Apart from specific versioning support, our tests should focus on the
behaviour of whatever version they're accompanying, regardless of when
they were written.

Application code should *not* do this.

This reverts commit 6d5b1fdf.

`eager_load` and `references` can include hashes, which won't match up
with `references`

A test case has been added to demonstrate the problem

There was a test remaining for PG only that was checking for an exact
limit clause

We currently generate an unbounded number of prepared statements when
`limit` or `offset` are called with a dynamic argument. This changes
`LIMIT` and `OFFSET` to use bind params, eliminating the problem.

`Type::Value#hash` needed to be implemented, as it turns out we busted
the query cache if the type object used wasn't exactly the same object.

This drops support for passing an `Arel::Nodes::SqlLiteral` to `limit`.
Doing this relied on AR internals, and was never officially supported
usage.

Fixes #22250.

Some backends allow `LIMIT 1,2` as a shorthand for `LIMIT 1 OFFSET 2`.
Supporting this in Active Record massively complicates using bind
parameters for limit and offset, and it's trivially easy to build an
invalid SQL query by also calling `offset` on the same `Relation`.

This is a niche syntax that is only supported by a few adapters, and can
be trivially worked around by calling offset explicitly.

The problem was that when saving an object, we would
call touch_later on the parent which wont be saved immediteally, and
it wont call any callbacks. That was working one level up because
we were calling touch, during the touch_later commit phase. However that still
didnt solve the problem when you have a 3+ levels of parents to be touched,
as calling touch would affect the parent, but it would be too late to run callbacks
on its grand-parent.

The solution for this, is instead, call touch_later upwards when the first
touch_later is called. So we make sure all the timestamps are updated without relying
on callbacks.

This also removed the hard dependency BelongsTo builder had with the TouchLater module.
So we can still have the old behaviour if TouchLater module is not included.

[fixes 5f5e6d92]
[related #19324]

Those are actually shortcuts for `after_commit`.

Before:

    after_commit :add_to_index_later, on: :create
    after_commit :update_in_index_later, on: :update
    after_commit :remove_from_index_later, on: :destroy

After:

    after_create_commit  :add_to_index_later
    after_update_commit  :update_in_index_later
    after_destroy_commit :remove_from_index_later

This removes the following warning which has been out in the case of a PostgreSQL 9.3 below.

```
activerecord/test/cases/adapters/postgresql/geometric_test.rb:265: warning: instance variable @connection not initialized
```

This will also get the defaults from attribute definitions like:

     attribute :type, :string, default: "SomethingElse"

`has_attribute?` method to check wether a given attribute has been defined.

This solves the following error:

     ActiveRecord::StatementInvalid: Could not find table 'guitars'

It seems that the table structure of the `Guitar` model has not been
necessary until now. Due to the wrong table name the model was not
correctly linked to the table.

fixes #17121

It appears that I missed this one when I delegated all the non-mutation
array methods that were not on Enumerable

As was pointed out by #17128, our blacklist of mutation methods was
non-exhaustive (and would need to be kept up to date with each new
version of Ruby). Now that `Relation` includes `Enumerable`, the number
of methods that we actually need to delegate are pretty small. As such,
we can change to explicitly delegating the few non-mutation related
methods that `Array` has which aren't on `Enumerable`

When same association is loaded in the model creation callback
The new object is inserted into association twice

`replace_named_bind_variables` and `replace_bind_variables` are
definded in `sanitization.rb`, so it is reasonable these tests are
on `sanitize_test.rb`.

Follow up to #21601.

If postgresql is being used and there are multiple schemas listed on the
`schema_search_path`, then `structure.sql` dumps (triggered by `rake
db:structure:dump` or `config.active_record.schema_format = :sql`) began
failing in Rails 4.2.5.

This is due to the changes made in
https://github.com/rails/rails/pull/17885 The problem is that multiple
schemas were getting getting passed to `Kernel.system` as a single,
space delimited string argument (for example, "--schema=foo
--schema=bar"). However, with the updated array style of calling
`Kernel.system`, these need to be passed as separate arguments (for
example, "--schema=foo", "--schema=bar"). If they get passed as a single
string, then the underlying pg_dump program isn't sure how to interpret
that single argument and you'll get an error reporting: "pg_dump: No
matching schemas were found"

This reverts commit 8246b593.

There was concern about this modifying the behavior of past migrations.
We're going to add an way to modify the migration generator instead.

It's often the case that you want to have an option that you cannot
specify at the database level, but want applied to *all* tables that you
create. For example, you might want to specify `ROW_FORMAT=DYNAMIC` to
not have to limit text columns to length 171 for indexing when using
utf8mb4. This allows an easy way to specify this in your database
configuration.

While this change affects both MySQL and MySQL2, the test only covers
MySQL2, as the legacy mysql adapter appears to always return ASCII
strings, and is tangential to what we're actually doing.

- key was a poor choice of name. A key implies something that will
  unlock a lock. The concept is actually more like a 'lock identifier'
- mysql documentation calls this a 'lock name'
- postgres documentation calls it a 'lock_id'
- Updated variable names to reflect the preferred terminology for the database in
  question

In b71e08f8 we started raising when nil or false was passed to merge to
fix #12264, however we should also do this for truthy values that are
invalid like true.

If argument of `build_record` has key and value which is same as
default value of database, we should also except the key from
`create_scope` in `initialize_attributes`.
Because at first `build_record` initialize record object with argument
of `build_record`, then assign attributes derived from Association's scope.
In this case `record.changed` does not include the key, which value is
same as default value of database, so we should add the key to except list.

Fix #21893.

From Ruby ( 2.3.0dev trunk 52520), `Hash#to_proc` is defined
(https://github.com/ruby/ruby/commit/fbe967ec02cb65a7efa3fb8f3d747cf6f620dde1),
and many tests have been failed with
`ArgumentError: wrong number of arguments (given 0, expected 1)`.
Because we call `Hash#to_proc` with no args in `#merge!`.

This commit changes order of conditionals to not call `Hash#to_proc`.

Reported on #21509, how views is treated by `#tables` are differ
by each adapters. To fix this different behavior, after Rails 5.0
is released, deprecate `#tables`.

And `#table_exists?` would check both tables and views.
To make their behavior consistent with `#tables`, after Rails 5.0
is released, deprecate `#table_exists?`.

And add code examples to `sanitize_sql_for_conditions`,
`sanitize_sql_for_assignment`, and `sanitize_sql_array`.

Skipping `marked_for_destruction?` when the associated object does not responds
to it make easier to validate virtual associations built on top of Active Model
objects and/or serialized objects that implement a `valid?` instance method.

The previous commit changes the state of the class, and while we are
cleaning up the database, I forgot to clean up the class

I've added a redundant test for this under the attributes API as well,
as that also causes this bug to manifest through public API (and
demonstrates that calling `reset_column_information` on the child
classes would be insufficient)

Since children of a class should always share a table with their parent,
just reloading the schema from the cache should be sufficient here.
`reload_schema_from_cache` should probably become public and
`# :nodoc:`, but I'd rather avoid the git churn here.

Fixes #22057

It goes expected, then actual.

Only changed this because the file was just touched (please don't submit pull requests :)).

These warings have been appeared from
https://github.com/rails/rails/commit/92bc8cdb0771bf6ffcfb31ef58dba529527b514c