Closed Bug 1067559 Opened 10 years ago Closed 7 years ago

Intermittent test_reserved.html | Should not be able to navigate off-domain parent by setting location

Categories

(Core :: Security, defect, P3)

x86
Linux
defect

Tracking

()

RESOLVED DUPLICATE of bug 1263213

People

(Reporter: RyanVM, Unassigned)

References

Details

Smells like a timing issue.

https://tbpl.mozilla.org/php/getParsedLog.php?id=48116154&tree=Mozilla-Inbound

Ubuntu VM 12.04 mozilla-inbound debug test mochitest-e10s-1 on 2014-09-15 11:08:58 PDT for push 8ed5d27728dc
slave: tst-linux32-spot-181

12:11:52     INFO -  2216 INFO TEST-PASS | /tests/docshell/test/navigation/test_reserved.html | Should be able to navigate off-domain top by setting location.
12:11:52     INFO -  2217 INFO TEST-PASS | /tests/docshell/test/navigation/test_reserved.html | Should be able to navigate off-domain top by calling window.open.
12:11:52     INFO -  2218 INFO TEST-PASS | /tests/docshell/test/navigation/test_reserved.html | Should be able to navigate off-domain top by submitting form.
12:11:52     INFO -  2219 INFO TEST-PASS | /tests/docshell/test/navigation/test_reserved.html | Should be able to navigate off-domain top by targeted hyperlink.
12:11:52     INFO -  2220 INFO TEST-UNEXPECTED-FAIL | /tests/docshell/test/navigation/test_reserved.html | Should not be able to navigate off-domain parent by setting location. - expected PASS
Hrm.  Is this a regression?

This test assumes that nothing other than it is loading content-type windows.  So if something in our chrome got changed to do that randomly, it could be screwing up this test...

I could add logging to see what's being loaded, if that would be helpful.
Flags: needinfo?(ryanvm)
(In reply to Boris Zbarsky [:bz] from comment #3)
> Hrm.  Is this a regression?

Comment 0 is the first instance I recall seeing. Note that debug mochitest-e10s-1 is relatively new in and of itself. That said, we seem to hit it around once a day per the TBPLbot comments above, so that does seem to confirm relatively recent breakage.

> I could add logging to see what's being loaded, if that would be helpful.

SGTM
Flags: needinfo?(ryanvm)
Oh, interesting.  This _only_ happens with debug e10s tests?
https://tbpl.mozilla.org/?tree=Try&rev=86e5066125c0 with some logging; will see if I can get it orange by retriggering.
Now that I read this test again it's really not making any sense to me.  In particular, navigation via location set is always allowed, so I'm not sure why this test is trying to test (presumably incorrectly!) that it's not allowed in some cases.
See Also: → 1263213
Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.
Priority: -- → P3
We seem to have moved on to using bug 1263213
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.