The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/90d7824a5136d081090646b7ec7f7158aaae8aca

commit 90d7824a5136d081090646b7ec7f7158aaae8aca
Author: jamiewalch <jamiewalch@chromium.org>
Date: Thu Jan 12 00:15:54 2017

Create proper object types when loading host cache.

I suspect that host.options was originally a simple dictionary, and so the cast was safe. When it got upgraded to an object with methods on the prototype it broke, but because the window of opportunity for failure is small (you have to connect without a call to HostList.refresh happening in the interim), it was difficult to track down.

BUG= 680302 

Review-Url: https://codereview.chromium.org/2621403004
Cr-Commit-Position: refs/heads/master@{#443072}

[modify] https://crrev.com/90d7824a5136d081090646b7ec7f7158aaae8aca/remoting/webapp/base/js/host.js
[modify] https://crrev.com/90d7824a5136d081090646b7ec7f7158aaae8aca/remoting/webapp/base/js/host_options.js
[modify] https://crrev.com/90d7824a5136d081090646b7ec7f7158aaae8aca/remoting/webapp/crd/js/host_list.js

Requesting merge of 90d7824a5136d081090646b7ec7f7158aaae8aca to M56. This change only affects Chrome Remote Desktop; there is no change to Chrome browser.

Approved for branch 2924 I guess.

How many of these merges do you need usually?  Does getting approval really help anything if the code has no impact on Chrome the browser?  If not we should just let you handle your own repos...

It looks like this is the 5th Chromoting-related merge we've requested for M56; is that too many? I'm not sure what you mean about getting approval--I was under the impression it was mandatory for all merges.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/881bcc9c0073e39ccd53444b19fb72a4def29a52

commit 881bcc9c0073e39ccd53444b19fb72a4def29a52
Author: Jamie Walch <jamiewalch@chromium.org>
Date: Fri Jan 13 02:04:20 2017

Create proper object types when loading host cache.

I suspect that host.options was originally a simple dictionary, and so the cast was safe. When it got upgraded to an object with methods on the prototype it broke, but because the window of opportunity for failure is small (you have to connect without a call to HostList.refresh happening in the interim), it was difficult to track down.

BUG= 680302 

Review-Url: https://codereview.chromium.org/2621403004
Cr-Commit-Position: refs/heads/master@{#443072}
(cherry picked from commit 90d7824a5136d081090646b7ec7f7158aaae8aca)

Review-Url: https://codereview.chromium.org/2627073005 .
Cr-Commit-Position: refs/branch-heads/2924@{#754}
Cr-Branched-From: 3a87aecc31cd1ffe751dd72c04e5a96a1fc8108a-refs/heads/master@{#433059}

[modify] https://crrev.com/881bcc9c0073e39ccd53444b19fb72a4def29a52/remoting/webapp/base/js/host.js
[modify] https://crrev.com/881bcc9c0073e39ccd53444b19fb72a4def29a52/remoting/webapp/base/js/host_options.js
[modify] https://crrev.com/881bcc9c0073e39ccd53444b19fb72a4def29a52/remoting/webapp/crd/js/host_list.js

Not seeing any automates tests in the CL, wondering whether we can automate to avoid regressions in future. 

We are working on a framework to allow end-to-end automation for Chrome Remote Desktop, but it's outside of Chromium and would likely not have helped with this bug anyway, which was very timing-sensitive.