The testing of error messages have been implemented wrongly a few times.
This is an attempt to fix it.

For example, some of these test should have failed with the new code.
The reason they are not failling with the new string is the fact they
were not being tested beforehand.

These methods are nodoc so we should not document them.

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

[ci skip]

The cause by which the test suite for the mysql adapter broke in 1502caef
(reverted 89ba5bb4) is because the precision was not extracted.

The rounding problem in mysql adapter has not been fixed, but `mysql_56`
helper tested only mysql2 adapter, its behavior was not apparent.

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.

Lowercase raw SQL has been replaced by 07b659ca already. This commit
replaces everything else of raw SQL.

/cc @yahonda

This makes it easier for third party adapters to run our tests,
even if that database does not support IF EXISTS.

This helper no longer makes sense as a separate method. Instead I'll
just have `deserialize` call `cast` by default. This led to a random
infinite loop in the `JSON` pg type, when it called `super` from
`deserialize`. Not really a great way to fix that other than not calling
super, or continuing to have the separate method, which makes the public
API differ from what we say it is.

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.

As per previous discussions, we want to give users the ability to
reference their own types with symbols, instead of having to pass the
object manually. This adds the class that will be used to do so.

    ActiveRecord::Type.register(:money, MyMoneyType)

...so it doesn't look like you *have* to use SQL strings for that case (not
anymore!). Would like to replace the SQL string example with something that
you cannot do with the "normal" query API, but I could not come up with a
short, realistic example. Suggestions welcome!

Using enum names in SQL strings doesn't actually work, the test was wrong (fixed
in 3dfd1bab).

Make sure we have coverage for both the find/build cases

Also updated the documentation about the new ability to query them normally,
and added test to make sure they work!

```
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]

The datetime precision tests for any adapters is duplicated.

This finally removes the warning "WARNING:  there is no transaction in progress"
when running Active Record tests using PostgreSQL.

Also removed some cruft in the `setup` and `teardown` methods.

"active_support/testing/stream" is already required in `test_case.rb`.
Furthermore the test "test/cases/migration_test.rb" could no longer be executed
directly.

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

As far as I can tell, the original reason that this behavior was added
has been sufficiently resolved elsewhere, as we no longer remove the
encoding of strings coming out of the database.