Also fix some wrong formatting.

Related discussion:
https://github.com/rails/rails/commit/ab72040b74f742b6676b2d2a5dd029bfdca25a7a#commitcomment-1525256

explains why the file update checker ignores mtimes in the future, plus a little refactor for the same price

Make file name and class name consistant

Related to dbc43bcc

Calculate the seconds since the UNIX epoch using the difference in
Julian day numbers from the epoch date. By reducing the Rational math
to just the offset component this gives a significant improvement.

Benchmark:

Calculating --------------------------------------------
       new     27733 i/100ms
   current     15031 i/100ms
       new     27737 i/100ms
   current     15549 i/100ms
--------------------------------------------------------
       new   548182.1 (±0.9%) i/s - 2745567 in 5.008943s
   current   216380.9 (±1.6%) i/s - 1082232 in 5.002781s
       new   510281.9 (±1.2%) i/s - 2551804 in 5.001525s
   current   219858.3 (±1.8%) i/s - 1103979 in 5.023039s

Introduced in pull request #6910, merged in 2ee3fa1a

connection.tables returns an array of strings

There appears to be a bug with DateTime#strftime("%s") on 32-bit platforms.
Bug report: http://bugs.ruby-lang.org/issues/6683

This reverts commit 210cd756.

Refactor migration test_helper

and create_join_table_test

Persist glob when replacing a path

Use Time.new to create times where the current offset is not zero or
not in the local time zone - closes #4847 and #6651.

Time#at no longer raises an error for large values so we can remove
the rescue clause from ActiveSupport::TimeZone#to_f.

Currently if the offset is not zero then to_time returns self which
can lead to errors where a developer assumes that the value is a
Time. To solve this we can use the native implementation of
DateTime#to_time in Ruby 1.9.3 as it handles offsets properly and
is faster than our override.

The native implementation of the seconds since the UNIX epoch in
strftime is significantly faster than our method.

Benchmark:
----------
require 'benchmark/ips'
require 'date'
require 'time'

date = DateTime.civil(1253,7,6,20,4,0)

Benchmark.ips do |x|
  x.report("strftime.to_i") { date.strftime('%s').to_i }
  x.report("ssue.to_i") { ((date - DateTime.civil(1970)) * 86400).to_i }

  x.report("strftime.to_f") { date.strftime('%s').to_f }
  x.report("ssue.to_f") { ((date - DateTime.civil(1970)) * 86400).to_f }
end

Output:
-------
Calculating -------------------------------------
       strftime.to_i     26480 i/100ms
           ssue.to_i     13818 i/100ms
       strftime.to_f     26561 i/100ms
           ssue.to_f     14479 i/100ms
-------------------------------------------------
       strftime.to_i   616937.3 (±2.4%) i/s - 3098160 in 5.024749s
           ssue.to_i   200108.8 (±6.9%) i/s -  994896 in 4.999278s
       strftime.to_f   553581.3 (±2.2%) i/s - 2788905 in 5.040397s
           ssue.to_f   204260.3 (±4.3%) i/s - 1028009 in 5.043072s

When Rails::Paths::Root's []= is used to replace a path it should persist the previous path's glob. Without passing the glob along we get gnarly bugs when trying to wire up things like engines.

    module FooEngine
      class Engine < ::Rails::Engine
        isolate_namespace FooEngine

        config.paths['config/initializers'] = "lib/foo_engine/initializers"
      end
    end

    ## Example of behaviour before this commit.
    #
    # Before the initializer override:
    >> FooEngine::Engine.config.paths["config/initializers"].glob
    => "**/*.rb"

    # After the initializer override:
    >> FooEngine::Engine.config.paths["config/initializers"].glob
    => nil

    ## Example of behaviour after this commit.
    #
    # Before the initializer override:
    >> FooEngine::Engine.config.paths["config/initializers"].glob
    => "**/*.rb"

    # After the initializer override:
    >> FooEngine::Engine.config.paths["config/initializers"].glob
    => "**/*.rb"

Remove Active Model dependency from Action Pack

Add :default and :null options to column_exists? method

Examples:

    column_exists?(:testings, :taggable_id, :integer, null: false)
    column_exists?(:testings, :taggable_type, :string, default: 'Photo')

Also fix failures in check options for nil

Conflicts:
	activemodel/lib/active_model/errors.rb

Av prepare2

Extract collation tasks to database tasks.

ActiveModel is used in ActionPack for ActiveModel::Naming for a few,
mostly optional aspects of ActionPack related to automatically converting
an ActiveModel compliant object into a key for params and routing. It uses
only three methods of ActiveModel (ActiveModel::Naming.route_key,
ActiveModel::Naming.singular_route_key and ActiveModel::Naming.param_key).

Remove ActiveRelation#inspect

make sure the inflection rules are loaded when cherry-picking active_support/core_ext/string/inflections.rb [fixes #6884]

Introduced in 1abe3167

The test was failing when running on isolation, because the extensions were
not being loaded, thus 1.year.from_now was failing. Just use mktime
instead, adding 1 year to Time.now.

Related with 5e7d6bba

Related with 5e7d6bba

Bug with FileUpdateChecker with wrong mtime

This reverts commit 6acebb38.

Usage of this feature did not reveal any improvement in existing apps.

Conflicts:

	actionpack/lib/action_dispatch/routing/mapper.rb
	guides/source/routing.textile
	railties/lib/rails/engine.rb
	railties/lib/rails/paths.rb
	railties/test/paths_test.rb

Since plugins were removed, we can clean up a few methods in engines.
We also use this opportunity to move `load_console`, `load_tasks` and
`load_runner` to Rails::Engine. This means that, if someone wants to
improve script/rails for engines to support console or runner commands,
part of the work is already done.