Unagi, 4/4/13 build. screenshot: http://cl.ly/image/022b3w0W0907 Notice that the "what homescreen am i on" indicator is *under* the notification. i could switch homescreens and it would slide around under the notification. the notification tray was not accessible at all. i had to reboot the phone to make the notification go away. STR: * set a daily recurring alarm on the clock for at least 8 hours into the future * do not answer the alarm when it goes off in the morning, because you left the phone downstairs on your desk * wake up because you hear the phone crash to the floor as it vibrates itself off the desk (probably trying to escape) The last step is optional.

Nominating for tef+ since a daily-use activity can break the homescreen UI, requiring restart.

Dietrich, The alarm mini-notification is an mini-style of the attention screen. If you click "home" button when the attention screen is showing, it will change to the style. Then, you could click the mini-attention screen to the full page style. The attention screen is needed to close by user.(Click 'Close' or 'Snooze' button) Do the above scenario work for you?

No, I tried everything - home button, touching the notification, etc. Nothing at all worked, had to reboot the phone.

Dietrich... how did you get a future build? I would like your time machine. :) On a serious note, which build did you mean? And are you able to reproduce this consistently?

(clearing tef? till we have more details on str, etc)

Not able to reproduce consistently - only saw once.

As I tested this case with latest build(b2g18-V1.0.1), I cannot reproduce it. My test result is as Ian mentioned. We could monitor this case. If the bug happens again, we could collect the log and report/raise it. Thanks. Test build: (mozilla-b2g18_v1_0_1-unagi-eng/2013-03-31-23-02-03) * Mercurial-Information: Gecko revision="b28463f2e718" Gaia revision="8c643de8c0fb" * Git-Information Gecko revision="a9e27af282b8d499ee49ae691918cea88b2f8571" Gaia revision="ddb38ac8a34f9e30e09d0ff3b5c1bfb9b664b7c3"

Thanks William. let's close this for now if we don't have consistent STR. When we reproduce this again, let's file a new bug.