428952 - crash [@ libnecko.so@0x3bd79] restarting for addon upgrade

Status: RESOLVED WORKSFORME

(SeaMonkey :: General, defect)

Description

[Tony Mechelynck [:tonymec]]

Restarting after enabling upgrade to Adblock Plus, Breakpad came up.
bp-60dcef19-0a25-11dd-b26f-001321b13766

Circumstances are ominously similar to those of bug 428796 but the stack-top is different, so, dupe or no dupe?

Here comes the crashing thread stack:
Frame  	Module  	Signature  	Source
0 	libnecko.so 	libnecko.so@0x3bd79 	
1 	libnecko.so 	libnecko.so@0x24b78 	
2 	libnecko.so 	libnecko.so@0x25f11 	
3 	libnecko.so 	libnecko.so@0x24ae1 	
4 	libpipnss.so 	nsHTTPDownloadEvent::Run() 	mozilla/security/manager/ssl/src/nsNSSCallbacks.cpp:119
5 	libxpcom_core.so 	libxpcom_core.so@0x582eb 	
6 	libxpcom_core.so 	libxpcom_core.so@0x26d4a 	
7 	libnecko.so 	libnecko.so@0x39df5 	
8 	libnecko.so 	libnecko.so@0x39fb7 	
9 	libnecko.so 	libnecko.so@0x1d65a 	
10 	libnecko.so 	libnecko.so@0x83342

Comment 1

[Tony Mechelynck [:tonymec]]

P.S. nsHTTPDownloadEvent + one address below and three above are identical in both crashes (cannot check further down because AFAICT current crash reports are truncated at depth 11).

Comment 2

[timeless]

w/o symbols you're not going to get particularly valid stack traces (and eventually stack crawlers will give up).

standard question:
are the datestamps for lib(necko/pipnss/xpcom_core).so similar?

Comment 3

[Tony Mechelynck [:tonymec]]

(In reply to comment #2)
> w/o symbols you're not going to get particularly valid stack traces (and
> eventually stack crawlers will give up).

IIUC there's a problem about posting symbols where the crash reporting system can find them -- I try to do my best, but...

> 
> standard question:
> are the datestamps for lib(necko/pipnss/xpcom_core).so similar?
> 

I think they were from nightlies from successive days... let me check... yes, this one was from Sm 2.0a1pre (Linux-i686 en-US) 2008041401 and the other from the 2008041301 build of the same.

Comment 4

[Tony Mechelynck [:tonymec]]

...and BTW, the 10-stack-frames-only problem was bug 428682 (fixed on 2008-04-18).

Comment 5

[Serge Gautherie (:sgautherie)]

Is this reproducible, or was it a "one time only" ?

Comment 6

[Tony Mechelynck [:tonymec]]

One time only, or maybe "two times only" if duped with bug 428796.

Comment 7

[Wayne Mery (:wsmwk)]

Tony, how's it looking now?

Comment 8

[Tony Mechelynck [:tonymec]]

Mozilla/5.0 (X11; Linux i686; rv:2.0b8pre) Gecko/20101212 Firefox/4.0b8pre SeaMonkey/2.1b2pre - Build ID: 20101212003013

haven't seen it at any recent time