Closed Bug 790174 Opened 12 years ago Closed 12 years ago

Test failure 'Underlined title matches entered title - got 'communit', expected 'community'' in /testAwesomeBar/testCheckItemHighlight.js

Categories

(Mozilla QA Graveyard :: Mozmill Tests, defect, P2)

defect

Tracking

(firefox15 affected, firefox16 affected, firefox17 affected, firefox18 affected, firefox-esr10 unaffected)

RESOLVED FIXED
Tracking Status
firefox15 --- affected
firefox16 --- affected
firefox17 --- affected
firefox18 --- affected
firefox-esr10 --- unaffected

People

(Reporter: AndreeaMatei, Assigned: AndreeaMatei)

References

()

Details

(Whiteboard: [mozmill-test-failure])

Attachments

(2 files, 4 obsolete files)

testCheckItemHighlight.js fails at checking the underlined URL matches the entered URL. We have there a controller.assert() which needs to be changed.
Assignee: nobody → andreea.matei
Attached patch patch v1 (obsolete) — Splinter Review
Replaced controller.assert() with controller.waitFor() and did testrun on Ubuntu and Mac OS X 10.7.4.
Attachment #660018 - Flags: review?(hskupin)
Attachment #660018 - Flags: review?(dave.hunt)
Whiteboard: [mozmill-test-failure]
Comment on attachment 660018 [details] [diff] [review] patch v1 Review of attachment 660018 [details] [diff] [review]: ----------------------------------------------------------------- ::: tests/functional/testAwesomeBar/testCheckItemHighlight.js @@ +93,4 @@ > > // Check that the underlined URL matches the entered URL > underlined.forEach(function (element) { > + controller.waitFor(function() { Do we really need a waitFor() call here? The item is visible for sure so I think using assert is still the right way to go.
Attachment #660018 - Flags: review?(hskupin)
Attachment #660018 - Flags: review?(dave.hunt)
Attachment #660018 - Flags: review-
Attached patch patch v1.1 (obsolete) — Splinter Review
Used assert.equal as requested.
Attachment #660018 - Attachment is obsolete: true
Attachment #660026 - Flags: review?(hskupin)
Attachment #660026 - Flags: review?(dave.hunt)
OS: Mac OS X → All
Comment on attachment 660026 [details] [diff] [review] patch v1.1 Review of attachment 660026 [details] [diff] [review]: ----------------------------------------------------------------- ::: tests/functional/testAwesomeBar/testCheckItemHighlight.js @@ +94,5 @@ > > // Check that the underlined URL matches the entered URL > underlined.forEach(function (element) { > + assert.equal(element.toLowerCase(), LOCAL_TEST_PAGES[0].name, > + "Underlined URL matches entered URL"); Sorry, but I was wrong here. Looking again at the test failure it was really thrown for one of those underlined elements. So we should indeed try with waitFor(). Similar to your last patch but please update the message so that it also contains the aType parameter.
Attachment #660026 - Flags: review?(hskupin)
Attachment #660026 - Flags: review?(dave.hunt)
Attachment #660026 - Flags: review-
Attached patch patch v1.3 (obsolete) — Splinter Review
Switched back to waitFor().
Attachment #660026 - Attachment is obsolete: true
Attachment #660030 - Flags: review?(hskupin)
Attachment #660030 - Flags: review?(dave.hunt)
Comment on attachment 660030 [details] [diff] [review] patch v1.3 Review of attachment 660030 [details] [diff] [review]: ----------------------------------------------------------------- ::: tests/functional/testAwesomeBar/testCheckItemHighlight.js @@ +97,3 @@ > return element.toLowerCase() === LOCAL_TEST_PAGES[0].name; > }, "Underlined " + aType + " matches entered " + aType + " - got '" + > element.toLowerCase() + "', expected '" + LOCAL_TEST_PAGES[0].name + "'"); As I have mentioned on one of the other bugs lately we do not have a chance to print out the current value. Please adjust that message accordingly.
Attachment #660030 - Flags: review?(hskupin)
Attachment #660030 - Flags: review?(dave.hunt)
Attachment #660030 - Flags: review-
Attached patch patch v2Splinter Review
Updated the message, leaving aType parameter and the expected value.
Attachment #660030 - Attachment is obsolete: true
Attachment #660069 - Flags: review?(hskupin)
Attachment #660069 - Flags: review?(dave.hunt)
Comment on attachment 660069 [details] [diff] [review] patch v2 Review of attachment 660069 [details] [diff] [review]: ----------------------------------------------------------------- I think that looks fine. I will get it in right now.
Attachment #660069 - Flags: review?(hskupin)
Attachment #660069 - Flags: review?(dave.hunt)
Attachment #660069 - Flags: review+
Landed on default: http://hg.mozilla.org/qa/mozmill-tests/rev/56c01340f467 If we do not regress anything lets get it backported tomorrow.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Attached patch debugging dumps (obsolete) — Splinter Review
Dump statements in toolbars.js for the getUnderlinedText method. The description path was checked and it's accurate. My return for the values array is "community", all the time.
Attachment #660838 - Flags: review?(hskupin)
Attachment #660838 - Flags: review?(dave.hunt)
Comment on attachment 660838 [details] [diff] [review] debugging dumps Review of attachment 660838 [details] [diff] [review]: ----------------------------------------------------------------- I think we should add even more debug here. You could dump the type that's passed to getUnderlinedText so we can easily distinguish between the title and url calls. Also, could you add the number of the current iteration to the dump statements - it's currently quite hard to follow and looks like length does not match the number of dump statements. It might also help to use newlines more sparingly, to group dump statements together. I'd also be curious to see what the following would return at the point of failure, to see if getting the first richlistitem again helps: dump(locationBar.autoCompleteResults.getUnderlinedText(locationBar.autoCompleteResults.getResult(0), aType) + "\n");
Attachment #660838 - Flags: review?(hskupin)
Attachment #660838 - Flags: review?(dave.hunt)
Attachment #660838 - Flags: review-
Attached patch debugging dumpsSplinter Review
Indeed the length did not match the number of dumps, that is because childNodes has 3 main nodes (0, 1, 2), plus others like item, iterator and length. And the length takes only the 3 principal nodes. In our for each (node in description.childNodes) we take all the nodes present. For debugging, I replaced it with a simple for that would go from 0 to description.childNodes.length, so we can see the details on the main nodes and also to dump the current step. Great idea to try again to get the richlistItem underlined. Tried on a separate function to dump, which would be called in the error message but because of all the functions wrapped there I got undefined. So I have put that into the error message and tested. Rearranged the new lines for a clearer view. Hopefully will have more data about the failure with all this.
Attachment #660838 - Attachment is obsolete: true
Attachment #661227 - Flags: review?(hskupin)
Attachment #661227 - Flags: review?(dave.hunt)
Comment on attachment 661227 [details] [diff] [review] debugging dumps Review of attachment 661227 [details] [diff] [review]: ----------------------------------------------------------------- Landed as: http://hg.mozilla.org/qa/mozmill-tests/rev/7e5bccbef014 (default) Let's see what the next failures tell us...
Attachment #661227 - Flags: review?(hskupin)
Attachment #661227 - Flags: review?(dave.hunt)
Attachment #661227 - Flags: review+
It failed here, both on Mac OS X 10.7.4: * http://mozmill-ci.blargon7.com/#/functional/report/671677a5d9d5ca25f3cf5ae1c4c18bc4 * http://mozmill-ci.blargon7.com/#/functional/report/671677a5d9d5ca25f3cf5ae1c4c06e8f It seems it's not retrieving the whole string on the second try either. I am not able to see the dumps results, could you or Henrik please help with that?
All the dump() output will be visible in the console on the Jenkins instance. You can't find it in the report. So work together with Vlad here.
Status: REOPENED → ASSIGNED
The dump statements should not have been put into the toolbars module but in the test itself. Given the caching of output there will be no good way to compare the output for a specific test. I backed this out: http://hg.mozilla.org/qa/mozmill-tests/rev/afe7826d71ab All I was able to find was: Type: title Nodes length: 3 Iteration number: 0 Node name: #text innerHTML: undefined Iteration number: 1 Node name: span innerHTML: Communit I got here Iteration number: 2 Node name: #text innerHTML: undefined Values: Communit Type: title Nodes length: 3 Iteration number: 0 Node name: #text innerHTML: undefined Iteration number: 1 Node name: span innerHTML: Communit I got here Iteration number: 2 Node name: #text innerHTML: undefined Values: Communit
Depends on: 792394
Priority: -- → P2
This should be fixed now by the landing of the patch on bug 792394.
Status: ASSIGNED → RESOLVED
Closed: 12 years ago12 years ago
Resolution: --- → FIXED
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: