Use generated binstubs in guides examples. [ci skip]

[ci skip]

While running Sqlite3 memory tests I encountered the following error:

```
Finished in 69.416366s, 58.0267 runs/s, 162.3681 assertions/s.

  1) Error:
ActiveRecord::Migration::ChangeSchemaTest#test_add_column_with_timestamp_type:
NoMethodError: undefined method `type' for nil:NilClass
    /Users/senny/Projects/rails/activerecord/test/cases/migration/change_schema_test.rb:244:in `test_add_column_with_timestamp_type'

4028 runs, 11271 assertions, 0 failures, 1 errors, 1 skips
```

This was because the table `testings` was used in multiple test-cases.
This resulted in a wrongly cached schema on `ActiveRecord::Base.schema_chae`.

/cc @zuhao

Closes #15122
Closes #15107

Un-define :to_json for Customer class after stubbing.

Deregister csv renderer after test to prevent leak.

This reverts commit 5c224de9.

Conflicts:
	actionpack/lib/action_dispatch/journey/visitors.rb

5c224de9 introduced a bug in the
formatter.  This commit includes a regression test.

Remove :timestamp column type

The `:timestamp` type for columns is unused. All database adapters treat
them as the same database type. All code in `ActiveRecord` which changes
its behavior based on the column's type acts the same in both cases.
However, when the type is passed to code that checks for the `:datetime`
type, but not `:timestamp` (such as XML serialization), the result is
unexpected behavior.

Existing schema definitions will continue to work, and the `timestamp`
type is transparently aliased to `datetime`.

Default config.assets.digests to true in development

Fix migrations that use enable_extension with table_name_prefix/suffix

Conflicts:
	activerecord/CHANGELOG.md
	activerecord/lib/active_record/migration.rb

Delegate `Column#type` to the injected type object

The decision to wrap type registrations in a proc was made for two
reasons.

1. Some cases need to make an additional decision based on the type
  (e.g. a `Decimal` with a 0 scale)
2. Aliased types are automatically updated if they type they point to is
  updated later. If a user or another adapter decides to change the
  object used for `decimal` columns, `numeric`, and `number` will
  automatically point to the new type, without having to track what
  types are aliased explicitly.

Everything else here should be pretty straightforward. PostgreSQL ranges
had to change slightly, since the `simplified_type` method is gone.

Return early on delete and destroy methods

Make `:index` in migrations work with all column types

Use destructured arguments when looping through pairs

Partial revert of c0bfc3f4

Reorganize ActiveRecord tasks [Arun Agrawal & Abd ar-Rahman Hamidi]

This is a reaction to https://github.com/rails/rails/commit/7ca75f3789e316b6f3dc1f8000ec4f85c8ba32ea#commitcomment-6303828

PostgreSQL timestamps should always be datetimes

I didn't want to do this, FNM_EXTGLOB is defined on 2.1.x, but Dir.glob
returns the wrong value on Ruby less than 2.2.0.  Checking for a
case-insensitive FS seems too hard, so just check Ruby version  Checking
for a case-insensitive FS seems too hard, so just check Ruby version.

Remove cruft now that strong_parameters is default [ci skip]

remove ecosystem link that is broken [ci skip]

[ci skip]

When delete or destroy is called on all records nothing
is deleted or destroyed. Intead of running through the code and still
not deleteing anything, we should early return

Add a type object to Column constructor

Part of #15134. In order to perform typecasting polymorphically, we need
to add another argument to the constructor. The order was chosen to
match the `oid_type` on `PostgreSQLColumn`.

Minor refactoring of looping behavior for aggregation