562583 - test_update.js fails run_test_14

Status: VERIFIED FIXED

(Toolkit :: Add-ons Manager, defect)

Description

[Robert Strong (they/them - no direct email)]

Attachment: console output with an interactive test run

Using a debug build. I had to run the test interactively to get output since it didn't write to the test's log file

************************************************************
* Call to xpconnect wrapped JSObject produced this error:  *
[Exception... "'JavaScript component does not have a method named: "readIntValue
"' when calling method: [nsIWindowsRegKey::readIntValue]"  nsresult: "0x80570030
 (NS_ERROR_XPC_JSOBJECT_HAS_NO_FUNCTION_NAMED)"  location: "JS frame :: file:///
d:/moz/addonsmgr/ff-dbg/dist/bin/modules/AddonUpdateChecker.jsm :: UpdateParser
:: line 393"  data: no]
************************************************************

I suspect it is caused by the following call to readIntValue since the test and the add-ons code doesn't call readIntValue.
http://mxr.mozilla.org/mozilla-central/source/toolkit/system/windowsproxy/nsWindowsSystemProxySettings.cpp#224

line 393 of AddonUpdateChecker.jsm is when the xhr is opened
this.request.open("GET", aUrl, true);

Comment 1

[Robert Strong (they/them - no direct email)]

I initially thought that the test not completing was due to this failure but since it occurs several times it may actually be due to something else causing the following
************************************************************
* Call to xpconnect wrapped JSObject produced this error:  *
[Exception... "'[JavaScript Error: "XPIProvider.installLocationsByName is null" {file: "file:///d:/moz/addonsmgr/ff-dbg/dist/bin/modules/XPIProvider.jsm" line: 2766}]' when calling method: [mozIStorageStatementCallback::handleResult]"  nsresult: "0x80570021 (NS_ERROR_XPC_JAVASCRIPT_ERROR_WITH_DETAILS)"  location: "<unknown>"  data: yes]
************************************************************
************************************************************
* Call to xpconnect wrapped JSObject produced this error:  *
[Exception... "'[JavaScript Error: "this.addonListeners is null" {file: "file:///d:/moz/addonsmgr/ff-dbg/dist/bin/modules/AddonManager.jsm" line: 281}]' when calling method: [mozIStorageStatementCallback::handleCompletion]"  nsresult: "0x80570021 (NS_ERROR_XPC_JAVASCRIPT_ERROR_WITH_DETAILS)"  location: "<unknown>"  data: yes]
************************************************************
js> 

When the test is ran normally (e.g. not interactively) I get
$ make SOLO_FILE=test_update.js -C test check-one
make: Entering directory `/d/moz/addonsmgr/ff-dbg/toolkit/mozapps/extensions/test'
d:/moz/mozilla-build/python25/python2.5.exe -u /d/moz/addonsmgr/mozilla/config/pythonpath.py \
          -I/d/moz/addonsmgr/mozilla/build \
          /d/moz/addonsmgr/mozilla/testing/xpcshell/runxpcshelltests.py \
          --symbols-path=../../../../dist/crashreporter-symbols \
          --test-path=test_update.js \
          ../../../../dist/bin/xpcshell \
          ../../../../_tests/xpcshell/test_extensionmanager/xpcshell
Traceback (most recent call last):
  File "d:/moz/addonsmgr/mozilla/config/pythonpath.py", line 40, in <module>
    execfile(script, {'__name__': '__main__', '__file__': script})
  File "d:/moz/addonsmgr/mozilla/testing/xpcshell/runxpcshelltests.py", line 525, in <module>
    main()
  File "d:/moz/addonsmgr/mozilla/testing/xpcshell/runxpcshelltests.py", line 521, in main
    if not xpcsh.runTests(args[0], testdirs=args[1:], **options.__dict__):
  File "d:/moz/addonsmgr/mozilla/testing/xpcshell/runxpcshelltests.py", line 451, in runTests
    <<<<<<<""" % (test, self.getReturnCode(proc), stdout)
IOError: [Errno 12] Not enough space
make: *** [check-one] Error 1
make: Leaving directory `/d/moz/addonsmgr/ff-dbg/toolkit/mozapps/extensions/test

Comment 2

[Robert Strong (they/them - no direct email)]

This is passing on the addonsmgr tinderbox so I suspect it might have something to do with my system which is Win7 64bit.

Comment 3

[Robert Strong (they/them - no direct email)]

All tests in test_update.js pass except for run_test_14

Comment 4

[Robert Strong (they/them - no direct email)]

And flip flopping the add-ons along with their checks so [e-mail] is the one with applyBackgroundUpdates = false makes it so run_test_14 passes.

Comment 5

[Robert Strong (they/them - no direct email)]

Attachment: patch for reference

Comment 6

[Robert Strong (they/them - no direct email)]

Also for reference, the test never passed over several complete xpcshell runs against extensions as well as many more runs against the individual test. With the patch it has passed over several complete xpcshell runs as well as many more runs against the individual test. If it weren't for the consistency with how it fails on my system and it succeeds on the tinderbox I'd guess that which add-on finishes each step first is nondeterministic.

Comment 7

[Dave Townsend [:mossop]]

Something about test_update.js seems to be racy so I've disabled this for the trunk landing.

Comment 8

[Gary [:streetwolf52]]

Tried the latest hourly with the new Add-On manager.  Breaks alot of my Add-Ons.

Comment 9

[Robert Strong (they/them - no direct email)]

Please file a bug... this bug is specifically about one of the tests failing

Comment 10

[Gary [:streetwolf52]]

I'll wait until the next nightly.

Comment 11

[Robert Strong (they/them - no direct email)]

In the future please only comment in bugs that are for your specific problem or file a new bug if there isn't one. Thanks

Comment 12

[Dave Townsend [:mossop]]

Test disabled in http://hg.mozilla.org/mozilla-central/rev/5b027af3af29

Comment 13

[Dave Townsend [:mossop]]

Attachment: patch rev 1

The problem is that we never know whether the background check is going to find the new version for addon1 or addon8 first since it relies on network conditions (internal only but still). This gets around it by using a custom InstallListener that doesn't care which onNewInstall event arrives first, just cares that the subsequent download events are for the correct add-on.

Comment 14

[Dave Townsend [:mossop]]

landed: http://hg.mozilla.org/mozilla-central/rev/77358a9fb4a3

Comment 15

[Henrik Skupin [:whimboo][⌚️UTC+2] (away 02/17 - 02/21)]

No orangeness. Marking as verified fixed based on check-in.