Mac: M69 introduced change in how drag lock impacts Chrome windows |
||||||
Issue descriptionReport is from twitter. https://twitter.com/rhhackett/status/1053329591618678785 and subsequent DMs What steps will reproduce the problem? (1) In macOS trackpad settings, turn on "enable dragging" and "three finger drag" (2) Click on the top rail of a browser window—either the dark grey space above the tabs, or in the grey space to the right of them—that is in the background of another open browser window (3) Click elsewhere What is the expected result? nothing moves What happens instead? The Chrome window from (2) jumps to the spot clicked on in (3) The twitter user linked above had turned off "enable dragging" and will let me know if it still repros. Repro has never been consistent, and only started after M69. rpop comment: we should be consistent with other apps on the platform when drag lock is enabled.
,
Oct 19
Update: reporter says the browser is still jumping, even with "enable dragging" disabled.
,
Oct 20
What MacBook model? What OS? What Chrome version?
,
Oct 21
,
Oct 22
Chrome version 70.0.3538.67 for hardware and OS version, which I suspect are relevant, please see this doc: https://docs.google.com/document/d/1m6TDddzs6pRWKKmR1Gy-es-fBV6vGpomL2nOtpJ1yF8/edit
,
Oct 23
Tried testing issue on reported chrome version 70.0.3538.67(as mentioned in comment#5) on Mac OS 10.13.6 by following below steps. Steps: ===== 1.Launched chrome. 2.Changed the trackpad settings by turn on "enable dragging" to "three finger drag". 3.Clicked on the dark grey space above the tabs and in the grey space to the right of the window which is the background of another open browser window. 4.Clicked some another place in the window. 5.Observed that currently opened window is that which we have clicked in the background(action performed in step:2). Attached screencast for reference. @reporter: Could you please review attached screencast and let us know if anything is being missed here. If possible request you to provide a screenshot/screencast of the issue so that it would be really helpful for further triaging of the issue. Thanks.!
,
Oct 25
Related to bug 898953 ?
,
Oct 29
I think it probably is. Can someone try confirming on 10.10?
,
Oct 29
Issue 899578 has been merged into this issue.
,
Oct 29
lgrey@: The reporter of issue 899578 is also on 10.10. Maybe the repro steps there helps here too.
,
Oct 29
Good idea! For convenience, here they are: .. Steps to reproduce the problem: 1. Open a new Chrome window. Browse to any content (or just leave it on the start page). 2. Change focus to any other window. (Can be another Chrome window, or any other app.) 3. Change focus back to the Chrome window by clicking (MUST BE DONE AS A TOUCHPAD TAP) on the gray bar along the top of that window (anywhere to the right of the "+" which will open a new tab). 4. Then click anywhere within the HTML content area of the Chrome window. This can either be a touchpad tap or full-on click. 5. The Chrome window moves to a new location on the screen. What is the expected behavior? Window should remain in the same place. What went wrong? Window moved. ... (I don't have 10.10 on a trackpad machine immediately handy, so can't confirm)
,
Oct 29
Reporter of issue 899578 (and Xoogler) here. It seems like what's happening is that the touchpad tap (on the window drag bar, in Chrome M69+) is bringing the Chrome window to the foreground but not giving it focus. So, for example, I have noticed that keypresses do not go to the Chrome window when I tap it once... I have to either give it a full click, or two taps. This bug is present in both 69.0.3497.100 (stable channel) and 72.0.3594.0 (canary). I will be happy to test builds for you if you want. Just let me know.
,
Oct 29
Issue 898953 is also 10.10. Wonder if something is going in general with input there.
,
Nov 1
I heard offline that this repros on 10.10 with "tap to click" enabled.
,
Nov 2
I ended up figuring out what was going on here in the course of looking into another issue. This was introduced by https://chromium-review.googlesource.com/c/chromium/src/+/1000212 When "Tap to Click" is enabled, a tap comes in as a mouseDown: (which AFAICT is at this point indistinguishable from an actual trackpad click). Then something or other happens that prevents the mouseUp, which arrives single-digit ms layer, from killing the drag. If we can't find a way to detect a tap vs. a click, maybe we should consider waiting a little bit before initiating the drag.
,
Nov 21
,
Dec 19
,
Dec 26
|
||||||
►
Sign in to add a comment |
||||||
Comment 1 by a...@chromium.org
, Oct 19