowly with mouse wheel on Mac (which has smaller wheel deltas than other platforms). I don't know a good solution to this, and it's mainly cosmetic, so I'm leaning towards WontFix. Ojan/Yash WDYT?
,
Oct 27 2016
This seems like a case where our SANACLAP heuristic doesn't work. That is, 1) an anchor's ancestor has not changed a layout affecting CSS property and 2) The "position" of anything above the anchor has not changed. This is sort of another variation of the WSJ bug, where sticky is implemented using display rather than position. I can't think of a good solution either; I'm okay with WontFix. Though, I do worry that there are other sites in the wild that probably to sticky the same way. Perhaps we may want to think about the following heuristics: - suppress if some ancestor that has position:fixed changes its display property - like we suppress any changes to position, suppress any changes to display. I don't particularly like any of them.
,
Oct 31 2016
,
Nov 4 2016
Lenovo has been notified of the issue. I think there is no further action for Chrome here.
,
Jan 11 2017
|
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by brajkumar@chromium.org
, Oct 27 2016Owner: skobes@chromium.org
Status: Assigned (was: Unconfirmed)