Point to future 4-2-stable branch for detailed changes. [ci skip]

Sync Active Support changelogs with release notes [ci skip]

4.2 Release Notes is a work in progress [ci skip]

Assorted fixes for AS release notes [ci skip]

* `Hash#deep_transform_keys` has already been backported
* Mentioning `ActiveSupport::SafeBuffer#prepend` in the "Deprecations" section would be enough
* Fix syntax in `Kernel#concern` example

Sync AS release notes with the latest commits [ci skip]

Synced 4.2 release notes with commits in railties [ci skip]

Sync AP Deprecations and Notable changes from CHANGELOG [ci skip]

Removed 4.2 release notes from the index [ci skip]

This deprecation was reverted in 489a8f2a [ci skip]

Synced Active Record changelogs for 4.2 release notes [ci skip]

Synced Active Model changelog with the release notes [ci skip]

Added cc26b6b7 as a 'Removal' item in AP release notes [ci skip]

Use 'character' instead of 'symbol' to avoid confusion [ci skip]

Edit pass on 4.2 release notes [ci skip]

📝 💄 Use `.xyz` and `#xyz` appropiately [ci skip]

- image_tag is an instance method
- explain possible results when asset pipeline is used

Related to a discussion on #15508

[ci skip] Clarify ActiveModel::Model docs

ActiveModel attributes initialized with String return String values

This was also removed from #15542

These were removed from #15542 unfortunately :(

Set flash in test session when necessary.

Fix parsed token value with header `Authorization token=`.

PG arrays should type cast user input

We guarantee that `model.value` does not change after
`model.save && model.reload`. This requires type casting user input for
non-string types.

The 401 status should be set first because setting the response body in
a live controller also closes the response to further changes.

Fixes #14229.

`to_session_value` returns nil when empty.

[ci skip] BigDecimal#duplicable? returns false in 1.9.x

Add a test to ensure we don't get a performance regression on #hash

Use a conditional rather than early return in `id`

Defer to super, rather than re-implementing Object#hash

Fix performance regression on preloading HABTM associations

We'd spend a lot of time calling `hash` and `eql?` on the join model,
which has no primary key. Calling `id` with no primary key is a really
slow way to get back `nil`, so we can improve the performance there.
However, even with the escape clause, we *still* weren't getting high
enough performance, as we were checking the primary key too much. `hash`
will always return `nil.hash` for records with no id, and `==` will
always return `false`. We can optimize those cases in the HABTM join
model.

In tag helper, honor html_safe on arrays; also make safe_join more similar to Array.join

fix bug in email with name example code

Correct output of #assert_not

Update test data which doesn't reflect expected usage

In tag helper, honor html_safe on array parameters; also make safe_join more similar to Array.join by first calling flatten.

Topics call `serialize :content`, which means that the values in the
database should be YAML encoded, and we would only expect to receive
YAML strings to `update_column` and `update_columns`.

if the subdomain wasn't specified, it's the same as if specifying
:subdomain as `true`, so we can default the value to `true` safely.

`normalize_host` already calls `named_host?`, so there is no reason to
test `named_host?` again in the `extract_domain` method.

ActiveRecord::Migration.maintain_test_schema! doesn't work with structure.sql

Conflicts:
	activerecord/CHANGELOG.md