This inlines casting for the most obvious types. The rest will
follow eventually. I need to put some tests in place, to make sure
that the inlining is not causing regressions.

/cc @sgrif

the issue tracker. See also #15455 [ci skip]

New records should remain new after yaml serialization

Create Changing Columns on migrations guides [ci skip]

duplicating API reference in guides. Also ✂ [ci skip]

Add Total of grouped items to Active Record query interface on guides [ci skip]

Remove most code related to serialized properties

Nearly completely implemented in terms of custom properties.
`_before_type_cast` now stores the raw serialized string consistently,
which removes the need to keep track of "state". The following is now
consistently true:

- `model.serialized == model.reload.serialized`
- A model can be dumped and loaded infinitely without changing
- A model can be saved and reloaded infinitely without changing

Don't change values in `@raw_attributes` during serialization

During `init_with`, the attributes given to the coder will be placed
into `@raw_attributes`. As such, we should read from `@raw_attributes`
when encoding, rather than `@attributes`, which has been type cast.

Swallow bundler exceptions when running tests

[ci skip]

[ci skip]

config.assets.digests -> config.assets.digest

in Asset Pipeline guide.

[skip ci]

be published and requires more attention. [ci skip]

[ci skip] correct doc for ActionView::Helpers::DebugHelper

[ci skip] Correct output of Hash#symbolize_keys

Clear all caches calculated based on `@columns` when `@columns` changes

[ci skip] Correct documentation of HashWithIndifferentAccess#dup

Update documentation for setting asset_host to a Proc [ci skip]

Update documentation for setting asset_host to a Proc, mention that the request parameter might not be supplied [ci skip]

Add missing test case for writing unknown attributes

HashWithIndifferentAccess#dup doc [ci skip]

The phrase "exact copy" in the existing docmentation is somewhat
misleading.

Update some bad test cases for serialized columns

The first case was not testing what the issue mentioned actually was (A
subclass of a class with serialized attributes does not serialize when
they come from the database).

The second case was a bad coder. It would fail if the model was `dup`ed,
or if the the model was loaded from the database and then saved again.

The third case wasn't testing anything that wasn't covered by the second
(and was also a bad coder for the same reasons as the second).

Rename attribute related instance variables to better express intent

Allow specifying a default value in overloaded properties