Good build: 52.0.2729.0
Bad build: 52.0.2730.0

Marking the above issue as RB-Stable as this is a recent regression.

Thank you!

Issue still observed on latest canary 52.0.2738.0.

My change is unlikely to cause this but I can help bisect on Mac. I will look at this shortly.

Trying to repro with public Chromium I get login failure with related log messages below. Does this require an internal build?

[5108:1295:0517/140339:WARNING:CONSOLE(2274)] "Unrecognized message from GAIA: showConfirmCancel", source: chrome://chrome-signin/gaia_auth_host.js (2274)
[5108:1295:0517/140346:ERROR:service_registry_impl.cc(95)] autofill::mojom::AutofillDriver not found
[5108:1295:0517/140346:ERROR:service_registry_impl.cc(95)] autofill::mojom::AutofillDriver not found
[5108:1295:0517/140352:WARNING:CONSOLE(2274)] "Unrecognized message from GAIA: clearOldAttempts", source: chrome://chrome-signin/gaia_auth_host.js (2274)
[5108:1295:0517/140354:WARNING:CONSOLE(2274)] "Unrecognized message from GAIA: showConfirmCancel", source: chrome://chrome-signin/gaia_auth_host.js (2274)
[5108:1295:0517/140412:INFO:CONSOLE(0)] "Error in event handler for (unknown): TypeError: Cannot read property 'send' of null
    at Object.onAPICall_ (chrome-extension://mfffpogegjflfpflabcdkioaeobkgjik/background.js:440:22)
    at Object.invokeMessageCallbacks_ (chrome-extension://mfffpogegjflfpflabcdkioaeobkgjik/channel.js:85:42)
    at Object.onMessage_ (chrome-extension://mfffpogegjflfpflabcdkioaeobkgjik/channel.js:110:12)", source: chrome-extension://mfffpogegjflfpflabcdkioaeobkgjik/_generated_background_page.html (0)
[5108:1295:0517/140413:WARNING:CONSOLE(2274)] "Unrecognized message from GAIA: showConfirmCancel", source: chrome://chrome-signin/gaia_auth_host.js (2274)
[5108:1295:0517/140435:WARNING:gaia_auth_fetcher.cc(705)] ClientLogin failed with 
[5108:1295:0517/140435:WARNING:gaia_auth_fetcher.cc(726)] Incomprehensible response from Google Accounts servers.

Just to update:

Able to reproduce the issue on mac 10.11 chrome version 52.0.2740.0 

I built a Chrome-branded binary and can now repro at ToT. Looking further.

Reverted my change:

https://chromium.googlesource.com/chromium/src/+/ec218fb8a6013ab21b83590acf77530cbbce1097

and issue still persists. Will have to bisect with my local build. This is an experimental flag. I'm assuming the intent is to ship this feature? Just to make sure it is worth the time to bisect.

I've bisected part way but I'm out of time and away until Monday. Current bisect log below. If this has to be addressed Thu/Fri suggest checking with ccameron@ or erikchen@ to find someone else who can bisect an official Chrome build else I will pick up on Monday.

wkorman-macpro:src wkorman$ git bisect bad
Bisecting: 238 revisions left to test after this (roughly 8 steps)
[34d31b534808e8d41265300b4dcc2f83375da7aa] Roll WebGL fe36f23..a7670bc

(haven't build or tested the above commit point yet)

wkorman-macpro:src wkorman$ git bisect log
git bisect start
# bad: [f43094f56012798a8bb0144b0e17076abc8c8dd0] [bgmode] Make push notifications register KeepAlives
git bisect bad f43094f56012798a8bb0144b0e17076abc8c8dd0
# good: [7b287abf66c88a5ac187b301c7dc4c23fd27b792] Roll src/third_party/catapult/ 1bfce2f49..8d108d1cc (1 commit).
git bisect good 7b287abf66c88a5ac187b301c7dc4c23fd27b792
# bad: [7d739009588058584119d03613a940c12b6e6ba5] Convert set zoom calls to go through RenderView.
git bisect bad 7d739009588058584119d03613a940c12b6e6ba5
# bad: [6b771d78012f595785897aed80f597a04126470f] Roll src/third_party/skia/ c12160712..554784cd8 (1 commit).
git bisect bad 6b771d78012f595785897aed80f597a04126470f

Issue is not reproducible on Mac 10.11.5 using 53.0.2747.0.
jshanbal@ : Could you please confirm if its working fine.

Bisected to the change below. Passing ownership to jshanbal@ for feedback. If no longer reproducible perhaps tapted@ or others have already fixed, the change in question has a note as:

// TODO(tapted): Find a better fix (this is horrible).

21610ba731345e89dcb62229b3eceac977f6425b is the first bad commit
commit 21610ba731345e89dcb62229b3eceac977f6425b
Author: tapted <tapted@chromium.org>
Date:   Sun May 8 22:40:49 2016 -0700

    Mac: Convince the WindowServer to update the print preview shadow on retina screens
    
    The OSX WindowServer currently forgets to draw the window shadow on
    print preview when it's shown on retina displays. This started flakily
    in 10.10 and now always happens. It's probably something to do with the
    print preview dialog being wholly a layer-hosting NSView with the
    WebContents showing print preview WebUI. This, combined with the private
    APIs we currently use to show the dialog, seems to confuse the shadow
    generation code in the WindowServer.
    
    To fix, it seems we can make a call to -[NSWindow setFrame:] once the
    animation completes. Other ways of invalidating the shadow didn't have
    an effect. The setFrame call can't be a no-op, but it can be immediately
    reversed. This doesn't seem to negatively impact other dialogs.
    
    It does mean the dialog still doesn't have a shadow whilst it is
    animating in, but that's less noticeable.
    
    BUG= 436884 
    TEST=Show Print Preview (Cmd+p) on a retina screen. It should have a
    border/shadow.
    
    Review-Url: https://codereview.chromium.org/1956283002
    Cr-Commit-Position: refs/heads/master@{#392286}

Yah - so that was to fix a different problem with the GAIA signin popup -- Issue 608920.

I can probably tweak the hack to make this go away.

But I think the scrollbars will go away if the CSS that gets served sets the thing that disables scrolling. That's server side, so probably easier to roll out.

Also, this seems pretty minor. Is it worth blocking? or even merging?

Pretty obscure/minor. I don't think this is worth merging a fix for, and we can't simply revert r392286 since it will break the signin popup in a worse way (i.e. Issue 608920)

Moving this nonessential bug to the next milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

This issue has been moved once and is lower than Pri-1. Removing the milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

With response to comment #10, above issue is still reproducible on latest canary version 54.0.2794.0 on Mac OS.
Please refer the attached screenshot.

Deleted: Actual_video.mov 5.2 MB

Actual_video.mov 5.2 MB Download

--Chrome Identity automated triaging--

This bug is Assigned and has gone one month without any activity, so it is being moved to Available to indicate that it is not actively being worked on. If you are working on this bug, please mark yourself as the owner and move back to Assigned. Please see https://goo.gl/78kbny for more details. Please remove the Services>SignIn or UI>Browser>Profiles components if this bug isn't related to Chrome Identity.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

--Chrome Identity automated triaging--

This bug is Assigned and has gone one month without any activity, so it is being moved to Available to indicate that it is not actively being worked on. If you are working on this bug, please mark yourself as the owner and move back to Assigned. Please see https://goo.gl/78kbny for more details. Please remove the Services>SignIn or UI>Browser>Profiles components if this bug isn't related to Chrome Identity.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot