I’m renaming all instances of `use_transcational_fixtures` to
`use_transactional_tests` and “transactional fixtures” to
“transactional tests”.

I’m deprecating `use_transactional_fixtures=`. So anyone who is
explicitly setting this will get a warning telling them to use
`use_transactional_tests=` instead.

I’m maintaining backwards compatibility—both forms will work.
`use_transactional_tests` will check to see if
`use_transactional_fixtures` is set and use that, otherwise it will use
itself. But because `use_transactional_tests` is a class attribute
(created with `class_attribute`) this requires a little bit of hoop
jumping. The writer method that `class_attribute` generates defines a
new reader method that return the value being set. Which means we can’t
set the default of `true` using `use_transactional_tests=` as was done
previously because that won’t take into account anyone using
`use_transactional_fixtures`. Instead I defined the reader method
manually and it checks `use_transactional_fixtures`. If it was set then
it should be used, otherwise it should return the default, which is
`true`. If someone uses `use_transactional_tests=` then it will
overwrite the backwards-compatible method with whatever they set.

Versions of the pg gem earlier than 0.18.0 cannot be used safely with Ruby 2.2.
Specifically, pg 0.17 when used with Ruby 2.2 has a known bug that causes
random bits to be added to the end of strings. Further explanation here:
https://bitbucket.org/ged/ruby-pg/issue/210/crazy-bytes-being-added-to-record

Various behaviors needed by associations (such as creating the through
record) are lost when `where` is called, since we stop having a
`CollectionProxy` and start having an `AssociationRelation` which does
not contain this behavior. I *think* we should be able to rm
`AssociationRelation`, but we have tests saying the changes required to
do that would be bad (Without saying why. Of course. >_>)

Fixes #19073.

MySQL unicode support is not only `utf8mb4`.
Then, The index length problem is not only `utf8mb4`.

http://dev.mysql.com/doc/refman/5.6/en/charset-unicode.html

    SELECT * FROM information_schema.character_sets WHERE maxlen > 3;
    +--------------------+----------------------+------------------+--------+
    | CHARACTER_SET_NAME | DEFAULT_COLLATE_NAME | DESCRIPTION      | MAXLEN |
    +--------------------+----------------------+------------------+--------+
    | utf8mb4            | utf8mb4_general_ci   | UTF-8 Unicode    |      4 |
    | utf16              | utf16_general_ci     | UTF-16 Unicode   |      4 |
    | utf16le            | utf16le_general_ci   | UTF-16LE Unicode |      4 |
    | utf32              | utf32_general_ci     | UTF-32 Unicode   |      4 |
    +--------------------+----------------------+------------------+--------+

[Toby Ovod-Everett & Andrey Nering & Yves Senn]

Closes #17726.
Closes #10939.

This patch makes three distinct modifications:

1. no longer fall back to disabling user triggers if system triggers can't be disabled
2. warn the user when referential integrity can't be disabled
3. restore aborted transactions when referential integrity can't be disabled

The motivation behind these changes is to make the behavior of Rails
transparent and less error-prone. To require superuser privileges is not optimal
but it's what Rails currently needs. Users who absolutely rely on disabling user triggers
can patch `disable_referential_integrity`.

We should investigate `SET CONSTRAINTS` as a possible solution which does not require
superuser privileges.

/cc @matthewd

Example:

    create_table :foos, id: :primary_key, limit: 8 do |t|
    end

    # or

    create_table :foos, id: false do |t|
      t.column :id, limit: 8
    end

Deprecate `required` option in favor of `optional` for belongs_to.

It is also necessary to format a time column like a datetime column.

[ci skip]

Fixes #18905. `#touch` now takes time as an option. Setting the option
saves the record with the updated_at/on attributes set to the current time
or the time specified. Updated tests and documentation accordingly.

This reverts commit 1502caef.

The test suite for the mysql adapter broke when this commit was used
with MySQL 5.6.

Conflicts:
	activerecord/CHANGELOG.md

Deprecated passing of `start` value to `find_in_batches` and `find_each` in favour of `begin_at` value.

`has_secure_token` hasen't been released yet. No need to track
every change in the CHANGELOG.

```
user = User.create(token: "custom-secure-token")
user.token # => "custom-secure-token"
```

Use SQL COUNT and LIMIT 1 queries for none? and one? methods if no block or limit is given,
instead of loading the entire collection to memory. The any? and many? methods already
follow this behavior.

[Eugene Gilburg & Rafael Mendonça França]

In addition to cleaning up the implementation, this allows type casting
behavior to be applied consistently everywhere. (#where for example). A
good example of this was the previous need for handling value to key
conversion in the setter, because the number had to be passed to `where`
directly. This is no longer required, since we can just pass the string
along to where. (It's left around for backwards compat)

Fixes #18387

It looks like the only reason `current_scope` was thread local on
`base_class` instead of `self` is to ensure that when we call a named
scope created with a proc on the parent class, it correctly uses the
default scope of the subclass. The reason this wasn't happening was
because the proc captured `self` as the parent class, and we're not
actually defining a real method. Using `instance_exec` fixes the
problem.

Fixes #18806

MySQL rejects to remove an index which is used in a foreign key constraint:

```
ActiveRecord::StatementInvalid: Mysql2::Error: Cannot drop index 'index_copies_on_title_id': needed in a foreign key constraint: ALTER TABLE `copies` DROP `title_id`
```

Removing the constraint before removing the column (and the index) solves this problem.

that complements the `start`parameter to specify where to stop batch processing

If timestamp column have the precision, it need to format according to
the precision of timestamp column.

The charset of `version` column in `schema_migrations` table is depend
on the database default charset and collation rather than the encoding
of the connection.

nil or false should not be valid argument to the merge method.

Closes #12264

While we query the proper columns, we go through normal handling for
converting the value to a primitive which assumes it should use the
table's primary key. If the association specifies a different value (and
we know that we're working with an association), we should use the
custom primary key instead.

Fixes #18813.

When ```becomes``` changes @attributes it should also change
@changed_attributes. Otherwise we'll experience a kind of split head situation
where attributes are coming from ```self```, but changed_attributes is coming
from ```klass.new```. This affects the inheritance_colmn as it's changed by new
for example.

Fixes #16881

The cache name should be converted to a string when given, not compared
as a symbol. This edge case is already adequately covered by our tests,
but was masked by another issue where we were incorrectly updating the
counter cache twice. When paired with a bug where we didn't update the
counter cache because we couldn't find a match with the name, this made
it look like everything was working fine.

Fixes #10865.

Fixes #17621. This 5 year old (or older) issue causes validations to fire
when a parent record has `validate: false` option and a child record is
saved. It's not the responsibility of the model to validate an
associated object unless the object was created or modified by the
parent.

Clean up tests related to validations

`assert_nothing_raised` is not benefiting us in these tests
Corrected spelling of "respects"
It's better to use `assert_not_operator` over `assert !r.valid`

The new association error belongs to Active Record, not Active Model.
See #18700 for reference.

    Post.where('id = 1').or(Post.where('id = 2'))
    # => SELECT * FROM posts WHERE (id = 1) OR (id = 2)

[Matthew Draper & Gael Muller]

These callbacks will already have been defined when the association was
built. The check against `reflection.autosave` happens at call time, not
at define time, so simply modifying the reflection is sufficient.

Fixes #18704