Merge lp:~ken-vandine/gwibber/lp_1058672 into lp:gwibber/3.4

[Ken VanDine]

Use webkit's onload-event instead of title changed to detect successful facebook login. Facebook changed the login_success.html file from html to plain text which does not have a title set. (LP: #1058672)

[Michael Terry (mterry)]

Makes sense. The if "/login_success" in url bit makes me slightly nervous about false positives. It seems unlikely that /login_success would appear randomly, but... I dunno. /login_success?success=false maybe?

You know facebook's login better than me. If that's a bogus concern, go ahead. Else, approved with that fixed if it is a real concern.

review: Approve

Unmerged revisions

1321. By Ken VanDine on 2012-10-18

Use webkit's onload-event instead of title changed to
detect successful facebook login. Facebook changed
the login_success.html file from html to plain text
which does not have a title set. (LP: #1058672)

=== modified file 'gwibber/microblog/plugins/facebook/gtk/facebook/__init__.py'
--- gwibber/microblog/plugins/facebook/gtk/facebook/__init__.py	2012-06-07 14:03:13 +0000
+++ gwibber/microblog/plugins/facebook/gtk/facebook/__init__.py	2012-10-18 15:00:31 +0000
@@ -105,7 +105,7 @@
       "redirect_uri": "http://www.facebook.com/connect/login_success.html",
     })
     web.load_uri("https://graph.facebook.com/oauth/authorize?" + url)
-    web.connect("title-changed", self.on_facebook_auth_title_change)
+    web.connect("onload-event", self.on_facebook_auth_title_change)
 
     self.scroll = Gtk.ScrolledWindow()
     
@@ -126,7 +126,7 @@
     self.dialog.infobar_content_area.show()
 
     url = web.get_main_frame().get_uri()
-    if title.get_title() == "Success":
+    if "/login_success" in url:
       try:
         self.account["access_token"] = str(urlparse.parse_qs(url.split("#", 1)[1])["access_token"][0])
       except: