Closed Bug 888811 Opened 11 years ago Closed 5 years ago

Re-enable browser_dbg_on-pause-raise.js on Linux

Categories

(DevTools :: Debugger, defect, P3)

Product:
DevTools
Component:
Debugger
Platform:
x86
Linux
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED INVALID

People

(Reporter: philor, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: intermittent-failure, Whiteboard: [leave open][test disabled on linux]) 

https://tbpl.mozilla.org/php/getParsedLog.php?id=24771185&tree=Mozilla-Central
Ubuntu VM 12.04 x64 mozilla-central pgo test mochitest-browser-chrome on 2013-06-30 20:46:03 PDT for push 7463e5548797
slave: tst-linux64-ec2-026

21:03:56     INFO -  TEST-INFO | chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js | switching to toolbox window.
21:04:25  WARNING -  TEST-UNEXPECTED-FAIL | chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js | Test timed out
(screenshot)
21:04:29     INFO -  DBG-FRONTEND: Destroying the DebuggerView
21:04:29     INFO -  DBG-FRONTEND: Destroying the ToolbarView
21:04:29     INFO -  DBG-FRONTEND: Destroying the OptionsView
21:04:29     INFO -  DBG-FRONTEND: Destroying the FilterView
21:04:29     INFO -  DBG-FRONTEND: Destroying the FilteredSourcesView
21:04:29     INFO -  DBG-FRONTEND: Destroying the FilteredFunctionsView
21:04:29     INFO -  DBG-FRONTEND: Destroying the ChromeGlobalsView
21:04:29     INFO -  DBG-FRONTEND: Destroying the StackFramesView
21:04:29     INFO -  DBG-FRONTEND: Destroying the SourcesView
21:04:29     INFO -  DBG-FRONTEND: Destroying the WatchExpressionsView
21:04:29     INFO -  DBG-FRONTEND: Destroying the GlobalSearchView
21:04:29     INFO -  DBG-FRONTEND: Destroying the DebuggerView panes
21:04:29     INFO -  DBG-FRONTEND: Destroying the DebuggerView editor
21:04:29     INFO -  DBG-FRONTEND: SourceScripts is disconnecting...
21:04:29     INFO -  DBG-FRONTEND: StackFrames is disconnecting...
21:04:29     INFO -  DBG-FRONTEND: ThreadState is disconnecting...
21:04:29     INFO -  INFO TEST-END | chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js | finished in 30296ms
21:04:29     INFO -  TEST-INFO | checking window state
21:04:29     INFO -  TEST-INFO | chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js | must wait for focus
21:04:29  WARNING -  TEST-UNEXPECTED-FAIL | chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js | uncaught exception - ReferenceError: info is not defined at chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js:46
21:04:29     INFO -  Stack trace:
21:04:29     INFO -      JS frame :: chrome://mochikit/content/tests/SimpleTest/SimpleTest.js :: simpletestOnerror :: line 1136
21:04:29     INFO -      native frame :: <unknown filename> :: <TOP_LEVEL> :: line 0
Any idea why this test doesn't work reliably in optimized Linux builds?
Flags: needinfo?(scrapmachines)
I know what happens to break the test. When the toolbox window is created (as in when you undock the toolbox), the .then method is called which adds an event handler on the main window to get focus and then it focuses the main window [method focusMainWindow in the file]. Somehow, the two events of main window getting focused and the newly created toolbox window getting focus are interfering with each other such that the event handler is not getting fired. :/

My understanding was that the approach that I took should work on linux. I guess I was wrong.

I have another approach which will surely work on opt and pgo and will not lead to any race condition on them, but will lead to race condition on debug and normal builds. It means that my test flow will have to be different for pgo + opt as compared to debug + normal builds. I have no idea how to do that :/
Flags: needinfo?(scrapmachines)
(In reply to Girish Sharma [:Optimizer] from comment #29)
> I have another approach which will surely work on opt and pgo and will not
> lead to any race condition on them, but will lead to race condition on debug
> and normal builds. It means that my test flow will have to be different for
> pgo + opt as compared to debug + normal builds. I have no idea how to do
> that :/

Care to elaborate on that other approach? Regular user builds are opt/pgo FWIW.
(In reply to Panos Astithas [:past] from comment #30)
> Care to elaborate on that other approach? Regular user builds are opt/pgo
> FWIW.

The other approach is in toolbox/tests/browser_toolbox_raise.js:50@testWindowHost

basically, here I wait until the newly opened toolbox window gets focus. Then I focus the main browser window.

But somehow, using this same approach in the debugger test was causing failures.
Test disabled on Linux.
https://hg.mozilla.org/integration/mozilla-inbound/rev/e9c5b7db433b
Whiteboard: [leave open][test disabled on linux]
Thanks!
Tweaking summary to avoid false positives with another frequent JS exception (dep of bug 920191).
Summary: Intermittent browser_dbg_bug883220_raise_on_pause.js | Test timed out, uncaught exception - ReferenceError: info is not defined at chrome://mochitests/content/browser/browser/devtools/debugger/test/browser_dbg_bug883220_raise_on_pause.js:46 → Intermittent browser_dbg_bug883220_raise_on_pause.js | Test timed out| uncaught exception...
Summary: Intermittent browser_dbg_bug883220_raise_on_pause.js | Test timed out| uncaught exception... → Intermittent browser_dbg_on-pause-raise.js | Test timed out
Hasn't happened in 3 months. Closing.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WORKSFORME
Nope, [leave open].
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Summary: Intermittent browser_dbg_on-pause-raise.js | Test timed out → Re-enable browser_dbg_on-pause-raise.js on Linux
Priority: -- → P3
Blocks: dbg-test
Product: Firefox → DevTools

This test file no longer exists.

Status: REOPENED → RESOLVED
Closed: 10 years ago → 5 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.