This is a recent regression for the testSessionStore/testUndoTab.js test module which started May 6th. A Nightly from May 5th works like expected. Pushlog: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=4c93d46ab92f&tochange=60349cbc3d4e Failures: "fail": { "message": "'Undo Close Tab' is disabled - got 'false'", "fileName": "file:///mozilla/code/mozmill-tests/nightly/firefox/tests/functional/testSessionStore/testUndoTab.js", "name": "testUndoTabFromContextMenu", "lineNumber": 49 } "message": "'Recently Closed Tabs' sub menu has to be empty - '2' should equal '0'", "lineNumber": 54, "name": "Error", "fileName": "file:///mozilla/code/mozmill-tests/nightly/firefox/tests/functional/testSessionStore/testUndoTab.js" "message": "'Recently Closed Tabs' sub menu entries have changed - '3' should equal '2'", "lineNumber": 855, "name": "Error", "fileName": "file:///mozilla/code/mozmill-tests/nightly/firefox/lib/tabs.js"

For now lets skip this test.

Comment on attachment 8603193 [details] [diff] [review] skip patch https://hg.mozilla.org/qa/mozmill-tests/rev/7e895028f703 (default)

Tim, do you have any idea if your patch on bug 1160556 could have caused this different behavior? Otherwise I cannot find anything else related to sessionstore in the pushlog.

That's more likely caused by bug 1109875 that made state collection for closing tabs asynchronous. This shouldn't really matter in normal usage but would probably break a few tests that close a tab and in the same tick expect all state to be saved and "undo-closeable".

Indeed, that sounds like it is the problem here. Thanks Tim! Given the amount of work I would have to put in here, I will keep this test skipped. Once we convert the sessionstore tests to Firefox UI Tests, we will pick up this information and handle the event properly.

Mozmill tests have been superseded by Marionette tests.