Update documentation for v3

6 files changed, 28 insertions, 25 deletions

diff --git a/CHANGES b/CHANGES
index 6a81a72..9b7cf7f 100644
--- a/CHANGES
+++ b/CHANGES
@@ -1,6 +1,12 @@
 Statsd Changelog
 ================
 
+Version 3.0
+-----------
+
+- Moved default client instances out of __init__.py. Now find them in
+  the `statsd.defaults.{django,env}` modules.
+
 Version 2.1.2
 -------------
 
diff --git a/docs/conf.py b/docs/conf.py
index d0ec1dc..4e75a4b 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -48,9 +48,9 @@ copyright = u'2014, James Socol'
 # built documents.
 #
 # The short X.Y version.
-version = '2.1'
+version = '3.0'
 # The full version, including alpha/beta/rc tags.
-release = '2.1.2'
+release = '3.0'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.
diff --git a/docs/configure.rst b/docs/configure.rst
index 752898a..bc945e3 100644
--- a/docs/configure.rst
+++ b/docs/configure.rst
@@ -77,39 +77,35 @@ Here are the settings and their defaults::
 
 You can use the default ``StatsClient`` simply::
 
-    from statsd import statsd
+    from statsd.defaults.django import statsd
 
     statsd.incr('foo')
 
-This instance will use the settings, if provided by Django. If no Django
-settings can be imported, it won't be available.
-
 
 From the Environment
 ====================
 
 Statsd isn't only useful in Django or on the web. A default instance
-will also be available if you configure at least two environment
-variables. These do not have defaults.
+can also be configured via environment variables.
 
-You can set these variables in the environment::
+Here are the environment variables and their defaults::
 
-    STATSD_HOST
-    STATSD_PORT
-    STATSD_PREFIX
-    STATSD_MAXUDPSIZE
+    STATSD_HOST=localhost
+    STATSD_PORT=8125
+    STATSD_PREFIX=None
+    STATSD_MAXUDPSIZE=512
 
 and then in your Python application, you can simply do::
 
-    from statsd import statsd
+    from statsd.defaults.env import statsd
 
     statsd.incr('foo')
 
 .. note::
 
-    To make this default instance available, you will need to set at
-    least ``STATSD_HOST`` and ``STATSD_PORT``, even if using the default
-    values of ``localhost`` and ``8125``.
+    As of version 3.0, this default instance is always available,
+    configured with the default values, unless overridden by the
+    environment.
 
 .. _statsd: https://github.com/etsy/statsd
 .. _Django: https://www.djangoproject.com/
diff --git a/docs/contributing.rst b/docs/contributing.rst
index 08cf6e5..f700d5b 100644
--- a/docs/contributing.rst
+++ b/docs/contributing.rst
@@ -36,8 +36,8 @@ You can also run the tests with tox::
 
     $ tox
 
-Tox will run the tests in Pythons 2.5, 2.6, 2.7, 3.2, 3.3, and PyPy, if
-they're available.
+Tox will run the tests in Pythons 2.5, 2.6, 2.7, 3.2, 3.3, 3.4, and
+PyPy, if they're available.
 
 
 Writing Tests
diff --git a/docs/reference.rst b/docs/reference.rst
index 651944e..0c0202e 100644
--- a/docs/reference.rst
+++ b/docs/reference.rst
@@ -277,3 +277,4 @@ stats.
 
 .. _statsd: https://github.com/etsy/statsd
 .. _0ed78be: https://github.com/etsy/statsd/commit/0ed78be7
+.. _1c10cfc0ac: https://github.com/etsy/statsd/commit/1c10cfc0ac
diff --git a/docs/timing.rst b/docs/timing.rst
index 051f42e..ed85481 100644
--- a/docs/timing.rst
+++ b/docs/timing.rst
@@ -96,7 +96,7 @@ better than nested.)
     # Do something fun.
     foo_timer.stop()
 
-When :py:meth:`statsd.client.Timer.stop` is called, a `timing stat
+When :py:meth:`statsd.client.Timer.stop` is called, a :ref:`timing stat
 <timer-type>`_ will automatically be sent to StatsD. You can over ride
 this behavior with the ``send=False`` keyword argument to ``stop()``::
 
@@ -109,11 +109,11 @@ ready.
 .. _timer-direct-note:
 
 .. note::
-   This use of timers is compatible with `Pipelines <pipeline-chapter>`_
-   but be careful with the ``send()`` method. It *must* be called for
-   the stat to be included when the Pipeline finally sends data, but
-   ``send()`` will *not* immediately cause data to be sent in the
-   context of a Pipeline. For example::
+   This use of timers is compatible with :ref:`Pipelines
+   <pipeline-chapter>`_ but be careful with the ``send()`` method. It
+   *must* be called for the stat to be included when the Pipeline
+   finally sends data, but ``send()`` will *not* immediately cause data
+   to be sent in the context of a Pipeline. For example::
 
     with statsd.pipeline() as pipe:
         foo_timer = pipe.timer('foo').start()