This will solve the issue that abort the connection transaction when we
skip the tests.

This implements the support to encode/decode JSON
data to/from database and creating columns of type
JSON using a native type [1] supported by PostgreSQL
from version 9.2.

[1] http://www.postgresql.org/docs/9.2/static/datatype-json.html

The previous implementation had the strange requirement that db/structure.sql contain only CREATE TABLE sql statements, one per table, separated by double newlines. SQLite3 and PostgreSQL database tasks, on the other hand, simply spawn 'sqlite3' and 'psql' binaries to load the file directly. The new implementation follows this and attempts to respect all current MySQL configuration settings.

The string_to_dummy_time method was blindly parsing the dummy time string
with Date._parse which returns a hash for the date part regardless
of whether the time part is an invalid time string.

All tests with a custom inheritance_column use the `Vegtable` model.
The field ruby_type on the Company models is no longer needed

previously the tests with and without a custom `inheritance_column`
used the same models. Since the model then has both fields this can lead
to false positives.

I had to create a new table because I needed an STI table,
which does not have both a "type" and a "custom_type"

the test fails with:
  1) Error:
test_alt_becomes_works_with_sti(InheritanceTest):
NoMethodError: undefined method `type=' for #<Cabbage id: 1, name: "my cucumber", custom_type: "Cucumber">
    /Users/username/Projects/rails/activemodel/lib/active_model/attribute_methods.rb:432:in `method_missing'
    /Users/username/Projects/rails/activerecord/lib/active_record/attribute_methods.rb:100:in `method_missing'
    /Users/username/Projects/rails/activerecord/lib/active_record/persistence.rb:165:in `becomes'
    test/cases/inheritance_test.rb:134:in `test_becomes_works_with_sti'
    test/cases/inheritance_test.rb:140:in `test_alt_becomes_works_with_sti'

This is a performance/GC optimisation.

In theory, this could be optimised by the implementation (last time I
checked, this would have no effect on JRuby). But in practise, this make
attribute access faster.

* Loop rather than recurse in retrieve_connection_pool
* Key the hash by class rather than class name. This avoids creating
  unnecessary strings.

This test https://github.com/rails/rails/blob/master/activerecord/test/cases/base_test.rb#L381
is failing because is not setting `CreditCard.pluralize_table_name` to `false`.

In this case, i prefer to change to another model that is not in the `GUESSED_CLASSES`
array.

Ensure class variable is set to nil.

It prevents the following test to fail:

    def test_do_not_run_the_converter_when_nil_was_set
      customers(:david).non_blank_gps_location = nil
      assert_nil Customer.gps_conversion_was_run
    end

Check https://github.com/rails/rails/blob/master/activerecord/test/models/customer.rb#L7
for more information.

serialized attribute value is not from the same class

This reverts commit a7f4b0a1.

Conflicts:
	activerecord/lib/active_record/associations/has_one_association.rb
	activerecord/lib/active_record/persistence.rb
	activerecord/test/cases/base_test.rb
	activerecord/test/cases/dirty_test.rb
	activerecord/test/cases/timestamp_test.rb

This functionality will be available from gem
`active_record-session_store` instead.

This is a cleaner version of #6916.

Closes #3165.

reset_counters() was crashing when there were multiple belongs_to associations with the same foreign key.

This closes #5200.

Reason since MySQL 5.6.6-m9 the `sql_mode` default value is
`NO_ENGINE_SUBSTITUTION`.
This default parameter change is out of control from Rails.

This test verifies Rails not overriding the default `@@GLOBAL.sql_mode` value
by checking if `@@GLOBAL.sql_mode` is the same as `@@SESSION.sql_mode`.

This is a real fix (as compared to the band-aid in b127d86c), which uses
the recently-added equality methods for ARel nodes. It has the side
benefit of simplifying the merge code a bit.

Fixes #7374