424188 - [FIX]Possible to exploit relative xul:script URIs in signed jars

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

See bug 418996 comment 1 and bug 418996 comment 21.

Comment 1

[Boris Zbarsky [:bzbarsky]]

The problem is presumably that XUL doesn't use the scriptloader for <xul:script> and hence doesn't do the downgrading that the scriptloader does?

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: Fix

Comment 4

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Comment on attachment 310841 [details] [diff] [review]
Fix

Looks good

Comment 5

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 310841 [details] [diff] [review]
Fix

Extend to XUL the protection HTML already had.  Only affects non-chrome XUL served inside a signed jar.  Such XUL can no longer keep its signed status if it includes unsigned scripts.

Might be worth beta exposure.

Comment 6

[Damon Sicore (:damons)]

Can we get a test for this?

Comment 7

[Boris Zbarsky [:bzbarsky]]

I'm not going to have time to write one in time for beta...  We need some tests for bug 418996 too, and to test this we need to either copy the server-side stuff Collin set up or (better) come up with some custom signed jars that mochitests can use...

Comment 8

[Damon Sicore (:damons)]

Comment on attachment 310841 [details] [diff] [review]
Fix

Can I get a promise that we'll get a test case for this and bug 418996?  :)

a1.9+ & a1.9beta5+=damons

Comment 9

[Boris Zbarsky [:bzbarsky]]

> Can I get a promise that we'll get a test case for this and bug 418996?  :)

Absolutely.  It's on my short-list of bugs to write tests for as soon as I have the time.  I'm just not sure that will be before 1.9 ship...

If someone picks this up in the meantime, great.  If not, once I finish this whole dissertation thing, I'll just do it.

Comment 10

[Boris Zbarsky [:bzbarsky]]

Filed bug 424488 on having a decent way to test this in a good controlled manner.

Comment 11

[Boris Zbarsky [:bzbarsky]]

Checked in.  Marking fixed in the sense that XUL and HTML now behave the same, though Collin found bug 424426, which affects both for now.

Comment 12

[Johnny Stenback (:jst)]

The branch patch in bug 424426 fixes this bug.

Comment 13

[Johnny Stenback (:jst)]

Fixed on the branch by the fix for bug 424426.

Comment 14

[Al Billings [:abillings - ex-MoCo]]

(In reply to comment #0)
> See bug 418996 comment 1 and bug 418996 comment 21.
> 

I tested the fix with the linked test case in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.15pre) Gecko/2008061005 BonEcho/2.0.0.15pre and the case doesn't repro like it does for 2.0.0.14. Is there additional testing that we should do to verify this?