585943 - Crash [@ nsHtml5TreeOperation::Init][@ TouchBadMemory][@ nsHtml5TreeBuilder::appendElement(nsIContent**, nsIContent**) ] on OOM

Status: RESOLVED FIXED

(Core :: DOM: HTML Parser, defect, P2)

Description

[Bob Clary [:bc] (inactive)]

Attachment: nsHtml5TreeOperation::Init crash report

1. http://redtaggy.com/
2. crash trunk mac/windows

related socorro signature:
nsHtml5TreeBuilder::appendElement(nsIContent**, nsIContent**) 

various assertions depending on the phase of the moon.

ASSERTION: Tree op allocation failed.: 'treeOp' /work/mozilla/builds/2.0.0/mozilla/parser/html/nsHtml5TreeBuilderCppSupplement.h

ASSERTION: bad method name: 'Error' /work/mozilla/builds/2.0.0/mozilla/js/src/xpconnect/src/xpcwrappednativeinfo.cpp

Comment 1

[Bob Clary [:bc] (inactive)]

Attachment: TouchBadMemory crash report

Comment 2

[Johnny Stenback (:jst)]

We should fix this for final, and by fix I mean make sure we crash in a safe way if we're OOM.

Comment 3

[Henri Sivonen (:hsivonen)]

All the nsTArrays in the HTML5 parser need to become infallible, but bug 550611 isn't ready yet.

Comment 4

[Johnny Stenback (:jst)]

bug 550611 is fixed now. Henri, can you work on a patch to make all nsTArray's in the parser explicitly use the infallible version?

Comment 5

[Henri Sivonen (:hsivonen)]

(In reply to comment #4)
> bug 550611 is fixed now. Henri, can you work on a patch to make all nsTArray's
> in the parser explicitly use the infallible version?

OK.

Comment 6

[Henri Sivonen (:hsivonen)]

Can this be considered fixed now that bug 610823 is fixed? That is, is it necessary to use the explicitly infallible versions now?

Comment 7

[Johnny Stenback (:jst)]

IMO yes.

Comment 8

[Johnny Stenback (:jst)]

Marking fixed per previous comments.