Closed Bug 701903 Opened 13 years ago Closed 13 years ago

Failure in testSearchSelection.js::testSearchSelectionViaContextMenu

Categories

(Mozilla QA Graveyard :: Mozmill Tests, defect)

defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: u279076, Assigned: remus.pop)

References

()

Details

(Whiteboard: [mozmill-test-failure])

Attachments

(6 files, 4 obsolete files)

(deleted), patch
u279076
: review+
Details | Diff | Splinter Review
(deleted), patch
u279076
: review+
Details | Diff | Splinter Review
(deleted), patch
u279076
: review+
Details | Diff | Splinter Review
(deleted), patch
u279076
: review+
Details | Diff | Splinter Review
(deleted), patch
whimboo
: review+
Details | Diff | Splinter Review
(deleted), patch
whimboo
: review+
Details | Diff | Splinter Review
MODULE: tests/functional/testSearch/testSearchSelection.js TEST: testGetMoreEngines FAIL: TimeoutError("controller.waitForEval: Timeout exceeded for 'subject.selectedTabIndex == subject.expectedIndex'")@mozmill-tests/tests/functional/testSearch/testSearchSelection.js:83 BRANCH: mozilla-aurora, default
Assignee: nobody → vlad.mozbugs
Status: NEW → ASSIGNED
A Pivotal Tracker story has been created for this Bug: https://www.pivotaltracker.com/story/show/20968059
Attached patch patch fix v1.0 - initial proposed fix (obsolete) (deleted) — Splinter Review
Changed waitForEval to waitFor
Attachment #574287 - Flags: review?(anthony.s.hughes)
Comment on attachment 574287 [details] [diff] [review] patch fix v1.0 - initial proposed fix >+const LOCAL_TEST_FOLDER = collector.addHttpResource('../../../data/'); >+const LOCAL_TEST_PAGE = LOCAL_TEST_FOLDER + 'layout/mozilla_mission.html'; > > const TIMEOUT = 5000; > >-const LOCAL_TEST_FOLDER = collector.addHttpResource('../../../data/'); >-const LOCAL_TEST_PAGE = LOCAL_TEST_FOLDER + 'layout/mozilla_mission.html'; >- >-var setupModule = function() { >+function setupModule() { These changes are unnecessary. Please only make changes which directly fix the failure. Making these changes now breaks test consistency and will impede broad-sweep refactoring efforts. The same goes for any following code which you "refactored" for style. >+ controller.waitFor(function () { >+ return (tabs.selectedIndex === tabIndex); >+ }, "Selected tab: '" + tabIndex + "'"); Can you refactor the message to read: "Selected tab X, expected tab Y" >+ controller.waitFor(function () { >+ return (tabs.selectedIndex === tabIndex + 1); >+ }, "Selected tab: '" + (tabIndex + 1) + "'"); Same here.
Attachment #574287 - Flags: review?(anthony.s.hughes) → review-
This test is currently failing on latest Aurora and latest Nightly because those builds do not open a tab in background anymore. I am uploading a mozmill work patch to demonstrate this. If you want to test manually, do it with following STR: 1. Open Firefox Aurora or Nightly 2. Navigate to google.com and search "Firefox" 3. Go to suggestions and mouse select the word "Firefox" from anywhere on the page 4. Right click and use the context menu to search (Search "ebay" for "word") A new tab is opened in foreground, even if the pref 'browser.tabs.loadInBackground' is set to true. The question is - Do we want this? If this is an expected behavior we'll change the test, if not, we have ourselves a Firefox bug. What do you think ?
Please apply the patch locally to test on, and watch the console for dumping messages. Compare the results with a firefox release or beta build and a firefox aurora or nightly build.
Vlad, the behaviour you got now is the expected one as of bug 695482. Also see bug 696747 for reference to get the wanted behaviour when that lands.
(In reply to Remus Pop (:RemusPop) from comment #6) > Vlad, the behaviour you got now is the expected one as of bug 695482. > > Also see bug 696747 for reference to get the wanted behaviour when that > lands. Even if bug 695482 is landed reading the comments shows that people have second opinions on this still. Bug 696747 has an r- patch so still needs work. We can definitely fix this failing test but my question is now: - Do we provide a separate fix for aurora and nightly and just fix beta and release at merges time? (can be tricky) - How do we proceed? My personal idea is to disable the test on aurora and nightly and keep them for beta and release which are more important to have this test, and then after 12 weeks make the change on all branches.
Update the test so it aligns with the new behavior and watch the other bug so we can update the test later. Same applies to the Litmus test which needs an update.
release and beta branch do not need a fix yet.
Attachment #574287 - Attachment is obsolete: true
Attachment #574582 - Attachment is obsolete: true
Attachment #575450 - Flags: review?(anthony.s.hughes)
(In reply to Henrik Skupin (:whimboo) from comment #8) > Update the test so it aligns with the new behavior and watch the other bug > so we can update the test later. Same applies to the Litmus test which needs > an update. Updated litmus test for aurora branch https://litmus.mozilla.org/show_test.cgi?id=15707 We don't have the default (nightly) branch there so no update in litmus for that. Hope I didn't do anything stupid.
Attachment #575450 - Flags: review?(anthony.s.hughes) → review+
Attachment #575450 - Attachment description: patch v1.0 - fix for aurora and nightly branch only → patch v1.0 - aurora/nightly [checked-in]
Please mark VERIFIED and update the spreadsheet after checking tomorrow's results.
Status: ASSIGNED → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
As given on the dependent bug the new feature has been backed out. That means that we should also backout this patch from default and aurora.
Status: VERIFIED → REOPENED
Resolution: FIXED → ---
Comment on attachment 575450 [details] [diff] [review] patch v1:default/aurora [backed-out] Backed out: http://hg.mozilla.org/qa/mozmill-tests/rev/2e459e48fd4d (default) http://hg.mozilla.org/qa/mozmill-tests/rev/7176167b1648 (mozilla-aurora) We will need to revisit this once bug 695482 is resolved.
Attachment #575450 - Attachment description: patch v1.0 - aurora/nightly [checked-in] → patch v1:default/aurora [backed-out]
Vlad, please submit a patch immediately which marks this test disabled.
Status: REOPENED → ASSIGNED
Block should be the other way around...this bug depends on the other bug.
No longer blocks: 695482
Depends on: 695482
skip patch as requested - we'll need to address a fix once the dependency is fixed
Attachment #580027 - Flags: review?(anthony.s.hughes)
Attachment #580027 - Flags: review?(anthony.s.hughes) → review+
Attachment #580027 - Attachment description: skip patch 1.0 → disable test (default/aurora) [checked-in]
Summary: Failure in testSearchSelection.js: Timeout exceed for 'subject.selectedTabIndex == subject.expectedIndex' → Failure in testSearchSelection.js::testSearchSelectionViaContextMenu
Whiteboard: [mozmill-test-failure] → [mozmill-test-failure][mozmill-test-skipped]
The dependency is shown as fixed, will investigate if the fix allows me to provide a fix patch for this test.
(In reply to Maniac Vlad Florin (:vladmaniac) from comment #20) > The dependency is shown as fixed, will investigate if the fix allows me to > provide a fix patch for this test. The dependency patch fixes our failure for nightly, still the other branches are still affected and the test should remain disabled for aurora, beta and firefox release branch. I suggest reopening this test once bug 592275 is fixed. Adding dependency
Depends on: 592275
Lets enable the test again on default.
(In reply to Henrik Skupin (:whimboo) from comment #22) > Lets enable the test again on default. Feel free to go on and enable it on default branch
(In reply to Maniac Vlad Florin (:vladmaniac) from comment #23) > (In reply to Henrik Skupin (:whimboo) from comment #22) > > Lets enable the test again on default. > > Feel free to go on and enable it on default branch Forget I said that - needs a little more investigation. I'll follow up shortly
It's time to get this test back in. This patch fixes our failures for the default branch
Attachment #605684 - Flags: review?(anthony.s.hughes)
Enable the test as it is on mozilla-aurora branch The new pref which was causing our failure is not landed for mozilla-aurora.
Attachment #605688 - Flags: review?(anthony.s.hughes)
Comment on attachment 605688 [details] [diff] [review] [mozilla-aurora][mozilla-beta][mozilla-release] enable test v1.0 The patch is eligible and tested also on mozilla-beta and mozilla-release branch. We can enable also on those branch. Dependencies were fixed so we can re-use this test
Attachment #605688 - Attachment description: [mozilla-aurora] enable test v1.0 → [mozilla-aurora][mozilla-beta][mozilla-release] enable test v1.0
Fixed commit message to inform about the changes in all branches, not just in aurora
Attachment #605688 - Attachment is obsolete: true
Attachment #605688 - Flags: review?(anthony.s.hughes)
Attachment #605691 - Flags: review?(anthony.s.hughes)
The fix patch passes locally on the functional testrun for nightly http://mozmill-crowd.blargon7.com/#/functional/report/e438d6e3916b2b636037d77445c613bd
Attachment #605684 - Flags: review?(anthony.s.hughes) → review+
Comment on attachment 605684 [details] [diff] [review] [default] fix v1.0 [landed:default] Landed: http://hg.mozilla.org/qa/mozmill-tests/rev/9f645bb2a217 (default)
Attachment #605684 - Attachment description: [default] fix patch v1.0 → [default] fix v1.0 [landed:default]
Attachment #605691 - Attachment description: [mozilla-aurora][mozilla-beta][mozilla-release] enable the test v1.1 → [aurora,beta,release] enable v1.1
Attachment #605691 - Flags: review?(anthony.s.hughes) → review+
Attachment #605691 - Attachment description: [aurora,beta,release] enable v1.1 → [aurora,beta,release] enable v1.1 [landed]
Please verify that the test passes on all branches and reopen if not.
Status: ASSIGNED → RESOLVED
Closed: 13 years ago13 years ago
Resolution: --- → FIXED
Whiteboard: [mozmill-test-failure][mozmill-test-skipped] → [mozmill-test-failure]
Status: RESOLVED → REOPENED
Depends on: 727131
Resolution: FIXED → ---
Before we flip a pref I would suggest that we use the default behavior of Firefox to test it.
Attached patch fix v2 (aurora) (obsolete) (deleted) — Splinter Review
Changed the pref and also moved it to a constant
Assignee: vlad.mozbugs → remus.pop
Status: REOPENED → ASSIGNED
Attachment #609664 - Flags: review?(vlad.mozbugs)
Attachment #609664 - Flags: feedback?(hskupin)
Comment on attachment 609664 [details] [diff] [review] fix v2 (aurora) >+const PREF_LOAD_IN_BACKGROUND = "browser.search.context.loadInBackground"; > const TIMEOUT = 5000; Looks good but please separate both from each other. Those constants are from different areas.
Attachment #609664 - Flags: feedback?(hskupin) → feedback+
Comment on attachment 609664 [details] [diff] [review] fix v2 (aurora) ++
Attachment #609664 - Flags: review?(vlad.mozbugs)
Attachment #609664 - Flags: review?(anthony.s.hughes)
Attachment #609664 - Flags: review+
Comment on attachment 609664 [details] [diff] [review] fix v2 (aurora) As I have mentioned in my feedback please get the patch updated. It removes one unnecessary round of reviews.
Attachment #609664 - Flags: review?(anthony.s.hughes) → review-
Attached patch patch v3 (aurora) [landed] (deleted) — Splinter Review
Added a empty line for delimiting the new constant.
Attachment #609664 - Attachment is obsolete: true
Attachment #610068 - Flags: review?(hskupin)
Comment on attachment 610068 [details] [diff] [review] patch v3 (aurora) [landed] So on which branches do I have to land this patch? Default and Aurora? the patch description is misleading here.
Attachment #610068 - Flags: review?(hskupin) → review+
Attached patch patch v2 (default) [landed] (deleted) — Splinter Review
This fixes the teardownModule which was not clearing the right preference. I also added a pref constant as in the aurora fix.
Attachment #610130 - Flags: review?(hskupin)
Attachment #610130 - Flags: review?(hskupin) → review+
Status: ASSIGNED → RESOLVED
Closed: 13 years ago13 years ago
Keywords: checkin-needed
Resolution: --- → FIXED
Attachment #610068 - Attachment description: patch v3 (aurora) → patch v3 (aurora) [landed]
Attachment #610130 - Attachment description: patch v2 (default) → patch v2 (default) [landed]
Thanks Henrik. Remus, please verify this is fixed with tomorrow's testruns.
Product: Mozilla QA → Mozilla QA Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: