916788 - crash on motherjones.com starting with ASSERTION: shouldn't be called with a null inner window: 'pwin', file content/base/src/nsScriptLoader.cpp, line 918

Status: RESOLVED DUPLICATE of bug 915625

(Core :: DOM: Core & HTML, defect)

Description

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

Attachment: end of browser stdout/stderr, followed by gdb debugging of problem

This is a similar assertion sequence as bug 897399, but I have the fix for that in the build where I'm seeing this.

I reproduced the following crash twice with the following steps, in a debug build based on changeset b9029b1de410 plus my patch queue (local sourcestamp is f345c7f9ec70 if I need to figure out which patches), and in a session with enough tabs open to be somewhat bogged down:
 1. load http://www.motherjones.com/politics/2013/08/eric-schlosser-command-control-excerpt-nuclear-weapons?page=2
 2. click the left arrow to go to page 1

###!!! ASSERTION: shouldn't be called with a null inner window: 'pwin', file content/base/src/nsScriptLoader.cpp, line 918

###!!! ASSERTION: windows must be global objects: 'globalObject', file content/base/src/nsScriptLoader.cpp, line 921

###!!! ABORT: You can't dereference a NULL nsCOMPtr with operator->().: 'mRawPtr != 0', file ../../../dist/include/nsCOMPtr.h, line 819

Comment 1

[Boris Zbarsky [:bzbarsky]]

David, do you have the fix for bug 915625 in your tree?  If not, do you have the patch for bug 906371, which caused bug 915625?

Comment 2

[Boris Zbarsky [:bzbarsky]]

Ah, the stack says you have the patch for bug 906371.  So I expect this is a duplicate of bug 915625.

Comment 3

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

(In reply to Boris Zbarsky [:bz] from comment #2)
> Ah, the stack says you have the patch for bug 906371.  So I expect this is a
> duplicate of bug 915625.

See also comment 0:

(In reply to David Baron [:dbaron] (needinfo? me) (busy through Sept. 14) from comment #0)
> I reproduced the following crash twice with the following steps, in a debug
> build based on changeset b9029b1de410 plus my patch queue (local sourcestamp
> is f345c7f9ec70 if I need to figure out which patches), and in a session

$ hg log -r 'ancestor(b9029b1de410, e64327542d6b)' --template="{node|short}\n"
cc0af9e0a324

which says I do not have the patch for bug 915625.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Ah, sorry I missed the changesets in comment 0.  :(