Use whatever adapter-provided means we have available to ensure forked
children don't send quit/shutdown/goodbye messages to the server on
connections that belonged to their parent.

Add new error class `StatementTimeout` which will be raised when statement timeout exceeded (#31129)

We are sometimes using The MAX_EXECUTION_TIME hint for MySQL depending
on the situation. It will prevent catastrophic performance down by wrong
performing queries.

The new error class `StatementTimeout` will make to be easier to handle
that case.

https://dev.mysql.com/doc/refman/5.7/en/optimizer-hints.html#optimizer-hints-execution-time

Follow up of #30360.

Attribute modules (`Attribute`, `Attributes`, `AttributeSet`) uses
`Type`, but referencing `Type` before the modules still fail.

```
% ./bin/test -w test/cases/attribute_test.rb -n test_with_value_from_user_validates_the_value
Run options: -n test_with_value_from_user_validates_the_value --seed 31876

E

Error:
ActiveModel::AttributeTest#test_with_value_from_user_validates_the_value:
NameError: uninitialized constant ActiveModel::AttributeTest::Type
    /Users/kamipo/src/github.com/rails/rails/activemodel/test/cases/attribute_test.rb:233:in `block in <class:AttributeTest>'

bin/test test/cases/attribute_test.rb:232

Finished in 0.002985s, 335.0479 runs/s, 335.0479 assertions/s.
1 runs, 1 assertions, 0 failures, 1 errors, 0 skips
```

Probably we need more autoloading at least `Type`.

This fixes an issue where competing threads deadlock each other.

- Thread A holds the load interlock but is blocked on getting the DB lock
- Thread B holds the DB lock but is blocked on getting the load interlock (for example when there is a `Model.transaction` block that needs to autoload)

This solution allows for dependency loading in other threads while a thread is waiting to acquire the DB lock.

Fixes #31019

```
[NoMethodError]: undefined method `state' for nil:NilClass  Method:[rescue in block in refresh]
```

In `within_new_transaction`, there is the possibility that
`begin_transaction` returns a `nil`. (i.e.: so `transaction = nil`)
So this method is checking `transaction` for nil in 2 spots.

Unfortunately, there is one line that is not checking `transaction` for `nil`
That line, `commit_transaction`, throws an exception for us in AR 5.0.0.1

The problem with the method is finally realized in the error checking itself.
it calls `transaction.state` (i.e.: nil.state) and that is the final exception
raised.

The actual underlying (user) issue is hidden by this line.

Solution is test transaction for nil.

Replace the primary key type `integer` in docs with `bigint`.

ref #26266

and abstract `SchemaDumper#extensions` is now an empty method.

Since #30337, every database adapter has its own `SchemaDumper`.
`extensions` are only supported by PostgreSQL database and postgresql database adapter.

since the minimum version of PostgreSQL currently Rails supports is 9.1,
there is no need to handle if `supports_extensions?`

Refer https://www.postgresql.org/docs/9.1/static/sql-createextension.html
"CREATE EXTENSION"

Although the sqlite adapter supports index sort orders, they
weren't being written to db/schema.rb.

`supports_disable_referential_integrity?` used to handle
if PostgreSQL database supports
`ALTER TABLE <table name> DISABLE/ENABLE TRIGGER` statements.
Refer https://github.com/rails/rails/commit/9a947af0e79cfb8692eb7e5ae94c1b8c40756f49

These statements have been documented since 8.1.

https://www.postgresql.org/docs/8.1/static/sql-altertable.html
> DISABLE/ENABLE TRIGGER

Now Rails supports PostgreSQL 9.1 or higher only.
No need to handle `supports_disable_referential_integrity?` anymore.
Also, this method does not exist in any other adapters including AbstractAdapter.

Remove deprecated methods `initialize_schema_migrations_table` and `initialize_internal_metadata_table`

This basically reverts 9d4f79d3

We already found the longer sequence name, but we could not consider
whether it was the sequence name created by serial type due to missed a
max identifier length limitation. I've addressed the sequence name
consideration to respect the max identifier length.

Fixes #28332.

Currently `AUTO_INCREMENT` is implicitly used in the default primary key
definition. But `AUTO_INCREMENT` is not only used for single column
primary key, but also for composite primary key. In that case,
`auto_increment: true` should be dumped explicitly in the
`db/schema.rb`.

Fixes #30894.

s/Action Record/Active Record/

* When the adapter is missing, raise an exception that points out config
  typos and missing Gemfile entries. (We can assume that a non-builtin
  adapter was used since these are always available.)
* When loading an adapter raises a LoadError, prefix its error message
  to indicate that the adapter is likely missing an optional dependency.

Implemented by #22911
Related to #30677

PostgreSQL 9.1+ introduced range types, and Rails added support for
using this datatype in ActiveRecord. However, the serialization of
`PostgreSQL::OID::Range` was incomplete, because it did not properly
quote the bounds that make up the range. A clear example of this is a
`tsrange`.

Normally, ActiveRecord quotes Date/Time objects to include the
milliseconds. However, the way `PostgreSQL::OID::Range` serialized its
bounds, the milliseconds were dropped. This meant that the value was
incomplete and not equal to the submitted value.

An example of normal timestamps vs. a `tsrange`. Note how the bounds
for the range do not include their milliseconds (they were present in
the ruby Range):

    UPDATE "iterations" SET "updated_at" = $1, "range" = $2 WHERE
    "iterations"."id" = $3
    [["updated_at", "2017-09-23 17:07:01.304864"],
    ["range", "[2017-09-23 00:00:00 UTC,2017-09-23 23:59:59 UTC]"],
    ["id", 1234]]

`PostgreSQL::OID::Range` serialized the range by interpolating a
string for the range, which works for most cases, but does not work
for timestamps:

    def serialize(value)
      if value.is_a?(::Range)
        from = type_cast_single_for_database(value.begin)
        to = type_cast_single_for_database(value.end)
        "[#{from},#{to}#{value.exclude_end? ? ')' : ']'}"
      else
        super
      end
    end

    (byebug) from = type_cast_single_for_database(value.begin)
    2010-01-01 13:30:00 UTC

    (byebug) to = type_cast_single_for_database(value.end)
    2011-02-02 19:30:00 UTC

    (byebug) "[#{from},#{to}#{value.exclude_end? ? ')' : ']'}"
    "[2010-01-01 13:30:00 UTC,2011-02-02 19:30:00 UTC)"

@sgrif (the original implementer for Postgres Range support) provided
some feedback about where the quoting should occur:

  Yeah, quoting at all is definitely wrong here. I'm not sure what I
  was thinking in 02579b56, but what this is doing is definitely in the
  wrong place. It should probably just be returning a range of
  subtype.serialize(value.begin) and subtype.serialize(value.end), and
  letting the adapter handle the rest.

`Postgres::OID::Range` now returns a `Range` object, and
`ActiveRecord::ConnectionAdapters::PostgreSQL::Quoting` can now encode
and quote a `Range`:

    def encode_range(range)
      "[#{type_cast(range.first)},#{type_cast(range.last)}#{range.exclude_end? ? ')' : ']'}"
    end

    ...

    encode_range(range)
    #=> "['2010-01-01 13:30:00.670277','2011-02-02 19:30:00.745125')"

This commit includes tests to make sure the milliseconds are
preserved in `tsrange` and `tstzrange` columns

I got this error in production using Puma in multi-threaded mode:

```
RuntimeError: Digest::Base cannot be directly inherited in Ruby
from active_support/security_utils.rb:23:in `variable_size_secure_compare'
from active_support/security_utils.rb:23:in `hexdigest'
from active_support/security_utils.rb:23:in `digest'
```

Looks like Digest uses const_missing to load Digest::SHA256 (https://github.com/ruby/ruby/blob/trunk/ext/digest/lib/digest.rb#L8)

- https://bugs.ruby-lang.org/issues/9494
- https://github.com/ruby/ruby/commit/c02fa39463a0c6bf698b01bc610135604aca2ff4

Currently the normalization only exists in `primary_key` shorthand. It
should be moved to `new_column_definition` to also affect to
`add_column` with primary key.

`change_column_{default,null,comment}` in mysql2 adapter are passing
`column.sql_type` as `type` to `change_column` to intend keeping
previous type. But `column_for` requires extra query, so use passing
`nil` to `type` explicitly in the internal for the purpose.

If collided named sequence already exists, newly created serial column
will generate alternative sequence name. Fix sequence name detection to
allow the alternative names.

Because `quoted_table_name` doesn't respect table alias. We should use
`arel_attribute` for that, so I added `column_name_from_arel_node` to
generate column name from an arel node.

Fixes #30539.

Otherwise using reserved words as composite primary key names will be
failed as an invalid SQL.

Fixes #30518.