388241 - shutdown crash at ~nsCOMPtr coming from nsBindingValues::ClearBindingSet

Status: RESOLVED WORKSFORME

(SeaMonkey :: General, defect)

Description

[Robert Kaiser]

Attachment: stacktrace

Build identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a7pre) Gecko/200707151919 SeaMonkey/2.0a1pre

I had a shutdown crash in ~nsCOMPtr, coming from nsBindingValues::ClearBindingSet today, this is an IRC snippet from the attached stacktrace:

<KaiRo> can someone tell me where this shutdown crash is actually crashing? http://pastebin.mozilla.org/140352
<timeless> freeing a xul template rdf node?
<timeless> line 7/6
<biesi> probably a released object in that comptr
<biesi> guess some code forgot an addref

I don't know where to file this correctly or what might be the cause, so filing in General for now...

Comment 1

[Wayne Mery (:wsmwk)]

I was about to ask Robert about maybe this WFM, but I found 4 crashes in the past 11 months. Robert, have you seen this crasher since report?

3 FF in http://crash-stats.mozilla.com/report/list?query_search=signature&query_type=exact&query=nsBindingValues%3A%3AClearBindingSet%28%29&date=11%2F10%2F2009&range_value=4&range_unit=weeks&do_query=1&signature=nsBindingValues%3A%3AClearBindingSet%28%29 (not an exact match to robert's stack, eg bp-9c45ea0a-a2df-4e8a-9d71-94abb2091106 ). but  bp-97bc3581-0ee8-4d47-8f20-f60c72091116 FF 3.0 does matches Robert's stack. None of the crashes have comments

I searched SM 2.0x crashes with nsBindingValues::ClearBindingSet in top 10 frames of last 2 weeks, and 4 weeks in any product with nsBindingValues::ClearBindingSet in the top frame, plus 10 months of top frame exact match on  nsBindingValues::ClearBindingSet() 

Robert's stack
#6  nsCOMPtr<nsIRDFNode>::~nsCOMPtr nsCOMPtr.h:581
#7  0xb67b1c1a in nsBindingValues::ClearBindingSet nsRDFBinding.cpp:227
#8  0xb67b1d18 in nsBindingValues::~nsBindingValues nsRDFBinding.cpp:219
#9  0xb67d447c in nsXULTemplateResultRDF::~nsXULTemplateResultRDF nsXULTemplateResultRDF.cpp:70

Comment 2

[Robert Kaiser]

(In reply to comment #1)
> Robert, have you seen this crasher since report?

I haven't explicitely looked at any crashers sine quite some time, as I have other priorites now and the quote of anything happening with my reports was low enough that I decided to not care any more.

Comment 3

[Wayne Mery (:wsmwk)]

I don't find any examples in crash-stats for last ~2 months for FF or SM