Comment 61 for bug 552227

I think I've discovered where the problem that's stopping facebook is:

in /usr/lib/python-2.6/gwibber/lib/gtk/facebook.py on lines 100 and 101 are the following:

url = web.get_main_frame().get_uri()
data = json.loads(urlparse.parse_qs(url.split("?", 1)[1])["session"][0])

Line 100 sets url to
http://www.facebook.com/connect/login_success.html
instead of
http://www.facebook.com/connect/login_success.html?session={"session_key":"2.3O...(rest of session key etc not shown)
which is what I get when I manually run the URL through a browser, which can be shown by inserting a "print url"

From what I can tell line 101 fails because split returns only 1 item, and the second is dereferenced - this trips the try, so the rest of the block is not executed, which includes the setting of username.

It's quite possible that this is a bug in an upstream package, as I can't see any reference to webkit (or wrappers therof) not returning the query string or having a method to get it

Hope this is helpful to a developer somewhere! - I'm continuing to look into it...