Closed Bug 1041544 Opened 10 years ago Closed 10 years ago

browser_certViewer.js leaks until shutdown when run as a directory instead of a full suite

Categories

(Core :: Security: PSM, defect)

Product:
Core
Component:
Security: PSM
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED DUPLICATE of bug 1041537

People

(Reporter: jmaher, Unassigned)

References

Details

in bug 992911 we are working on running our tests per directory instead of as a full suite.  When we do that, browser_certViewer.js fails on linux64 debug (bc2) and windows debug (bc3):
https://tbpl.mozilla.org/?tree=Try&rev=e5fd341d6175

You can see a link to the raw logs here:
https://tbpl.mozilla.org/php/getParsedLog.php?id=44146991&tree=Try

We are planning on switching to this mode of running tests in 2 weeks and would like to do what we can to fix the tests.
David, can you look at this and see what we might be able to do?
Flags: needinfo?(dkeeler)
of course I forgot we have an almost identical failure (linux debug bc2, windows debug bc3) in browser_certificateManagerLeak.js.

Unfortunately there are only 3 tests in this directory and this accounts for 2 of the 3 :(
Is there something other than "--run-by-dir" I need to do to reproduce this locally? I'm on a linux 64-bit machine, and nothing is leaking when I run `./mach mochitest-browser --run-by-dir security/manager/ssl/tests/mochitest/'.
Flags: needinfo?(dkeeler)
are you using a debug build?

here is a link to the build that I had on try (from friday):
http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/jmaher@mozilla.com-e5fd341d6175/try-linux64-debug/firefox-33.0a1.en-US.linux-x86_64.tar.bz2

by running it with --test-path=security/manager/ssl/tests/mochitest/browser it should reproduce, maybe you have to run it 3 or 4 times to see it, but if you run it 5 times and it doesn't fail, we should look at solving this another way (either on try, or with a loaner).

I have learned over time that our machines we run tests on end up finding issues that we can never get to reproduce locally- although half the time we can solve it locally.
I used the build you linked to from try, but I still can't reproduce locally. As it happens, I requested a loaner test machine in bug 1042198 to investigate another bug I can't reproduce locally, so maybe I'll have more luck with that.
Depends on: 1042198
Hey David, did you have any luck with the loaner machine?
Flags: needinfo?(dkeeler)
Well, I did manage to vnc into the loaner and get the test suite running, but even with the build you provided, I wasn't able to reproduce the issue. Is there a set of instructions I can follow?
Flags: needinfo?(dkeeler)
Flags: needinfo?(jmaher)
It appears that the work that ttaubert has done in bug 1041537 fixes this.  Here is a try server with this test running:
https://tbpl.mozilla.org/?tree=Try&rev=0b5eb184c71d
Status: NEW → RESOLVED
Closed: 10 years ago
Flags: needinfo?(jmaher)
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.