557809 - [MacOSX SeaMonkey] mochitest-chrome: "leaked 4224 bytes" related to isp rdf data (and nsMsgServiceProviderService)

Status: NEW

(SeaMonkey :: MailNews: Account Configuration, defect)

Description

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1270637578.1270639106.28683.gz
OS X 10.5 comm-central-trunk debug test mochitest-other on 2010/04/07 03:52:58
{
Leaked URLs:
  file:///builds/slave/comm-central-trunk-macosx-debug-unittest-mochitest-other/build/SeaMonkeyDebug.app/Contents/MacOS/isp/movemail.rdf
  file:///builds/slave/comm-central-trunk-macosx-debug-unittest-mochitest-other/build/SeaMonkeyDebug.app/Contents/MacOS/isp/rss.rdf


TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 6028 bytes during test execution
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of AtomImpl with size 20 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of CompositeDataSourceImpl with size 68 bytes each (136 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 3 instances of InMemoryDataSource with size 152 bytes each (456 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 16 instances of LiteralImpl with size 12 bytes each (192 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of LocalStoreImpl with size 40 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of RDFServiceImpl with size 276 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of RDFXMLDataSourceImpl with size 68 bytes each (136 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 37 instances of RDF_Assertion with size 28 bytes each (1036 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of nsBaseURLParser with size 12 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of nsGenericFactory with size 20 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of nsLocalFile with size 1048 bytes each (2096 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of nsMsgServiceProviderService with size 16 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of nsNameSpaceMap with size 4 bytes each (8 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 4 instances of nsNameSpaceMap::Entry with size 20 bytes each (80 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 29 instances of nsRDFResource with size 28 bytes each (812 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of nsStandardURL with size 184 bytes each (368 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 36 instances of nsStringBuffer with size 8 bytes each (288 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 9 instances of nsVoidArray with size 4 bytes each (36 bytes total)
}

KaiRo, I suggest to use a leak threshold until bug 427198 is fixed.

Comment 1

[Robert Kaiser]

As we can't apply leak thresholds as nicely as previously and the change wouldn't even make things sctually green, I won't set any leak thresholds for now.

Comment 2

[Serge Gautherie (:sgautherie)]

(In reply to comment #1)

> As we can't apply leak thresholds as nicely as previously and the change

Ah :-( Why is that?

> wouldn't even make things sctually green, I won't set any leak thresholds for
> now.

Well, it would at least make this suite fully pass ... and let us see if there are other (intermittent) leaks...

Comment 3

[Robert Kaiser]

(In reply to comment #2)
> (In reply to comment #1)
> 
> > As we can't apply leak thresholds as nicely as previously and the change
> 
> Ah :-( Why is that?

Because the packaged test factory doesn't accept leak threshold params for everything, and before packaged factories we used our own separate factory that did.

> > wouldn't even make things sctually green, I won't set any leak thresholds for
> > now.
> 
> Well, it would at least make this suite fully pass ... and let us see if there
> are other (intermittent) leaks...

Let's concentrate on getting the rest to pass - and hope a fix for this leak can be found in time.

Comment 4

[Serge Gautherie (:sgautherie)]

(In reply to comment #3)

> Because the packaged test factory doesn't accept leak threshold params for
> everything, and before packaged factories we used our own separate factory that
> did.

Oh. Is there a bug filed for this c-c regression / m-c enhancement?

> Let's concentrate on getting the rest to pass

(To me, less summaries with 'LEAK' is better concentration...)

Comment 5

[Robert Kaiser]

(In reply to comment #4)
> (In reply to comment #3)
> 
> > Because the packaged test factory doesn't accept leak threshold params for
> > everything, and before packaged factories we used our own separate factory that
> > did.
> 
> Oh. Is there a bug filed for this c-c regression / m-c enhancement?

I don't know, and I'm not even sure how far patches for that would be accepted, I just don't have time to care about everything. <rant> I wanted to get to bed earlier today, and now it's past 3am again and I still sit here working...</rant>

> > Let's concentrate on getting the rest to pass
> 
> (To me, less summaries with 'LEAK' is better concentration...)

I understand, but I on the other hand always feel that LEAK is printed everywhere when we are cheating away any of them with a "threshold" sign (which is just another way of saying "we ignore some amount of leaks by pretending they are not there at all"...

Comment 6

[Serge Gautherie (:sgautherie)]

(In reply to comment #0)

Leak has a little changed in the meantime:

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1283190776.1283192510.4911.gz&fulltext=1
OS X 10.5 comm-central-trunk debug test mochitest-other on 2010/08/30 10:52:56
{
...

WARNING: Leaking the RDF Service.: file /builds/slave/comm-central-trunk-macosx-debug/build/mozilla/rdf/build/nsRDFModule.cpp, line 198

Leaked URLs:
file:///builds/slave/comm-central-trunk-macosx-debug-unittest-mochitest-other/build/SeaMonkeyDebug.app/Contents/MacOS/isp/movemail.rdf
file:///builds/slave/comm-central-trunk-macosx-debug-unittest-mochitest-other/build/SeaMonkeyDebug.app/Contents/MacOS/isp/rss.rdf

TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 4224 bytes during test execution
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of AtomImpl with size 20 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of CompositeDataSourceImpl with size 68 bytes each (136 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of GenericFactory with size 16 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 3 instances of InMemoryDataSource with size 152 bytes each (456 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 16 instances of LiteralImpl with size 12 bytes each (192 bytes total)
}


(In reply to comment #1)
> .. and the change wouldn't even make things sctually green,
> I won't set any leak thresholds for now.

Eventually, all perma-orange are fixed.
Only this leak keeps this suite orange.

What do you suggest we do about it?
(As there hasn't been progress on bug 427198 yet :-|)

Comment 7

[Robert Kaiser]

(In reply to comment #6)
> What do you suggest we do about it?

I suggest we find and fix the leak.