This reverts commit 28977f1f.

If select clause is specified and last column has a column alias,
additional column alias causes a statement invalid.

Add test coverage for counting a single column with NULL values.

Fixes #27676, #27682, and #27705.

Follow up to #26266.

The default type of `primary_key` and `references` were changed to
`bigint` since #26266. But legacy migration and sqlite3 adapter should
keep its previous behavior.

`ColumnDumper#migration_keys` was extracted to customize keys for
standardized column arguments widths. But the feature was removed in
df84e986. The internal method is no longer used for that.

Since text default treated as an empty string in non-strict mode has
been removed in #26154, `MySQL::Column` behaves like any other column.
Also, The difference between strict and non-strict mode is covered by
`test_mysql_not_null_defaults_strict` and `test_mysql_not_null_defaults_non_strict`.

```
go get -u github.com/client9/misspell/cmd/misspell
misspell  -w -error -source=text .
```

Some custom primary key tests (added at #17631, #17696, #18220, #18228)
were lost at #26266. Restore the tests to improve test coverage.

The PR #27384 changed migration compatibility behaviour.

```ruby
class CreateMasterData < ActiveRecord::Migration[5.0]
  def change
    create_table :master_data, id: :integer do |t|
      t.string :name
    end
  end
end

```

Previously this migration created non-autoincremental primary key
expected. But after the PR, the primary key changed to autoincremental,
it is unexpected.

This change restores the behaviour of the compatibility layer.

And raise `ArgumentError` when passing string to define callback.

The PR #27384 changed integer-like primary key to be autoincrement
unless an explicit default. This means that integer-like primary key is
restored as autoincrement unless dumping the default nil explicitly.
We should dump integer-like primary key with default nil correctly.

When the association's primary key is manually set as a symbol and called with an array of strings, CollectionAssociation#ids_writer
fails to cast the ids to integers. This is because
AssociationReflection#association_primary_key_type and
ThroughReflection#association_primary_key_type return the incorrect
type, since ModelSchema.type_for_attribute only accepts a string. The
result is an ActiveRecord::RecordNotFound error.

MySQL generated columns: https://dev.mysql.com/doc/refman/5.7/en/create-table-generated-columns.html
MariaDB virtual columns: https://mariadb.com/kb/en/mariadb/virtual-computed-columns/

Declare virtual columns with `t.virtual name, type: …, as: "expression"`.
Pass `stored: true` to persist the generated value (false by default).

Example:

    create_table :generated_columns do |t|
      t.string  :name
      t.virtual :upper_name, type: :string,  as: "UPPER(name)"
      t.virtual :name_length, type: :integer, as: "LENGTH(name)", stored: true
      t.index :name_length  # May be indexed, too!
    end

Closes #22589

Fixes #27666.

Fixes #27826
Refer https://github.com/rails/rails/commit/f7b317175430a2d9300d9c4acfc1f34f4fdb2fbc

This is an alternative of #27714.

If `has_one :through` association has set `nil`, `through_record` is
destroyed but still remain loaded target in `through_proxy` until
`reload` or `reset` explicitly.

If `through_proxy` is not reset (remain destroyed (frozen) target),
setting new record causes `RuntimeError: Can't modify frozen hash`.

To prevent `RuntimeError`, should reload `through_record` that has been
destroyed in `create_through_record`.

This message could be generated by `assert_predicate`, https://github.com/rails/rails/commit/1853a4f2c8b2bddfbde6aae80abb08310295201c#commitcomment-20546113
but I'd rather handwrite the message string if I could reduce magic from the code by doing so.

`assert_raise` does not check error message. However, in some tests,
it seems like expecting error message checking with `assert_raise`.
Instead of specifying an error message in `assert_raise`, modify to use
another assert to check the error message.

Follow up to 249f71a2

These internal initialize methods are no longer used internally.

This reverts commit 39c77eb1, reversing
changes made to 9f6f51be.

These are followups for 307065f9,
but TBH I'm personally not very much confortable with this style.
Maybe we could override assert_equal in our test_helper not to warn?

Remove `initialize_schema_migrations_table` and `initialize_internal_metadata_table` internal public methods

These internal methods accidentally appeared in the doc, and so almost
useless. It is enough to create these internal tables directly, and
indeed do so in several places.

https://github.com/rails/rails/blob/v5.0.1/activerecord/lib/active_record/schema.rb#L55
https://github.com/rails/rails/blob/v5.0.1/activerecord/lib/active_record/railties/databases.rake#L6
https://github.com/rails/rails/blob/v5.0.1/activerecord/lib/active_record/tasks/database_tasks.rb#L230

https://www.sqlite.org/foreignkeys.html

If an adapter does not support foreign key feature, should be noop.

https://github.com/rails/rails/blob/v5.0.0.rc1/activerecord/test/cases/migration/foreign_key_test.rb#L288-L294
https://github.com/rails/rails/blob/v5.0.0.rc1/activerecord/test/cases/migration/references_foreign_key_test.rb#L208-L214

Raise `ActiveRecord::InvalidForeignKey` when a record cannot be inserted
or updated because it references a non-existent record for SQLite3
adapter.

This is a regression test for a fix included in
https://github.com/rails/rails/commit/bad9bfbea6d6af9dc28583e08a49492668087393.

Without that change, this test would fail with:

    ActiveRecord::StatementInvalid: SQLite3::SQLException: near "WHERE": syntax error: UPDATE "topics" SET "replies_count" = COALESCE("replies_count", 0) - 1,  WHERE "topics"."id" = ?

If a process is forked more than once, the pool was grabbing the oldest
spec, not the most recent spec. This wasn't noticed before because most
folks are lilely forking the process only once.

If you're forking the process multiple times however the wrong spec name
will be returned and an incorrect connection will be used for the
process.

This fixes the issue by reversing the list of spec names so we can grab
the most recent spec rather than the oldest spec.

Supports the `touch` option from update_counters.
The default behavior is not to update timestamp columns.

because Struct.new returns a Class, we just can give it a name and use it directly without inheriting from it