9730 - complicated RDF crash

Status: VERIFIED WORKSFORME

(Core Graveyard :: RDF, defect, P3)

Description

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Here are the conditions that need to be met to cause this crash:

You need to try to read a nonexistant RDF datasource by HTTP (not file:), and
then when RemoteDataSource::Refresh() throws an exception, also call
RDFService::GetDataSource().

Test case to be attached.

(And all I was trying to do was to get HasMoreElements to throw an exception so
I could test exception handling in eval()...)

Tested on:
* Linux mozilla.org opt build 1999-07-12-18-M8.

Comment 1

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: test page causing crash

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Oh yes, stack trace (is this useful from the opt builds?  nobody ever answers
that question...  at least it shows why I picked "RDF" component.):

#0  0x4034124f in nsCOMPtr_base::assign_with_AddRef ()
#1  0x40a9801f in RDFContentSinkImpl::Init ()
#2  0x400b4f55 in CNavDTD::CloseBody ()
#3  0x400b567a in CNavDTD::CloseContainer ()
#4  0x400b57a2 in CNavDTD::CloseContainersTo ()
#5  0x400b589c in CNavDTD::CloseContainersTo ()
#6  0x400b26d4 in CNavDTD::DidBuildModel ()
#7  0x400bbd00 in nsParser::DidBuildModel ()
#8  0x400bc51c in nsParser::ResumeParse ()
#9  0x400bc8c7 in nsParser::OnStopRequest ()

Comment 3

[Chris Waterson]

This is looking an _awful_ lot like 11112; or if it's not, then we're not even
getting far enough to see the problem because of 11112. I'm going to mark a
dependency for now; once 11112 is fixed, we can see if it's a dup or not.

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

This started well before bug 11112.

Comment 5

[chris hofmann]

got to move to m10, unless waterson can pull rabbits out of hats

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Verified fixed.