ActiveRecord::Observer was extracted into a separate gem so it
should not be referenced anymore (see https://github.com/rails/rails/commit/ccecab3)

hash filters should be accessed with symbols or strings

At present, ActiveRecord::Delegation compiles delegation methods on a
global basis. The compiled methods apply to all subsequent Relation
instances. This creates several problems:

1) After Post.all.recent has been called, User.all.respond_to?(:recent)
   will be true, even if User.all.recent will actually raise an error due
   to no User.recent method existing. (See #8080.)

2) Depending on the AR class, the delegation should do different things.
   For example, if a Post.zip method exists, then Post.all.zip should call
   it. But this will then result in User.zip being called by a subsequent
   User.all.zip, even if User.zip does not exist, when in fact
   User.all.zip should call User.all.to_a.zip. (There are various
   variants of this problem.)

We are creating these compiled delegations in order to avoid method
missing and to avoid repeating logic on each invocation.

One way of handling these issues is to add additional checks in various
places to ensure we're doing the "right thing". However, this makes the
compiled methods signficantly slower. In which case, there's almost no
point in avoiding method_missing at all. (See #8127 for a proposed
solution which takes this approach.)

This is an alternative approach which involves creating a subclass of
ActiveRecord::Relation for each AR class represented. So, with this
patch, Post.all.class != User.all.class. This means that the delegations
are compiled for and only apply to a single AR class. A compiled method
for Post.all will not be invoked from User.all.

This solves the above issues without incurring significant performance
penalties. It's designed to be relatively seamless, however the downside
is a bit of complexity and potentially confusion for a user who thinks
that Post.all and User.all should be instances of the same class.

Benchmark
---------

require 'active_record'
require 'benchmark/ips'

class Post < ActiveRecord::Base
  establish_connection adapter: 'sqlite3', database: ':memory:'
  connection.create_table :posts

  def self.omg
    :omg
  end
end

relation = Post.all

Benchmark.ips do |r|
  r.report('delegation')   { relation.omg }
  r.report('constructing') { Post.all }
end

Before
------

Calculating -------------------------------------
          delegation      4392 i/100ms
        constructing      4780 i/100ms
-------------------------------------------------
          delegation   144235.9 (±27.7%) i/s -     663192 in   5.038075s
        constructing   182015.5 (±21.2%) i/s -     850840 in   5.005364s

After
-----

Calculating -------------------------------------
          delegation      6677 i/100ms
        constructing      6260 i/100ms
-------------------------------------------------
          delegation   166828.2 (±34.2%) i/s -     754501 in   5.001430s
        constructing   116575.5 (±18.6%) i/s -     563400 in   5.036690s

Comments
--------

Bear in mind that the standard deviations in the above are huge, so we
can't compare the numbers too directly. However, we can conclude that
Relation construction has become a little slower (as we'd expect), but
not by a huge huge amount, and we can still construct a large number of
Relations quite quickly.

Keying these hashes by klass causes reloadable classes to never get
freed. Thanks to @thedarkone for pointing this out in
the comments on 221571be.

This doesn't seem to make a massive difference to performance.

Benchmark
---------

require 'active_record'
require 'benchmark/ips'

class Post < ActiveRecord::Base
  establish_connection adapter: 'sqlite3', database: ':memory:'
end

GC.disable

Benchmark.ips(20) do |r|
  r.report { Post.connection }
end

Before
------

Calculating -------------------------------------
                          5632 i/100ms
-------------------------------------------------
                       218671.0 (±1.9%) i/s -    4364800 in  19.969401s

After
-----

Calculating -------------------------------------
                          8743 i/100ms
-------------------------------------------------
                       206525.9 (±17.8%) i/s -    4039266 in  19.992590s

These are for internal use only and cannot be relied on as part of the
public API. See discussion on 8c2c6051.

Remove outdated monkey patch in mysql tests

This reverts commit a3024f81.

REASON: Since warning doesn't exist

Split Validations and Callbacks into separate guides.

Two big features that are only barely related in the same guide. Seems
bad.

I did not check references to these guides yet, so some links may need
to be updated.

Gemfile source url for plugins updated

Move instantiation responsibilities from Inheritance to Persistence. Have Inheritance#discriminate_class_for_record handle STI lookup duties.

Added examples to current_page? method [ci skip]

augmented the Rails 4 guides with links back to Github

Allows you to do BaseClass.new(:type => "SubClass") as well as
parent.children.build(:type => "SubClass") or parent.build_child
to initialize an STI subclass. Ensures that the class name is a
valid class and that it is in the ancestors of the super class
that the association is expecting.

Add #seconds_until_end_of_day to DateTime and Time

[ci skip]

Fix typo in the Observers deprecation message.

Rather than just raising a NoMethodError when copying the config, this
commit adds a warning message until either the rails-observers gem is
installed or the relevant config options are removed.

They was extracted from a plugin.

See https://github.com/rails/rails-observers

[Rafael Mendonça França + Steve Klabnik]

Simplify String#mb_chars and fix documentation.

This behavior mattered under Ruby 1.8, but that doesn't matter now
that we don't support it.

In addition, we don't want to proxy the #class method. A test was added
to prevent against regressions.

This documentation has been out of date.

Ensure that associations have a symbol argument.

Fixes #7418.

Specify type of singular assication during serialization

When serialising a class, specify the type of any singular associations, if
necessary. Rails already correctly specifies the :type of any enumerable
association (e.g. a has_many association), but made no attempt to do so for
non-enumerables (e.g. a has_one association).
We must specify the :type of any STI association. A has_one
association to a class which uses single-table inheritance is an example of
this type of association.

Fixes #7471

Revert "rails gem has no lib directory to require"

This reverts commit 6bcbd604.

The default value for require_paths is ['lib'], and Rails does not
need to include any lib/ file to build the gem, so the previous
commit 'makes sense'. However, gems will not be built at all if
require_paths is empty (see https://gist.github.com/1065473), so
it's better to stick with the default, which does not have any
side-effects