[Bug 23684] Gabble advertizes an avatar's sha1 in its presence stanza without following XEP-0153

[bugzilla-daemon at freedesktop.org]

http://bugs.freedesktop.org/show_bug.cgi?id=23684





--- Comment #12 from Will Thompson <will.thompson at collabora.co.uk>  2009-10-08 08:30:55 PST ---
+    sync_stream(q, stream) # Twice because the vCard request is done in
+    sync_stream(q, stream) # g_idle_add

Does this actually reliably give the idle a chance to run?

6f705c81fc7cec means that we leak presence->avatar_sha1 when doing
conflict resolution: self_avatar_resolve_conflict() sets it to NULL
without freeing it. Either it should free it, or not set it to NULL.

+          DEBUG ("Avatar conflict! Received hash '%s' and our cache is '%s'",
+            sha1, presence->avatar_sha1 == NULL ? "<NULL>" :
presence->avatar_sha1);

Style: The second line is too long

+    # if the server don't reply after the timeout and there is pending
+    # requests, Gabble must handle that correctly and not crash.
+    contacts = ['random_user_%s at bigserver.com' % i for i in range(1, 100) ]
+    handles = [conn.RequestHandles(1, [contact])[0] for contact in contacts]
+    conn.Avatars.RequestAvatars(handles)
+    try:
+        conn.Avatars.RequestAvatar(handles[-1])
+    except dbus.DBusException, e:
+        assertEquals(cs.NOT_AVAILABLE, e.get_dbus_name())
+    else:
+        assert False

I know how this works, but we should document what this is actually
doing. Something like:

  # First, ensure the pipeline is full...
  ...
  # Then, request yet another avatar. The request will time out before
  # the IQ is sent, which used to trigger a crash in Gabble
  # (LP#445847). So, we assert that the error is NotAvailable (rather
  # than the error returned when the service crashes).


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.