Warm up cache to prevent assertion failure.

Remove PG's definition of `type_cast`

Don't rely on the sql type to quote XML columns in PG

It always responds to mounted_helpers now

Avoid retrieving connection in ActiveRecord::QueryCache multiple times.

Remove the `text?` predicate from the type objects

Add callbacks documentation for upgrading to 4.1.

The previous code example for the new explicit block requirement when
setting callbacks was a little confusing. This commit makes the example
more obvious.

As discussed in #15304, we need to automate this process but for now,
having out-of-date versions is not ideal.

Since master targets 4.2.0, let's also update references to the last 4.1
version to 4.2.0.

Finally, let's remove mentions to versions when this is not needed. The
guides cover the features of the current version anyway.

[Juanito Fatas + Robin Dupret]

It is now expected in 4.1 to use an explicit block rather than implicit
when setting callbacks through ActiveSupport::Callbacks. This commit
highlights this new expectation as part of the upgrading documentation.

Fix ActiveSupport::Callbacks #set_callback docs.

A minor version breakage due to a rewrite of the callbacks code now
requires an explicit block to be passed to #set_callback. This amends
the documentation.

This was only used for uniqueness validations. The first usage was in
conjunction with `limit`. Types which cast to string, but are not
considered text cannot have a limit. The second case was only with an
explicit `:case_sensitive => true` option given by the user.

All cases except for `nil` in an array have been removed. `nil` in an
array is handled by the Array type object.

adapter, fixed from #16057 [ci skip]

[ci skip] /sqlite/i --> SQLite

[ci skip] add guide for ActiveModel::Naming

Prior to this commit shallow resources would only generate paths for
non-direct children (with a nested depth greater than 1).

Take the following routes file.

    resources :blogs do
      resources :posts, shallow: true do
        resources :comments do
          resources :tags
        end
      end
    end

This would generate shallow paths for `tags` nested under `posts`,
e.g `/posts/:id/tags/`, however it would not generate shallow paths
for `comments` nested under `posts`, e.g `/posts/:id/comments/new`.

This commit changes the behaviour of the route mapper so that it
generate paths for direct children of shallow resources, for example
if you take the previous routes file, this will now generate
shallow paths for `comments` nested under `posts`, .e.g
`posts/:id/comments/new`.

This was the behaviour in Rails `4.0.4` however this was broken in
@jcoglan's fix for another routes related issue[1].

This also fixes an issue[2] reported by @smdern.

[1] https://github.com/rails/rails/commit/d0e5963
[2] https://github.com/rails/rails/issues/15783

active_record: Type cast booleans and durations for string columns.

[ci skip] link directly to ActionMailer Previews documentation

[ci skip] Concludes changelog links in the bottom.

This way when we do next release notes, this would be easier to copy and change.

This reverts commit a0309775.

This needs more work before it would work correctly on master.

Fixed issue with ActiveRecord serialize object as JSON

Conflicts:
	activerecord/CHANGELOG.md
	activerecord/lib/active_record/attribute_methods/serialization.rb

If our connection is explicitly non-strict, tell MySQL

One of the author fixture we have ("david") references an author address by ID.
Since we disable FK checks when inserting fixtures, this is all fine until we
try to update it, at which point MySQL would complain about the missing row
referenced by the `author_address_id`.

[Godfrey Chan, Matthew Draper]

We default to making the connection strict, but have historically relied
on the MySQL default when we want it to be non-strict. On some (recent?)
versions of MySQL, new connections default to being strict, so if we've
been told 'strict:false', we're obliged to pass that on.

This fixes a test failure that we've seen turn up on relatively-new
development machines, so we do already have a test covering it.

This reverts commit e84799d9, e31104c1 and e6ca8e2c

Use the type object for quoting PG Ranges

Removed mentions of non-existent methods, also added a mention for
`Date.current`, which is analagous to `Time.current`. Closes #16061.