209701 - Permission denied to create wrapper for object when wrapping a "function" interface, such as TreeWalker.filter

Status: RESOLVED INACTIVE

(Core :: XPConnect, defect)

Description

[Alex Vincent [:WeirdAl]]

From JavaScript, it should be possible to reach the filter property of a
TreeWalker, per the DOM Level 2 Traversal/Range Recommendation from W3C. 
Currently, it is not, instead returning the above exception.

Testcase coming up.

Comment 1

[Alex Vincent [:WeirdAl]]

Attachment: Testcase

Comment 2

[Alex Vincent [:WeirdAl]]

Attachment: Testcase (revised)

Oops.  A couple small glitches with the testcase.

Comment 3

[Christopher Aillon (sabbatical, not receiving bugmail)]

So, caps is complaining here because it is not getting a classinfo from
xpconnect.  The top of the stack looks like:

#0  nsScriptSecurityManager::CanCreateWrapper(JSContext*, nsID const&,
nsISupports*, nsIClassInfo*, void**) (this=0x8110a00, cx=0x847eac0,
aIID=@0x813c058, aObj=0x86fcbf0, 
    aClassInfo=0x0, aPolicy=0x0) at
/usr/local/src/mozilla/trunk/mozilla/caps/src/nsScriptSecurityManager.cpp:2396
#1  0x408a02e3 in XPCWrappedNative::InitTearOff(XPCCallContext&,
XPCWrappedNativeTearOff*, XPCNativeInterface*, int) (this=0x8799780,
ccx=@0xbfffd260, aTearOff=0x87997a0, 
    aInterface=0x87617c0, needJSObject=0) at
/usr/local/src/mozilla/trunk/mozilla/js/src/xpconnect/src/xpcwrappednative.cpp:1425


NodeFilter does not exist in nsDOMClassInfo at all, so I tried adding it there,
with default flags and the default SH, and came up empty.  I'm not quite sure
why that did not work...

Comment 4

[Christopher Aillon (sabbatical, not receiving bugmail)]

An even more reduced testcase is in the URL

Comment 5

[David Bradley]

Attachment: Adds nsISecurityCheckedComponent to nsXPCWrappedjS

One potential solution, still evaluating the security implications

Comment 6

[Alex Vincent [:WeirdAl]]

*** Bug 300668 has been marked as a duplicate of this bug. ***

Comment 7

[Alex Vincent [:WeirdAl]]

*** Bug 302705 has been marked as a duplicate of this bug. ***

Comment 8

[Alex Vincent [:WeirdAl]]

*** Bug 310994 has been marked as a duplicate of this bug. ***

Comment 9

[Alex Vincent [:WeirdAl]]

*** Bug 228767 has been marked as a duplicate of this bug. ***

Comment 10

[Alex Vincent [:WeirdAl]]

*** Bug 307088 has been marked as a duplicate of this bug. ***

Comment 11

[Alex Vincent [:WeirdAl]]

The patch included here is over two years old, but surprisingly enough, has not
bitrotted!

What can we do to fix this?

I'm adding the blocking of bug 312630 and the assertion keyword, because this
bug blocks the assertions I add in that bug.

Comment 12

[Alex Vincent [:WeirdAl]]

I'm going to replace the assertion I would've added for bug 312630 with a
warning.  There's just too many instances to justify crashing the program.

Comment 13

[Martijn Wargers (dead)]

Well, the patch needs review, to see if it is any good, I guess.
It could have negative security implications, though, which is mentioned in
comment 5.

Comment 14

[Alex Vincent [:WeirdAl]]

Comment on attachment 127228 [details] [diff] [review]
Adds nsISecurityCheckedComponent to nsXPCWrappedjS

Requesting review to get the patch back on radar.

Comment 15

[timeless]

Comment on attachment 127228 [details] [diff] [review]
Adds nsISecurityCheckedComponent to nsXPCWrappedjS

src/xpcprivate.h
class nsXPCWrappedJS
won't compile ifndef XPC_USE_SECURITY_CHECKED_COMPONENT

Comment 16

[Gérard Talbot]

Hello, Can someone more knowledgeable than me verify/determine if bug 326248 is a duplicate of this bug here. Thanks.

Comment 17

[BMO Automation]

Per policy at https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Inactive_Bugs. If this bug is not an enhancement request or a bug not present in a supported release of Firefox, then it may be reopened.