Comment 17 for bug 585061

(In reply to comment #12)
> There may be a deeper problem going on ("why is mConnection NULL?", "why are
> these methods being called when mConnection is NULL?"), but for now, just
> sidestepping the NULL-dereferences would be great.
For what it's worth, I think we need to figure out those questions first. Blindly wallpapering over the issue is not the way to go. Clearly some invariants held by the code are being broken here, and it could mean that other things are broken as a result.