[Webkit-unassigned] [Bug 64802] Custom cursors cause the WebProcess to crash

Wed Jul 20 08:09:54 PDT 2011

https://bugs.webkit.org/show_bug.cgi?id=64802


--- Comment #7 from Amruth Raj <amruthraj at motorola.com>  2011-07-20 08:09:54 PST ---
(In reply to comment #6)
> (In reply to comment #5)
> > Darin, would you prefer this just fixed in WebCore to ensure we don't send null images?  I'd be happy to write that patch, but it wouldn't get fixed today.
> 
> I think the right question is what behavior we want when such a cursor is specified. Once we know what behavior we want, then we can figure out how to get that behavior. We could start by figuring out what behavior we have here in WebKit1. I’m thinking that a custom cursor with an empty image should probably result in an invisible cursor. Is that right?
In WebKit1, GTK handles this case. In case a NULL image is sent to gdk_window_set_cursor, it uses cursor of the parent window to display.
> 
> As far as implementation is concerned, it’s OK with me to fix the encoding machinery so it can successfully send an empty custom cursor across the process boundary. I object to the incorrect patch, not the very notion of what the patch is doing. The patch is wrong because if we encode one of these, the decoder will then try to decode fields that aren’t present and consume additional bytes and thus be unable to decode the rest of the stream. That’s a relatively simple mechanical thing to fix.
I will attach a patch shortly addressing this.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.