573055 - Reproducible [@ nsNPAPIPluginInstance::InitializePlugin()] crashes / Java

Status: RESOLVED WORKSFORME

(Core Graveyard :: Plug-ins, defect)

Description

[NoOp]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a5pre) Gecko/20100510 SeaMonkey/2.1a1
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a5pre) Gecko/20100510 SeaMonkey/2.1a1

I am getting reproducible crashes w/Java. Crash reports are here w/steps to reproduce in the first. Signatures are nsNPAPIPluginInstance::InitializePlugin():

<http://crash-stats.mozilla.com/report/index/bp-56525554-5f02-4c96-af58-954aa2100613>

Steps to reproduce: 1. Open 2.1a and point the brower to
http://java.com/en/ 2. Click on the 'Do I have Java' link directly below
the "Free Java Download" button. 3. On
http://java.com/en/download/installed.jsp click on the "Veryify Java
version" button & let it verify that you have: (Version 6 Update 20)."
4. Now click on the browser 'Back' button. Crash.

<http://crash-stats.mozilla.com/report/index/f1cb93ef-2b09-4e78-b04e-f41692100613>

Build identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a5pre)
Gecko/20100510 SeaMonkey/2.1a1


Reproducible: Always

Steps to Reproduce:
1. Open 2.1a and point the brower to http://java.com/en/
2. Click on the 'Do I have Java' link directly below the "Free Java Download" button.
3. On http://java.com/en/download/installed.jsp click on the "Veryify Java
version" button & let it verify that you have: (Version 6 Update 20)."
4. Now click on the browser 'Back' button.
Actual Results:  
Crash. 

Expected Results:  
Expect SeaMonkey to go to the previous browser page.

I cannot reproduce on SeaMonkey 2.0.4 or 2.0.5. I can reproduce with 2.1a1 on multiple machines. Confirmed as reproducible on other linux builds in 
Newsgroups: mozilla.dev.apps.seamonkey
Subject: [2.1a1] Reproducible crashes w/Java

Crash reports:
<http://crash-stats.mozilla.com/report/index/bp-56525554-5f02-4c96-af58-954aa2100613>

<http://crash-stats.mozilla.com/report/index/f1cb93ef-2b09-4e78-b04e-f41692100613>

Third repeat w/clean test profile:
<http://crash-stats.mozilla.com/report/index/bp-dc96d9ac-6571-48be-a431-c384a2100613>

Crash report submitted today just to ensure that I can still repeat since the initial testing on 06/13/2010:
<http://crash-stats.mozilla.com/report/index/bp-3b104bd5-c552-4a8d-93c6-624062100618>

Comment 1

[Hartmut Figge]

(In reply to comment #0)

> I cannot reproduce on SeaMonkey 2.0.4 or 2.0.5. I can reproduce with 2.1a1 on
> multiple machines. Confirmed as reproducible on other linux builds in 
> Newsgroups: mozilla.dev.apps.seamonkey
> Subject: [2.1a1] Reproducible crashes w/Java

That could be me. :)

I have now tried to reproduce this bug with the help of the archive of my selfcompiled builds. That showed this regression range:

hafi@e675 ~ $ fenster 1001260000-1001270724
2010-01-25 15:00:00 PDT
2010-01-26 22:24:00 PDT

Comment 2

[Jens Hatlak (:InvisibleSmiley)]

http://hg.mozilla.org/mozilla-central/pushloghtml?startdate=2010-01-25+15%3A00%3A00&enddate=2010-01-26+22%3A24%3A00
shows three checkins with "plugin":
http://hg.mozilla.org/mozilla-central/rev/aa9d33fc9493
http://hg.mozilla.org/mozilla-central/rev/623eefcb43d7
http://hg.mozilla.org/mozilla-central/rev/795803409745

FWIW, I'd place my bet on the first one. Note that I know nothing about that C++ stuff.

Comment 3

[Hartmut Figge]

(In reply to comment #2)

> http://hg.mozilla.org/mozilla-central/rev/aa9d33fc9493
> http://hg.mozilla.org/mozilla-central/rev/623eefcb43d7
> http://hg.mozilla.org/mozilla-central/rev/795803409745

None of them can backed out easily for testing purpose. Lots of eg. 'Hunk #6 FAILED at 765.'.

Comment 4

[[e-mail]]

(In reply to comment #3)
> (In reply to comment #2)
> > http://hg.mozilla.org/mozilla-central/rev/aa9d33fc9493
> > http://hg.mozilla.org/mozilla-central/rev/623eefcb43d7
> > http://hg.mozilla.org/mozilla-central/rev/795803409745
> None of them can backed out easily for testing purpose. Lots of eg. 'Hunk #6
> FAILED at 765.'.
Maybe you can make builds of rev 4d7bde383df5 and aa9d33fc9493 to compare?

Comment 5

[Hartmut Figge]

(In reply to comment #4)
> Maybe you can make builds of rev 4d7bde383df5 and aa9d33fc9493 to compare?

I was not successful trying it.

Comment 6

[Hartmut Figge]

After sleeping about it i was able to compile close enough builds with my SM 2.1 for testing. What a relief to see the old AOM again. Erm. *g*

No crash with 4d7bde383df5 and crash with aa9d33fc9493. To double check i have backed out aa9d33fc9493 from the latter, which was easy now, and got the crash again.

So it seems clear that the guess of Jens was right and aa9d33fc9493 is responsible for the crash.

Comment 7

[Hartmut Figge]

(In reply to comment #6)

> No crash with 4d7bde383df5 and crash with aa9d33fc9493. To double check i have
> backed out aa9d33fc9493 from the latter, which was easy now, and got the crash
> again.

Nonsense. The crash was gone after the backout of aa9d33fc9493.

Comment 8

[Jens Hatlak (:InvisibleSmiley)]

Confirming as per the findings above. Changeset aa9d33fc9493 was for bug 535643, will add a comment there.

Comment 9

[Robert Kaiser]

This is a core plugins code crash, though SeaMonkey is non-OOPP right now, so that may have an effect there.

Comment 10

[Marcia Knous [:marcia]]

There are several Firefox crashes reported in this stack as well, Windows and Mac: http://tinyurl.com/359s62v.

Comment 11

[NoOp]

One thing I should note - I cannot reproduce in the 64bit builds of 2.1a1 or 2.1a2. I can only reproduce in the 32bit builds.

Comment 12

[NoOp]

I can no longer duplicate. Closing as 'Worksforme'.