711963 - Crash [@ @0x0 | nsSVGTextFrame::GetCanvasTM]

Status: RESOLVED DUPLICATE of bug 704706

(Core :: SVG, defect)

Description

[Christian Holler (:decoder)]

This bug was filed from the Socorro interface and is 
report bp-c700f28a-c89e-44f2-b00e-2264a2111219 .
============================================================= 

The following SVG crashes Firefox Nightly (Build ID 20111218031140):


<svg xmlns="http://www.w3.org/2000/svg">
  <filter id="f_calcMode_linear_to" x="0%" y="0%" width="100%" height="100%">
    <feComponentTransfer>
      <feFuncR type="table" tableValues="0 2 2 0"/><text id="b">Text B</text></feComponentTransfer>
  </filter>
  <filter id="f_calcMode_discrete" x="0%" y="0%" width="100%" height="100%">
    <feComponentTransfer>
    </feComponentTransfer>
  </filter>
</svg>


Assuming sg:critical per severity rating guide, but I'm wondering why executing NULL should be relevant in a remote situation (it is under Windows when conducting local attacks but remote processes usually don't have this address mapped).

Comment 1

[Daniel Veditz [:dveditz]]

If we can prove that it's always EITHER a valid address OR null then it wouldn't be vulnerable, but the fact that our object is unexpectedly gone may mean we can't prove that. This testcase may always result in null, but until we know the cause we don't know if a slightly different one could end up with a different result.

Comment 2

[Daniel Veditz [:dveditz]]

I do not see a crash in Firefox 9 or 10a2.

Comment 3

[Robert Longson [:longsonr]]

Didn't bug 704706 fix this? Are we getting crashes from builds with that patch in? If not this is a duplicate of that.

Comment 4

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to Robert Longson from comment #3)
> Didn't bug 704706 fix this? Are we getting crashes from builds with that
> patch in? If not this is a duplicate of that.

Possibly so, I was testing with Dec 19, 2011 Mac nightly which crashed, and Dec 22, 2011 Mac nightly didn't.

Comment 5

[Robert Longson [:longsonr]]

The code change in that bug is designed to cover this case so if we're not getting crashes post that landing then this is a duplicate.

Comment 6

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

And as per Jesse mentions in bug 704706 comment 0, it may indeed be a regression from the last few days prior to Nov 22, 2011, or possibly bug 696078.

Comment 7

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Duping per comment 5.

Comment 8

[Robert Longson [:longsonr]]

(In reply to Gary Kwong [:gkw, :nth10sd] from comment #6)
> And as per Jesse mentions in bug 704706 comment 0, it may indeed be a
> regression from the last few days prior to Nov 22, 2011, or possibly bug
> 696078.

Jesse was spot on, bug 696078 was the cause.