585449 - [xpcshell] test_askuser.js fails: "ASSERTION: StoreMDNSentFlag failed: 'NS_SUCCEEDED(rv)', file .../mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp, line 1097"

Status: RESOLVED FIXED

(MailNews Core :: Testing Infrastructure, defect)

Description

[Robert Kaiser]

test_askuser.js fails on all platforms.

Example log:
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1281272587.1281273215.5505.gz

Last good build on Linux:
rev:8b1b4701dbab / moz:a9501fcd9c25

First bad build on Linux:
rev:af73207d80c1 / moz:f298567cdf2c

Last good build on Mac:
rev:8b1b4701dbab / moz:a9501fcd9c25

First bad build on Mac:
rev:af73207d80c1 / moz:2625ed42ed89

Last good build on Windows:
rev:8b1b4701dbab / moz:0e0e2516f04e

First bad build on Windows:
rev:af73207d80c1 / moz:4980d2c1252c

Comment 1

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1281333772.1281335988.14398.gz#err4
WINNT 5.2 comm-central-trunk debug test xpcshell on 2010/08/08 23:02:52
{
TEST-INFO | e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\xpcshell\tests\test_mdn\unit\test_askuser.js | running test ...
TEST-UNEXPECTED-FAIL | e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\xpcshell\tests\test_mdn\unit\test_askuser.js | test failed (with xpcshell return code: -2147483645), see following log:
  >>>>>>>
  ### XPCOM_MEM_LEAK_LOG defined -- logging leaks to c:\docume~1\seabld\locals~1\temp\tmpdjxsgc\runxpcshelltests_leaks.log
pldhash: for the table at address 019111A8, the given entrySize of 48 probably favors chaining over double hashing.
Directory request for: SysD that we (mailDirService.js) are not handling, leaving it to another handler.
Directory request for: MailD that we (mailDirService.js) are not handling, leaving it to another handler.
Directory request for: MFCaF that we (mailDirService.js) are not handling, leaving it to another handler.
Directory request for: DefRt that we (mailDirService.js) are not handling, leaving it to another handler.
nsNativeModuleLoader::LoadModule("e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\seamonkey\components\gkdebug.dll") - load FAILED, rv: 80520012, error:
	<unknown; can't get error from NSPR>
nsNativeModuleLoader::LoadModule("e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\seamonkey\components\jsperf.dll") - load FAILED, rv: 80520012, error:
	<unknown; can't get error from NSPR>
nsNativeModuleLoader::LoadModule("e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\seamonkey\components\xpcomsample.dll") - load FAILED, rv: 80520012, error:
	<unknown; can't get error from NSPR>
nsNativeModuleLoader::LoadModule("e:\builds\slave\comm-central-trunk-win32-debug-unittest-xpcshell\build\seamonkey\components\xpctest.dll") - load FAILED, rv: 80520012, error:
	<unknown; can't get error from NSPR>
WARNING: No valid default account found, just using first (FIXME): file e:/builds/slave/comm-central-trunk-win32-debug/build/mailnews/base/src/nsMsgAccountManager.cpp, line 776
TEST-INFO | (xpcshell/head.js) | test 1 pending
[e-mail]
###!!! ASSERTION: StoreMDNSentFlag failed: 'NS_SUCCEEDED(rv)', file e:/builds/slave/comm-central-trunk-win32-debug/build/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp, line 1097
nsStringStats
 => mAllocCount:           4252
 => mReallocCount:          437
 => mFreeCount:            2303  --  LEAKED 1949 !!!
 => mShareCount:           8859
 => mAdoptCount:            173
 => mAdoptFreeCount:        169  --  LEAKED 4 !!!
0x00000000063D679B
0x00000000063D660F
0x00000000063D6748
xpcom_core!mozilla::CallStack::`default constructor closure'+0x00000000000CD8AA
gklayout!DumpJSStack+0x0000000000C37A83
gklayout!DumpJSStack+0x0000000000C355B8
gklayout!DumpJSStack+0x0000000000C35249
gklayout!DumpJSStack+0x0000000000C44B81
mozjs!JS_GetPropertyDescriptorById+0x00000000000BEFE1
mozjs!JS_GetPropertyDescriptorById+0x00000000000BED28
mozjs!JS_GetPropertyDescriptorById+0x000000000009B909
mozjs!JS_GetPropertyDescriptorById+0x00000000000AC9E8
mozjs!JS_GetPropertyDescriptorById+0x000000000009C92F
mozjs!JS_GetPropertyDescriptorById+0x0000000000028E3C
mozjs!JS_GetPropertyDescriptorById+0x0000000000028F5F
0x0000000000405B58
0x00000000004049F9
0x0000000000412136
0x0000000000411F8D
kernel32!ProcessIdToSessionId+0x0000000000000209
###!!! ASSERTION: StoreMDNSentFlag failed: 'NS_SUCCEEDED(rv)', file e:/builds/slave/comm-central-trunk-win32-debug/build/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp, line 1097

  <<<<<<<
}

Comment 2

[Robert Kaiser]

Not that this log explains anything. And given our current look at tests, i.e. almost everything in perma-organe, no test failure is major unless we _know_ that it's caused by an actual functionality problem.

Comment 3

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1281340078.1281340840.991.gz&fulltext=1#err0
Linux comm-central-trunk debug test xpcshell on 2010/08/09 00:47:58
{
TEST-INFO | /builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/xpcshell/tests/test_mdn/unit/test_askuser.js | running test ...
TEST-UNEXPECTED-FAIL | /builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/xpcshell/tests/test_mdn/unit/test_askuser.js | test failed (with xpcshell return code: -6), see following log:
  >>>>>>>
  ### XPCOM_MEM_LEAK_LOG defined -- logging leaks to /tmp/tmpKl53F2/runxpcshelltests_leaks.log
nsNativeModuleLoader::LoadModule("/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgkdebug.so") - load FAILED, rv: 80004005, error:
	/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgkdebug.so: cannot open shared object file: No such file or directory
nsNativeModuleLoader::LoadModule("/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libjsperf.so") - load FAILED, rv: 80004005, error:
	/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libjsperf.so: cannot open shared object file: No such file or directory
nsNativeModuleLoader::LoadModule("/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libxpcomsample.so") - load FAILED, rv: 80004005, error:
	/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libxpcomsample.so: cannot open shared object file: No such file or directory
nsNativeModuleLoader::LoadModule("/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libxpctest.so") - load FAILED, rv: 80004005, error:
	/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libxpctest.so: cannot open shared object file: No such file or directory
pldhash: for the table at address 0x9baa8b0, the given entrySize of 48 probably favors chaining over double hashing.
Directory request for: MailD that we (mailDirService.js) are not handling, leaving it to another handler.
Directory request for: MFCaF that we (mailDirService.js) are not handling, leaving it to another handler.
Directory request for: DefRt that we (mailDirService.js) are not handling, leaving it to another handler.
TEST-INFO | (xpcshell/head.js) | test 1 pending
[e-mail]
###!!! ASSERTION: StoreMDNSentFlag failed: 'NS_SUCCEEDED(rv)', file /builds/slave/comm-central-trunk-linux-debug/build/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp, line 1097
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libmsgmdn.so +0x00001F19]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libmsgmdn.so +0x000051FD]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libmsgmdn.so +0x000066BE]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libmsgmdn.so +0x00006888]
NS_InvokeByIndex_P+0x0000002F [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libxpcom_core.so +0x000DC873]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgklayout.so +0x00D4B28C]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgklayout.so +0x00D4D989]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgklayout.so +0x00D48E40]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/components/libgklayout.so +0x00D5302D]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x000D48D3]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x000D0770]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x000D345D]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x000BE2EF]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x000D2A76]
JS_EvaluateUCScriptForPrincipals+0x00000135 [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x00025C43]
JS_EvaluateScriptForPrincipals+0x0000007F [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/libmozjs.so +0x00025D06]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/xpcshell +0x00005E3E]
UNKNOWN [/builds/slave/comm-central-trunk-linux-debug-unittest-xpcshell/build/seamonkey/xpcshell +0x00006A0D]
__libc_start_main+0x000000DC [/lib/libc.so.6 +0x00015DEC]
###!!! ASSERTION: StoreMDNSentFlag failed: 'NS_SUCCEEDED(rv)', file /builds/slave/comm-central-trunk-linux-debug/build/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp, line 1097

  <<<<<<<
}

Comment 4

[Serge Gautherie (:sgautherie)]

(In reply to comment #2)

Sorry to have disturbed you: I'll get back to my own business :-/

Comment 5

[Robert Kaiser]

No disturbance if you work on actually fixing this failures (I'll cheer a lot for that!) - or if someone else doing it can really figure out what the massive amount of cryptic stuff you have posted in here.

I pondered looking into the cause of this, but after that dump of non-understable stuff that looks like garbage to me at a first glance, though surely there's something useful in it, I'll first look into things that are cleaner to me.
I intentionally only did link the logs in comment #0 and pasted log stuff in comments is more unreadable than the logs themselves, in 80% of all cases I have looked into, the log messages were unhelpful at best, irritating at the worst, and on-hands debugging was needed to find out the the hell was up anyhow. The linked logs won't expire anyhow, AFAIK.

Comment 6

[Robert Kaiser]

The test has been added in bug 151244 on 2010-04-08, it started to fail when bug 558543 landed at 2010-06-29, adding the second test in this dir.

Comment 7

[Robert Kaiser]

Our MDN bar landed 2010-05-10 from bug 539066.

Still, bienvenu landed the patch that happened to break this test for us - any idea what's going on there?

Comment 8

[David :Bienvenu]

assertions in debug builds will kill xpcshell tests, though I don't know why you're seeing an assertion.

Comment 9

[Robert Kaiser]

(In reply to comment #8)
> assertions in debug builds will kill xpcshell tests, though I don't know why
> you're seeing an assertion.

Hmm, right. Are you implying that this just doesn't show up as a failure in Thunderbird because you guys don't test on debug?

Comment 10

[David :Bienvenu]

I run all the xpcshell tests with my debug builds very frequently and I don't see this problem. So it's not that simple - I'm just telling you a little bit more about why the test seemed to fail for SM.

Comment 11

[Robert Kaiser]

I guess the question is why we are running into the assertion at all - and possibly a lead could be that we didn't before bug 558543 landed.
The test passed correctly before that MDN patch landed - interestingly that one did only touch mailnews/ though.

Comment 12

[Serge Gautherie (:sgautherie)]

(In reply to comment #6)
> The test has been added in bug 151244 on 2010-04-08, it started to fail when
> bug 558543 landed at 2010-06-29, adding the second test in this dir.

(In reply to comment #7)
> Our MDN bar landed 2010-05-10 from bug 539066.

Fwiw,
http://brasstacks.mozilla.com/topfails/test/SeaMonkey?name=chrome://mochikit/content/browser/toolkit/content/tests/browser/browser_Services.js

"2010-05-08 02:30": oldest report (in the db)
though the error may have changed in the meantime...

"2010-06-29 11:55": was already failing until 2010-06-24 22:03,
I would assume the test suite just didn't run in that 5 days timeframe.

Comment 13

[Serge Gautherie (:sgautherie)]

Correct URL:
http://brasstacks.mozilla.com/topfails/test/SeaMonkey?name=xpcshell/tests/test_mdn/unit/test_askuser.js

Comment 14

[Justin Wood (:Callek)]

I don't feel comfortable blocking on this perma-orange until we know if its an actually issue in the application, or if its an issue in the test. That said, I wholehartedly want to devote resources to getting our tests green and this fixed.

Comment 15

[Bruno 'Aqualon' Escherl]

What happens is the following:

http://mxr.mozilla.org/comm-central/source/mailnews/extensions/mdn/test/unit/test_askuser.js?mark=46-46,54-55

There we have an undefined key, which results in a null key used as parameter for the call of StoreMDNSentFlag:

http://mxr.mozilla.org/comm-central/source/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp?mark=173-174,181-181#173

The MarkMDNSent call fails of course with a null key, so we hit the assertion at http://mxr.mozilla.org/comm-central/source/mailnews/extensions/mdn/src/nsMsgMdnGenerator.cpp?mark=1096-1097#1094

Comment 16

[David :Bienvenu]

Could that test be expecting that msgHdr.messageKey gets set to 0?

Comment 17

[Bruno 'Aqualon' Escherl]

Forgot to mention that Thunderbird debug also hits that assertion.

(In reply to comment #16)
> Could that test be expecting that msgHdr.messageKey gets set to 0?
Setting this explicitly in the test also results in a null key and thus the assertion gets triggered.

Comment 18

[Bruno 'Aqualon' Escherl]

Attachment: add message to local mail folder that we can process (checked in - comment #19)

Now I understood comment #16, the test needs to add a message to the local mail folder, just as test_mdnFlags.js does.

Comment 19

[Misak Khachatryan]

Pushed:
http://hg.mozilla.org/comm-central/rev/50cab7663d15