video and ui logs:

https://drive.google.com/file/d/1GzPHT7uTWRlDIKvK3CnEeJ9MYkAAkM5O/view

Deleted: ui.log.1 27.1 KB

ui.log.1 27.1 KB Download

Deleted: ui.log.2 1.5 KB

ui.log.2 1.5 KB Download

chrome: 67.0.3376.0
platform: 10504.0.0-18-03-20)

Are you able to get crash-id from the affected machine?

I only tested on test image. here are mini dumps

Deleted: chrome.20180320.160223.4874.dmp 363 KB

chrome.20180320.160223.4874.dmp 363 KB Download

Deleted: chrome.20180320.132837.8268.dmp 432 KB

chrome.20180320.132837.8268.dmp 432 KB Download

This didn't happen with another account. Looks like this depends on app you have installed. Lowering to P1 although we still need to fix it. I hope we get crash reports soon from canary.

Adding more (arc/mus)  folks in case they have an clue.

Actually that account didn't have ARC enabled, so it could be arc related. 
I could narrow the regression range down to 10492/67.0.3369.0(bad) - 10486/67.0.3366.3(good). I'll run bisect today.

I don't think 67.0.3369.0 had mus enabled. If I'm reading things right 67.0.3369.0 is commit 66edfc19fe65cccd4e0bb80b5ba9ce1a0259e72e and https://chromium.googlesource.com/chromium/src/+/66edfc19fe65cccd4e0bb80b5ba9ce1a0259e72e/testing/variations/fieldtrial_testing_config.json doesn't have WindowService-ChromeOS .

Do you have a stack trace?

FYI for other people, dates on those builds:
2018-03-15 11:12	10492.0.0
2018-03-13 11:10	10486.0.0

The log shows a bunch of errors, but I have no idea if those are expected or not.

Here is the regression range:
https://chromium.googlesource.com/chromium/src/+log/67.0.3366.3..67.0.3369.0?n=10000

I couldn't find anything suspicious about mus nor arc (yet). I'll update the bug later today.

And to be clear, mus is NOT on in any canaries yet.

If you can ssh into the device you might be able to force crash uploads (even on test image) with:

metrics_client -C  # force metrics consent
# make crash_sender try to upload when running a test image
touch /run/crash_reporter/crash-test-in-progress
crash_sender -e FORCE_OFFICIAL=1 -e SECONDS_SEND_SPREAD=1
grep crash_sender /var/log/messages

Login works OK for me with 10506.0.0 / 67.0.3376.0 dev-channel eve test

 Issue 822661  has been merged into this issue.

This is culprit:
https://chromium-review.googlesource.com/954355


0x0000571f75e15042	(chrome -ordinal.h:304 )	ChromeAppListModelUpdater::UpdateAppItemFromSyncItem(app_list::AppListSyncableService::SyncItem*, bool, bool)
0x0000571f75e0ff47	(chrome -app_list_syncable_service.cc:567 )	app_list::AppListSyncableService::ModelUpdaterDelegate::OnAppListItemAdded(ChromeAppListItem*)
0x0000571f75e15128	(chrome -chrome_app_list_model_updater.cc:481 )	ChromeAppListModelUpdater::OnFolderCreated(mojo::StructPtr<ash::mojom::AppListItemMetadata>)

This crash reports before and after

https://crash.corp.google.com/browse?q=product.name%3D%27Chrome_ChromeOS%27%20AND%20expanded_custom_data.ChromeCrashProto.ptype%3D%27browser%27%20AND%20product.Version%3D%2767.0.3369.0%27

https://crash.corp.google.com/browse?q=product.name%3D%27Chrome_ChromeOS%27%20AND%20expanded_custom_data.ChromeCrashProto.ptype%3D%27browser%27%20AND%20product.Version%3D%2767.0.3366.3%27


I confirmed on my device:
https://crash.corp.google.com/browse?stbtiq=b916676a393749a4

hejq@, can you look into this?

+xiyuan

The codereview mentions the CL causing login crashes but says they are fixed.

I think it's same issue.

The regression range had a few fixes, but wasn't probably eonugh.

I see. Let me quickly work on that issue first and check whether we have more.

Sorry I didn't mean to change the status. If you think that's same (and probably is), please merge it back.