Mention the Graph API Explorer. It is a handy tool for debugging issues
with the wire protocol.

https://bugzilla.gnome.org/show_bug.cgi?id=785726

Ever since version 2.3 of the Graph API was retired on 8th July, it
stopped including "email" in the JSON when a GET request is issued
against https://graph.facebook.com/me?access_token=...

Now it needs to be explicitly asked for. The changelog [1] for version
2.4 says:
  "To try to improve performance on mobile networks, Nodes and Edges
   in v2.4 requires that you explicitly request the field(s) you need
   for your GET requests. For example, GET /v2.4/me/feed no longer
   includes likes and comments by default, but GET
   /v2.4/me/feed?fields=comments,likes will return the data."

It seems that "id" is always included, but we mention it anyway.

The "username" fallback that was added in commit 10c77f17 has
been removed because requesting it leads to:
  {
    "error": {
      "message": "(#12) username field is deprecated for versions v2.0
                  and higher",
      "type": "OAuthException",
      "code": 12,
      "fbtrace_id": "AJYNXiIv00W"
    }
  }

We will have to investigate whether we need a new fallback for the
presentation identity.

[1] https://developers.facebook.com/docs/apps/changelog

https://bugzilla.gnome.org/show_bug.cgi?id=785726

The json_object_get_string_member API expects the member to be present.
Therefore, we should not use its return value to determine its
availability. Otherwise it will lead to:
  Json-CRITICAL **: json_object_get_string_member: assertion
    'node != NULL' failed

https://bugzilla.gnome.org/show_bug.cgi?id=785726

Google recently updated their web authentication UI. The ID and name
of the identity node has changed.

Spotted by Matthias Clasen.

https://bugzilla.gnome.org/show_bug.cgi?id=779548

https://bugzilla.gnome.org/show_bug.cgi?id=779548

It's good to consolidate all the scope level resource deallocation in
one place. Since we have an 'out' label now, let's make use of it.

https://bugzilla.gnome.org/show_bug.cgi?id=779548

https://bugzilla.gnome.org/show_bug.cgi?id=779548

https://bugzilla.gnome.org/show_bug.cgi?id=776897

The lifecycle of operation_result is controlled by the g_bus_watch_name
call and the watched_client_connections hash table. Therefore, we
should not unref it when there is an error in on_caller_watched. What
we should do is drop our reference to it right after starting to watch
the bus.

https://bugzilla.gnome.org/show_bug.cgi?id=776897

https://bugzilla.gnome.org/show_bug.cgi?id=765121

https://bugzilla.gnome.org/show_bug.cgi?id=765121

https://bugzilla.gnome.org/show_bug.cgi?id=765121

https://bugzilla.gnome.org/show_bug.cgi?id=776897

The list returned by goa_client_get_accounts() is (transfer full).
ensure_credentials_for_accounts() was not freeing it.

https://bugzilla.gnome.org/show_bug.cgi?id=776871

The vast majority, if not all, of *_finish functions out there transfer
ownership of the return value to the caller. This also matches with the
behaviour of new APIs like g_task_propagate_pointer. Therefore, let's
make this change for the sake of consistency and making it easier to
port to GTask.

https://bugzilla.gnome.org/show_bug.cgi?id=764157

Some bugs in evolution-data-server were causing us to hit Google's
daily limit for the CalDAV and Tasks APIs. The bugs have since been
fixed, but until every single user out there installs the fix,
everybody using the old keys will be susceptible even if they have a
fixed copy of evolution-data-server. This is because Google identifies
us by the API key and not the version of the code running on the
client.

We are therefore phasing out the old key, so that users updating their
systems have no connection to the one that was tainted by these bugs.

https://bugzilla.gnome.org/show_bug.cgi?id=771547
https://bugzilla.gnome.org/show_bug.cgi?id=773248

https://bugzilla.gnome.org/show_bug.cgi?id=756431

https://bugzilla.gnome.org/show_bug.cgi?id=756431

Fall out from 517fc9c4

https://bugzilla.gnome.org/show_bug.cgi?id=768808

https://bugzilla.gnome.org/show_bug.cgi?id=756431

https://bugzilla.gnome.org/show_bug.cgi?id=768845

https://bugzilla.gnome.org/show_bug.cgi?id=768845