Open Bug 1775712 Opened 2 years ago Updated 1 day ago

Intermittent browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | single tracking bug

Categories

(Firefox :: Toolbars and Customization, defect, P3)

defect

Tracking

()

REOPENED
Tracking Status
firefox108 --- affected

People

(Reporter: jmaher, Unassigned, NeedInfo)

References

Details

(Keywords: intermittent-failure, intermittent-testcase, leave-open, Whiteboard: [stockwell disabled])

Attachments

(4 files)

No description provided.

Additional information about this bug failures and frequency patterns can be found by running: ./mach test-info failure-report --bug 1775712

Blanket S4 for all the new tracking bugs because they don't seem to be high-frequency. We can increase the severity if frequency goes up (or clear it to re-surface this to triagers, please.)

Severity: -- → S4

There have been 30 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows10-32-2004-qr
  • windows10-64-2004-asan-qr
  • windows10-64-2004-qr
Flags: needinfo?(gijskruitbosch+bugs)

There have been 34 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows10-64-2004-qr
  • windows10-64-2004-asan-qr

There have been 37 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows10-64-2004-qr
  • windows10-64-2004-asan-qr
  • macosx1015-64-qr

I've tried looking into this a few times but keep not getting very far. The test passes in verify mode on my windows machine on an opt build (which is the configuration that fails on infra). It's not obvious to me from the logs what exactly is failing. The screenshots are... odd. For the most recent failure log, they show us failing in testBookmarkButtonPress which tries to open the star button, but the screenshot is from when the test continues in a new task that shows the bookmarks menu button, after timing out. That seems like a framework problem to me - the screenshot should be taken immediately. As it is, I don't understand why the bookmark button click didn't work - on the screenshot that did get taken, the button is present and visible.

Aryx, do you know who could help with the screenshots / test framework issue, so we get more reliable debugging information here? Also, any chance of some kind of regression window?

Finally, Marco, as this seems to reliably fail primarily for the bookmarks star button (though I've spotted 1 that fails for the reload/stop button), any idea why clicking that might not open the bookmarks panel? There is already code that tries to ensure the button has finished updating:

    // The button ignores activation while the bookmarked status is being
    // updated. So, wait for it to finish updating.
    await TestUtils.waitForCondition(
      () => BookmarkingUI.status != BookmarkingUI.STATUS_UPDATING
    );

https://searchfox.org/mozilla-central/rev/1564027eba41993868973e4a7b93bea942a16b78/browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js#214-235

Flags: needinfo?(mak)
Flags: needinfo?(aryx.bugmail)

The issue with screenshots out of sync with the logged test failure is tracked by bug 1759582 and might have always affected the test framework. We don't have anybody working on desktop test frameworks at the moment.

Flags: needinfo?(aryx.bugmail)

(In reply to :Gijs (he/him) from comment #19)

Finally, Marco, as this seems to reliably fail primarily for the bookmarks star button (though I've spotted 1 that fails for the reload/stop button), any idea why clicking that might not open the bookmarks panel? There is already code that tries to ensure the button has finished updating:

I looked at it and came to the same conclusions, I'm not sure why it would not open the panel and I couldn't find an obvious culprit.
The only thing that comes to my mind, is diffently from other buttons, the star button is hidden on empty pages, maybe the forceFocus call happens before it's fully visible.
I'll try to run with some logging on Try.

Update:

There have been 41 failures within the last 7 days:

  • 1 failure on OS X 10.15 WebRender debug
  • 1 failure on OS X 10.15 WebRender Shippable opt
  • 37 failures on Windows 10 x64 2004 asan WebRender opt
  • 1 failure on Windows 10 x64 2004 WebRender debug
  • 1 failure on Windows 10 x64 2004 WebRender Shippable opt

Recent failure log: https://treeherder.mozilla.org/logviewer?job_id=394889576&repo=mozilla-central&lineNumber=2198

[task 2022-10-30T22:35:37.351Z] 22:35:37     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Sidebar hidden after press - 
[task 2022-10-30T22:35:37.351Z] 22:35:37     INFO - Leaving test bound testSidebarsButtonPress
[task 2022-10-30T22:35:37.352Z] 22:35:37     INFO - Entering test bound testBookmarkButtonPress
[task 2022-10-30T22:35:37.353Z] 22:35:37     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use <!DOCTYPE html>." {file: "https://example.com/" line: 0}]
[task 2022-10-30T22:35:37.353Z] 22:35:37     INFO - Buffered messages finished
[task 2022-10-30T22:35:37.353Z] 22:35:37     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Test timed out - 
[task 2022-10-30T22:35:37.354Z] 22:35:37     INFO - Not taking screenshot here: see the one that was previously logged
[task 2022-10-30T22:35:37.354Z] 22:35:37     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Uncaught exception received from previously timed out test - focus listener on #editBookmarkPanel not removed before the end of test
[task 2022-10-30T22:35:37.355Z] 22:35:37     INFO - Entering test bound testBookmarksmenuButtonPress
[task 2022-10-30T22:35:37.355Z] 22:35:37     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Bookmarks Menu shown after toolbar button pressed - 

I am working to get tests running on windows 11 and I ran into this failure. On win/asan this test is already failing often (60%+ of the time) and on windows in general it has a fairly high failure rate.

for windows11, asan is perma fail, and the rest of windows I see this on and off, enough to consider disabling it.

I would vote for disabling on win/asan, and then seeing how this looks going forward.

(In reply to Joel Maher ( :jmaher ) (UTC -8) from comment #26)

for windows11, asan is perma fail, and the rest of windows I see this on and off, enough to consider disabling it.

can I run that on Try, or pass you a patch to test?
I can also try running locally on win11, though I'm having some issues getting the asan build

Flags: needinfo?(mak)
Flags: needinfo?(jmaher)
Flags: needinfo?(gijskruitbosch+bugs)

I am happy to test it out, or you can with a command like this:
./mach try fuzzy --no-artifact --worker-override="win10-64-2004=gecko-t/win11-64-2009-alpha" -q 'test-windows10 asan' browser/base/content/test/keyboard

Flags: needinfo?(jmaher)

ok, definitely the problem is that the panel doesn't open, hopefully with asan it happens often enough to reproduce it locally, even if I can't get a build I should be able to use the one generated by Try, passing it to mach mochitest.

Flags: needinfo?(mak)

I just saw a case where what it fails is testReloadButtonPress, it basically does the same thing as on the bookmarks button, but on the reload button. So the problem may be broader than just the bookmarks button.

I'm still unable to build windows 11 asan locally, rust is unhappy with loading some dlls and crashes, a non asan build works fine. on Try I could only confirm that the element is visible and waiting to focus it even longer doesn't help (document.activeElement is correct anyway). Repeating the key press also doesn't help. The element seems to properly receive the keypress event based on logging. It sounds more like a popupmanager issue. I'll try to get an interactive task and see if I can find more, but it's unlikely.

I think we can disable the test on windows ASAN anyway, considered how often it happens.

Assignee: nobody → ncsoregi
Status: NEW → ASSIGNED
Pushed by imoraru@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/0ee89dc44faf Disable browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js on Windows Asan. r?#intermittent-reviewers CLOSED TREE
Status: ASSIGNED → RESOLVED
Closed: 2 years ago
Resolution: --- → FIXED
Target Milestone: --- → 108 Branch

Should probably not close this just because we disabled the test...

Status: RESOLVED → REOPENED
Resolution: FIXED → ---

The last thing we could try here is a Try run with MOZ_PROFILER_STARTUP =1 and hopefully the profile timeline may show something interesting.
I have no further ideas, for now.

Flags: needinfo?(mak)

There have been 35 total failures in the last 7 days.
There are:

  • 1 failure on windows11-64-2009-shippable-qr opt
  • 17 failures on windows11-64-2009-qr debug
  • 1 failure on windows11-64-2009-ccov-qr opt
  • 13 failures on windows11-32-2009-qr debug
  • 3 failures on macosx1015-64-qr debug

Recent failure log.

Gijs, as the owner of this component, can you help us assign the bug to someone? Thank you.

Flags: needinfo?(gijskruitbosch+bugs)

(In reply to Marco Bonardo [:mak] from comment #33)

I'm still unable to build windows 11 asan locally, rust is unhappy with loading some dlls and crashes, a non asan build works fine. on Try I could only confirm that the element is visible and waiting to focus it even longer doesn't help (document.activeElement is correct anyway). Repeating the key press also doesn't help. The element seems to properly receive the keypress event based on logging. It sounds more like a popupmanager issue. I'll try to get an interactive task and see if I can find more, but it's unlikely.

Emilio, do you have any suggestions on what we could try to debug this further and/or wrt the popup manager and/or focus handling? The lack of success in reproducing locally, and gradually disabling on more platforms (first windows asan, now looks like windows debug is the next victim, and it's already disabled on all of Linux and win7), is a little frustrating.

Flags: needinfo?(gijskruitbosch+bugs) → needinfo?(emilio)

Probably something like MOZ_LOG=Focus:5,FocusNavigation:5 might be useful?

Flags: needinfo?(emilio)

(In reply to Emilio Cobos Álvarez (:emilio) from comment #70)

Probably something like MOZ_LOG=Focus:5,FocusNavigation:5 might be useful?

I set this up and got e.g. https://treeherder.mozilla.org/logviewer?job_id=418349606&repo=try&lineNumber=5214 but I'm struggling to make much out of that.

I also asked pernosco self-serve to repro the linux debug failures (I re-enabled on platforms currently disabled on central/autoland) but haven't heard back yet.

(In reply to :Gijs (he/him) from comment #73)

(In reply to Emilio Cobos Álvarez (:emilio) from comment #70)

Probably something like MOZ_LOG=Focus:5,FocusNavigation:5 might be useful?

I set this up and got e.g. https://treeherder.mozilla.org/logviewer?job_id=418349606&repo=try&lineNumber=5214 but I'm struggling to make much out of that.

I also asked pernosco self-serve to repro the linux debug failures (I re-enabled on platforms currently disabled on central/autoland) but haven't heard back yet.

Pernosco didn't get anywhere. :-(

Emilio, is the log useful to you at all?

Flags: needinfo?(emilio)

So it's pretty hard to follow because I don't see the log messages while the test is running:

[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Buffered messages logged at 22:01:42
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Entering test bound setPref
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Leaving test bound setPref
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Entering test bound testAppMenuButtonPress
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Focus inside app menu after toolbar button pressed - 
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Leaving test bound testAppMenuButtonPress
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - Entering test bound testAppMenuButtonWrongKey
[task 2023-06-06T22:03:12.452Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | App menu is closed after tab - 
[task 2023-06-06T22:03:12.453Z] 22:03:12     INFO - Leaving test bound testAppMenuButtonWrongKey
[task 2023-06-06T22:03:12.454Z] 22:03:12     INFO - Entering test bound testLibraryButtonPress
[task 2023-06-06T22:03:12.455Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Focus inside Library menu after toolbar button pressed - 
[task 2023-06-06T22:03:12.455Z] 22:03:12     INFO - Leaving test bound testLibraryButtonPress
[task 2023-06-06T22:03:12.456Z] 22:03:12     INFO - Entering test bound testDeveloperButtonPress
[task 2023-06-06T22:03:12.457Z] 22:03:12     INFO - Buffered messages logged at 22:01:43
[task 2023-06-06T22:03:12.459Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Focus inside Developer menu after toolbar button pressed - 
[task 2023-06-06T22:03:12.459Z] 22:03:12     INFO - Leaving test bound testDeveloperButtonPress
[task 2023-06-06T22:03:12.460Z] 22:03:12     INFO - Entering test bound testDeveloperButtonWrongKey
[task 2023-06-06T22:03:12.461Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Developer menu not open after tab - 
[task 2023-06-06T22:03:12.462Z] 22:03:12     INFO - Leaving test bound testDeveloperButtonWrongKey
[task 2023-06-06T22:03:12.463Z] 22:03:12     INFO - Entering test bound testPageActionsButtonPress
[task 2023-06-06T22:03:12.464Z] 22:03:12     INFO - Buffered messages logged at 22:01:44
[task 2023-06-06T22:03:12.465Z] 22:03:12     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use “<!DOCTYPE html>”." {file: "https://example.com/" line: 0}]
[task 2023-06-06T22:03:12.466Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Focus inside Page Actions menu after toolbar button pressed - 
[task 2023-06-06T22:03:12.467Z] 22:03:12     INFO - Leaving test bound testPageActionsButtonPress
[task 2023-06-06T22:03:12.468Z] 22:03:12     INFO - Entering test bound testBackForwardButtonPress
[task 2023-06-06T22:03:12.469Z] 22:03:12     INFO - Buffered messages logged at 22:01:45
[task 2023-06-06T22:03:12.470Z] 22:03:12     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use “<!DOCTYPE html>”." {file: "https://example.com/1" line: 0}]
[task 2023-06-06T22:03:12.471Z] 22:03:12     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use “<!DOCTYPE html>”." {file: "https://example.com/2" line: 0}]
[task 2023-06-06T22:03:12.471Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Location changed after back button pressed - 
[task 2023-06-06T22:03:12.472Z] 22:03:12     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Location changed after forward button pressed - 
[task 2023-06-06T22:03:12.472Z] 22:03:12     INFO - Buffered messages logged at 22:01:46
[task 2023-06-06T22:03:12.473Z] 22:03:12     INFO - Leaving test bound testBackForwardButtonPress
[task 2023-06-06T22:03:12.473Z] 22:03:12     INFO - Entering test bound testReloadButtonPress
[task 2023-06-06T22:03:12.473Z] 22:03:12     INFO - Buffered messages logged at 22:01:47
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use “<!DOCTYPE html>”." {file: "https://example.com/1" line: 0}]
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Waiting for button to be enabled.
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Focusing button
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Pressing space on the button
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Waiting for load.
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - Buffered messages finished
[task 2023-06-06T22:03:12.474Z] 22:03:12     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Test timed out - 
[task 2023-06-06T22:03:12.536Z] 22:03:12     INFO - GECKO(1583) | MEMORY STAT | vsize 3571MB | residentFast 584MB | heapAllocated 316MB
[task 2023-06-06T22:03:12.539Z] 22:03:12     INFO - TEST-OK | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | took 90254ms
[task 2023-06-06T22:03:12.541Z] 22:03:12     INFO - Not taking screenshot here: see the one that was previously logged
[task 2023-06-06T22:03:12.542Z] 22:03:12     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Found a tab after previous test timed out: https://example.com/1 - 

But afterwards I see a bunch of log stuff. Do you know why that might be? I would expect log messages interleaved between the INFO - stuff.

Flags: needinfo?(emilio) → needinfo?(gijskruitbosch+bugs)

(In reply to Emilio Cobos Álvarez (:emilio) from comment #75)

But afterwards I see a bunch of log stuff. Do you know why that might be? I would expect log messages interleaved between the INFO - stuff.

I think it's a function of log buffering. Because the entire test directory runs, we don't normally print individual test assertions or info calls. When the test fails, the framework dumps messages anyway. But the MOZ_LOG stuff is output immediately. The log stuff you see after the test is from after the test - it's the bits before the buffered dump that relate to what happened mid-test.

I think using requestCompleteLog will fix this by just printing test logs immediately instead of buffering them. I've repushed with that:

https://treeherder.mozilla.org/jobs?repo=try&revision=59aa61d4a113cbfdd1a8c425742eb48787684f18

(In reply to :Gijs (he/him) from comment #77)

https://treeherder.mozilla.org/jobs?repo=try&revision=59aa61d4a113cbfdd1a8c425742eb48787684f18

This appears to have worked (direct link), but I'm still not sure off-hand how to interpret the MOZ_LOG stuff. Emilio?

Flags: needinfo?(gijskruitbosch+bugs) → needinfo?(emilio)

Oh, so I see what's going on.

The test that times out starts here. That test focuses the button, waits for the UI to be ready, then presses space.

That doesn't do anything tho, because the focus has been cleared in between. And I see why. The problem is that forceFocus only makes the element focusable temporarily, but then removes the focusability. So then the focus fixup (step 15 here) comes in and clears the focus.

So fixes is either:

  • Make sure the element is focusable while waiting for the bookmarking UI.
  • Wait before calling forceFocus, rather than after.
Flags: needinfo?(emilio)

Do you have cycles to try either of those things? Otherwise I can though not today.

Flags: needinfo?(gijskruitbosch+bugs)

https://treeherder.mozilla.org/jobs?repo=try&revision=41fe896ab6a98a81829ce0e3b05beb79f6455f08

I don't really understand why the reload button case would fail but I've tried to address that by giving forceFocus an optional parameter that keeps the button focusable a bit longer.

Assignee: nobody → gijskruitbosch+bugs
Flags: needinfo?(gijskruitbosch+bugs)
Attachment #9339967 - Attachment description: Bug 1775712 - make browser_toolbarButtonKeyPress.js more reliable, r?mak → Bug 1775712 - make browser_toolbarButtonKeyPress.js more reliable by having forceFocus run a callback before removing focusability, r?mak
Pushed by gijskruitbosch@gmail.com: https://hg.mozilla.org/integration/autoland/rev/14edc9bcbe56 make browser_toolbarButtonKeyPress.js more reliable by having forceFocus run a callback before removing focusability, r=mak
Keywords: leave-open
Whiteboard: [stockwell disable-recommended] → [stockwell unknown]
Whiteboard: [stockwell disable-recommended] → [stockwell unknown]
Whiteboard: [stockwell disable-recommended] → [stockwell unknown]

There have been 61 total failures in the last 7 days.
There are:

  • 5 failures on Linux 18.04 x64 WebRender asan opt
  • 8 failures on Linux 18.04 x64 WebRender debug
  • 8 failures on Linux 18.04 x64 WebRender tsan opt
  • 5 failures on OS X 10.15 WebRender debug
  • 9 failures on Windows 11 x86 22H2 WebRender debug
  • 17 failures on Windows 11 x64 22H2 asan WebRender opt
  • 1 failure on Windows 11 x64 22H2 CCov WebRender opt
  • 7 failures on Windows 11 x64 22H2 WebRender debug
  • 1 failure on Windows 11 x64 22H2 WebRender opt

Failure lines that were classified with this bug are: Test timed out -, Uncaught exception received from previously timed out test bound testReloadButtonPress - at resource://testing-common/BrowserTestUtils.sys.mjs:519 - Error: The window unloaded while we were waiting for the browser to load - this should never happen. (which was originally Uncaught exception received from previously timed out test bound testBookmarkButtonPress - focus listener on #editBookmarkPanel not removed before the end of test before there were added changes) and Found a tab after previous test timed out: https://example.com/1 - ( which appeared after https://bugzilla.mozilla.org/show_bug.cgi?id=1775712#c84).

Recent failure log.

[task 2023-06-26T02:30:39.032Z] 02:30:39     INFO - TEST-PASS | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Location changed after forward button pressed - 
[task 2023-06-26T02:30:39.032Z] 02:30:39     INFO - Buffered messages logged at 02:28:40
[task 2023-06-26T02:30:39.033Z] 02:30:39     INFO - Leaving test bound testBackForwardButtonPress
[task 2023-06-26T02:30:39.033Z] 02:30:39     INFO - Entering test bound testReloadButtonPress
[task 2023-06-26T02:30:39.033Z] 02:30:39     INFO - Console message: [JavaScript Warning: "This page is in Quirks Mode. Page layout may be impacted. For Standards Mode use “<!DOCTYPE html>”." {file: "https://example.com/1" line: 0}]
[task 2023-06-26T02:30:39.033Z] 02:30:39     INFO - Waiting for button to be enabled.
[task 2023-06-26T02:30:39.034Z] 02:30:39     INFO - Focusing button
[task 2023-06-26T02:30:39.034Z] 02:30:39     INFO - Pressing space on the button
[task 2023-06-26T02:30:39.034Z] 02:30:39     INFO - Waiting for load.
[task 2023-06-26T02:30:39.034Z] 02:30:39     INFO - Buffered messages finished
[task 2023-06-26T02:30:39.034Z] 02:30:39     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Test timed out - 
[task 2023-06-26T02:30:39.300Z] 02:30:39     INFO - GECKO(2468) | MEMORY STAT | vsize 130551872MB | residentFast 1265MB
[task 2023-06-26T02:30:39.302Z] 02:30:39     INFO - TEST-OK | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | took 124240ms
[task 2023-06-26T02:30:39.305Z] 02:30:39     INFO - Not taking screenshot here: see the one that was previously logged
[task 2023-06-26T02:30:39.305Z] 02:30:39     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Found a tab after previous test timed out: https://example.com/1 - 
[task 2023-06-26T02:30:39.500Z] 02:30:39     INFO - checking window state
[task 2023-06-26T02:30:39.703Z] 02:30:39     INFO - TEST-START | browser/base/content/test/keyboard/browser_toolbarKeyNav.js
[task 2023-06-26T02:30:56.274Z] 02:30:56     INFO - GECKO(2468) | 1687746656273	Marionette	TRACE	Received observer notification browser-delayed-startup-finished
[task 2023-06-26T02:31:12.766Z] 02:31:12     INFO - GECKO(2468) | MEMORY STAT | vsize 130551856MB | residentFast 1299MB
[task 2023-06-26T02:31:12.768Z] 02:31:12     INFO - TEST-OK | browser/base/content/test/keyboard/browser_toolbarKeyNav.js | took 33064ms
[task 2023-06-26T02:31:12.876Z] 02:31:12     INFO - checking window state
[task 2023-06-26T02:31:15.596Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2618
[task 2023-06-26T02:31:15.604Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2563
[task 2023-06-26T02:31:15.641Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2993
[task 2023-06-26T02:31:15.654Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 3124
[task 2023-06-26T02:31:15.806Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2537
[task 2023-06-26T02:31:15.859Z] 02:31:15     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2685
[task 2023-06-26T02:31:17.136Z] 02:31:17     INFO - GECKO(2468) | Completed ShutdownLeaks collections in process 2468
[task 2023-06-26T02:31:17.137Z] 02:31:17     INFO - TEST-START | Shutdown
[task 2023-06-26T02:31:17.137Z] 02:31:17     INFO - Browser Chrome Test Summary
[task 2023-06-26T02:31:17.138Z] 02:31:17     INFO - Passed:  151
[task 2023-06-26T02:31:17.139Z] 02:31:17     INFO - Failed:  2
[task 2023-06-26T02:31:17.140Z] 02:31:17     INFO - Todo:    0
[task 2023-06-26T02:31:17.140Z] 02:31:17     INFO - Mode:    e10s
[task 2023-06-26T02:31:17.140Z] 02:31:17     INFO - *** End BrowserChrome Test Results ***
[task 2023-06-26T02:31:17.669Z] 02:31:17     INFO - GECKO(2468) | 1687746677668	Marionette	TRACE	Received observer notification quit-application
[task 2023-06-26T02:31:17.671Z] 02:31:17     INFO - GECKO(2468) | 1687746677669	Marionette	INFO	Stopped listening on port 2828
[task 2023-06-26T02:31:17.675Z] 02:31:17     INFO - GECKO(2468) | 1687746677674	Marionette	DEBUG	Marionette stopped listening
[task 2023-06-26T02:31:17.813Z] 02:31:17     INFO - Not taking screenshot here: see the one that was previously logged
[task 2023-06-26T02:31:17.814Z] 02:31:17     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js | Uncaught exception received from previously timed out test bound testReloadButtonPress - at resource://testing-common/BrowserTestUtils.sys.mjs:519 - Error: The window unloaded while we were waiting for the browser to load - this should never happen.
[task 2023-06-26T02:31:17.814Z] 02:31:17     INFO - Stack trace:
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - listener@resource://testing-common/BrowserTestUtils.sys.mjs:519:15
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - EventListener.handleEvent*browserLoaded/<@resource://testing-common/BrowserTestUtils.sys.mjs:534:27
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - browserLoaded@resource://testing-common/BrowserTestUtils.sys.mjs:484:12
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - testReloadButtonPress/<@chrome://mochitests/content/browser/browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js:174:37
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - async*withNewTab@resource://testing-common/BrowserTestUtils.sys.mjs:146:22
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - async*testReloadButtonPress@chrome://mochitests/content/browser/browser/base/content/test/keyboard/browser_toolbarButtonKeyPress.js:168:26
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - handleTask@chrome://mochikit/content/browser-test.js:1131:26
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - _runTaskBasedTest@chrome://mochikit/content/browser-test.js:1203:18
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - async*Tester_execTest@chrome://mochikit/content/browser-test.js:1345:14
[task 2023-06-26T02:31:17.815Z] 02:31:17     INFO - nextTest/<@chrome://mochikit/content/browser-test.js:1120:14
[task 2023-06-26T02:31:17.816Z] 02:31:17     INFO - SimpleTest.waitForFocus/<@chrome://mochikit/content/tests/SimpleTest/SimpleTest.js:1056:13
[task 2023-06-26T02:31:17.901Z] 02:31:17     INFO - GECKO(2468) | JavaScript error: chrome://mochikit/content/browser-test.js, line 1194: TypeError: can't access property "shift", currentScope.__tasks is null
[task 2023-06-26T02:31:21.404Z] 02:31:21     INFO - TEST-INFO | Main app process: exit 0

Hello Gijs! Could you please take another look at this bug?

Thank you!

Flags: needinfo?(gijskruitbosch+bugs)
Whiteboard: [stockwell disable-recommended] → [stockwell needswork:owner]
Pushed by csabou@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/f87590a1b732 Disable browser_toolbarButtonKeyPress.js for frequent failures. r=intermittent-reviewers,jmaher DONTBUILD
Whiteboard: [stockwell disable-recommended] → [stockwell disabled]

Emilio, any idea what's going on with the reload button bit? Here is a log with MOZ_LOG stuff.

It looks like focusing isn't doing anything? I wonder if the content page is stealing focus and we need to somehow move focus to the parent or something like this? But I'm struggling to parse the MOZ_LOG output successfully...

Flags: needinfo?(gijskruitbosch+bugs) → needinfo?(emilio)

RecvSetFocusedBrowsingContext does indicate that, yeah... Btw, maybe we should look into splitting up this test, so that we don't loose so much test coverage when there's a flaky one...

Flags: needinfo?(emilio)

Do we have a regression window for comment 160? I was on PTO so I missed it at the time but it looks like this test started failing a lot more - did we enable a new platform, or was this triggered by a code change?

Flags: needinfo?(aryx.bugmail)

The macOS AArch and Linux opt/Shippable platforms had been running the test before, this looks like an issue from a code change. The Pocket button popup is open but empty.

The frequency increased on 2024-09-05 around here. Could it be from bug 1768527?

Flags: needinfo?(aryx.bugmail)

(In reply to Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout) from comment #167)

The macOS AArch and Linux opt/Shippable platforms had been running the test before, this looks like an issue from a code change. The Pocket button popup is open but empty.

The frequency increased on 2024-09-05 around here. Could it be from bug 1768527?

That bug primarily changed tests and some mobile code. I don't know that the layout change could be responsible here. Emilio?

Flags: needinfo?(emilio)

Seems viewport related, so... maybe? Hiro?

Flags: needinfo?(emilio) → needinfo?(hikezoe.birchill)

(In reply to :Gijs (he/him) from comment #170)

(In reply to Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout) from comment #167)

The macOS AArch and Linux opt/Shippable platforms had been running the test before, this looks like an issue from a code change. The Pocket button popup is open but empty.

The frequency increased on 2024-09-05 around here. Could it be from bug 1768527?

That bug primarily changed tests and some mobile code. I don't know that the layout change could be responsible here. Emilio?

The code change also affects on desktops, that's said, it only affects the top level content documents, not the browser ones. That's odd.

Anyways, I am going to poke the test locally on my Linux box, hope the test fails there.

(In reply to Hiroyuki Ikezoe (:hiro) from comment #172)

Anyways, I am going to poke the test locally on my Linux box, hope the test fails there.

Any luck?

Assignee: gijskruitbosch+bugs → nobody

(In reply to :Gijs (he/him) from comment #175)

(In reply to Hiroyuki Ikezoe (:hiro) from comment #172)

Anyways, I am going to poke the test locally on my Linux box, hope the test fails there.

Any luck?

No luck locally, I just confirmed that the test actually failed on try, which means it's harder to diagnose. :/

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: