[Webkit-unassigned] [Bug 177577] REGRESSION(2.18.0): [GTK] Crash in WebKitWebContext disposal

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=177577

--- Comment #4 from Michael Catanzaro <mcatanzaro at igalia.com> ---
Actually, I see from https://bugs.eclipse.org/bugs/show_bug.cgi?id=522733#c7 that you really are creating a normal GTK+ widget hierarchy, so you should not need to dispose the WebKitWebView manually or do anything other than add it to the GtkContainer. After you create the WebKitWebView and add it to the GtkContainer, the GtkContainer has sunk the floating ref and the WebKitWebView is now owned by the GtkContainer.

Note: it's expected that dispose may be called multiple times, as that's required to break reference cycles. What's not OK is if *finalize* gets called multiple times. That should never happen, though, because g_object_unref() should spit obvious criticals out on stderr if it gets called when the refcount is already zero. So I don't think it's a WebKitWebView refcounting error unless you are seeing criticals like these.

Also, my question "does Eclipse use WebCookieManagerProxy directly" was pretty dumb, because that's a C++ object, not something exposed in our API. So forget about that. (I'm working a bit too late at night again. :)

(Sorry I haven't actually found the issue, but hopefully the above gives you something to work with. First thing I would do to investigate further would be 'bt full' in gdb.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20171026/895b8715/attachment.html>