Part of the larger refactoring to remove type casting from Arel. We can
inform it that we already have the right type by wrapping the value in
an `Arel::Nodes::Quoted`. This commit can be reverted when we have
removed type casting from Arel in Rail 5.1

Part of the larger refactoring to remove type casting from Arel.

Part of the larger refactoring to remove type casting from Arel. We can
inform it that we already have the right type by wrapping the value in
an `Arel::Nodes::Quoted`. This commit can be reverted when we have
removed type casting from Arel in Rail 5.1

There are several valid cases where right now we can't determine the
association's class in a call to `where`. In these cases, we can fall
back to casting by looking up the column from the connection adapter
(which is what happens right now when we fall through to Arel)

This is ugly, and since we're trying to separate the concept of a type
from a column, I'd like to remove it in the future. The problem
basically comes down to this:

    Liquid.joins(molecules: :electrons)
      .where("molecules.name" => "something", "electrons.name" => "something")

The hash in this case will turn into:

    {
      molecules: { name: "something" },
      electrons: { name: "something" },
    }

What we actually need is:

    {
      molecules: {
        name: "something",
        electrons: { name: "something" },
      }
    }

/cc @mrgilman

This code could use some much heavier refactoring. It looks like
`build_relation` duplicates most of the logic of `Relation#where` and
`PredicateBuilder` with regards to handling associations and attribute
aliases

Part of the larger refactoring to remove type casting from Arel. Since
we've already cast the value a few lines above, we don't need to re-cast
it later. We can inform Arel of this by wrapping it in an
`Arel::Nodes::Quoted`, which will no longer be required in Rails 5.1

We will always have the correct type for this query, so no casting is
needed. We inform Arel that we already have the right type by wrapping
it in an `Arel::Nodes::Quoted` (which we will no longer need to do in
Rails 5.1)

This will allow eager type casting to take place as needed. There
doesn't seem to be any particular reason that the `in` statement was
forced for single values, and the commit message where it was introduced
gives no context.

See
https://github.com/rails/rails/commit/d90b4e2615e8048fdeffc6dffe3246704adee01f

Part of a larger refactoring to remove type casting from Arel.

/cc @mrgilman

[Sean Griffin & Melanie Gilman]

A custom object is required for this, as you cannot build a range
object out of `Arel::Nodes::Quoted` objects. Depends on the changes
introduced in
https://github.com/rails/arel/commit/cf03bd45e39def057a2f63e42a3391b7d750dece

/cc @mrgilman

As part of the larger refactoring to remove type casting from Arel, we
need to do the casting of values eagerly. The predicate builder is the
closest place that knows about the Active Record class, and can
therefore have the type information.

/cc @mrgilman

[Sean Griffin & Melanie Gilman]

This class cares far too much about the internals of other parts of
Active Record. This is an attempt to break out a meaningful object which
represents the needs of the predicate builder. I'm not fully satisfied
with the name, but the general concept is an object which represents a
table, the associations to/from that table, and the types associated
with it. Many of these exist at the `ActiveRecord::Base` class level,
not as properties of the table itself, hence the need for another
object. Currently it provides these by holding a reference to the class,
but that will likely change in the future. This allows the predicate
builder to remain wholy concerned with building predicates.

/cc @mrgilman

I'm attempting to remove `klass` as a dependency of the predicate
builder, in favor of an object that better represents what we're using
it for. The only part of this which doesn't fit nicely into that picture
is the check for an association being polymorphic. Since I'm not yet
sure what that is going to look like, I've moved this logic into another
class in an attempt to separate things that will change from things that
won't.

This reduces the number of places which will need to care about single
value or range specific logic as we introduce type casting. The array
handler is only responsible for producing `in` statements.

/cc @mrgilman

[Sean Griffin & Melanie Gilman]

This will allow us to pass the predicate builder into the constructor of
these handlers. The procs had to be changed to objects, because the
`PredicateBuilder` needs to be marshalable. If we ever decide to make
`register_handler` part of the public API, we should come up with a
better solution which allows procs.

/cc @mrgilman

[Sean Griffin & Melanie Gilman]

We're accidentally documenting `PredicateBuilder` and `ArrayHandler`
since there's a constant which is missing `# :nodoc:`

Construction of relations can be a hotspot, we don't want to create one
of these in the constructor. This also allows us to do more expensive
things in the predicate builder's constructor, since it's created once
per AR::Base subclass

We don't memoize the relation instance

I think we should deprecate this behavior and just error if you tell us
to do a case insensitive comparison for types which are not case
sensitive. Partially reverts 35592307

Fixes #18195

`Rails` constant is added by rails-html-sanitizer leading to bugs in
non-Rails apps using ActiveRecord and ActionMailer

Clarify that query methods have a custom definition of whether a numeric value is present. [ci skip]

The way Active Record query methods handle numeric values is a special case, and is not part of Rails's standard definition of present. This update attempts to make this more clear in the docs, so that people don't expect Object#present? to return false if used on a number that is zero.

All cases are properly handled in CollectionAssociation
for all subclasses of this association

We were ignoring the `default_value?` escape clause in the serialized
type, which caused the default value to always be treated as changed.

Fixes #18169

The code for `TableDefinition#references` and
`SchemaStatements#add_reference` were almost identical both
structurally, and in terms of domain knowledge. This removes that
duplication into a common class, using the `Table` API as the expected
interface of its collaborator.

This isn't Seattle.rb, @senny. ;)

The code in `ConnectionPool#release` assumed that a single thread only
ever holds a single connection, and thus that releasing a connection
only requires the owning thread_id.

There is a trivial counterexample to this assumption: code that checks
out additional connections from the pool in the same thread. For
instance:

    connection_1 = ActiveRecord::Base.connection
    connection_2 = ActiveRecord::Base.connection_pool.checkout
    ActiveRecord::Base.connection_pool.checkin(connection_2)
    connection_3 = ActiveRecord::Base.connection

At this point, connection_1 has been removed from the
`@reserved_connections` hash, causing a NEW connection to be returned as
connection_3 and the loss of any tracking info on connection_1. As long
as the thread in this example lives, connection_1 will be inaccessible
and un-reapable. If this block of code runs more times than the size of
the connection pool in a single thread, every subsequent connection
attempt will timeout, as all of the available connections have been
leaked.

Reverts parts of 9e457a86 and
essentially all of 4367d2f0

I just wasted an absurd amount of time trying to figure out why my model
wasn't being deleted even though I was setting `_destroy` to true like
the instructions said. Making the documentation a little bit clear so
that someone like me doesn't waste their time in future.

Update Active Record's attribute query methods documentation to clarify that whether an attribute is present is based on Object#present?. This gives people a place to go see what the exact definition of presence is. [ci skip]

This bug occurs when an attribute of an ActiveRecord model is an
ActiveRecord::Type::Integer type or a ActiveRecord::Type::Decimal type (or any
other type that includes the ActiveRecord::Type::Numeric module. When the value
of the attribute is negative and is set to the same negative value, it is marked
as changed.

Take the following example of a Person model with the integer attribute age:

    class Person < ActiveRecord::Base
      # age          :integer(4)
    end

The following will produce the error:

    person = Person.new(age: -1)
    person.age = -1
    person.changes
    => { "age" => [-1, -1] }
    person.age_changed?
    => true

The problematic line is here:

    module ActiveRecord
      module Type
        module Numeric
          ...

          def non_numeric_string?(value)
            # 'wibble'.to_i will give zero, we want to make sure
            # that we aren't marking int zero to string zero as
            # changed.
            value.to_s !~ /\A\d+\.?\d*\z/
          end
        end
      end
    end

The regex match doesn't accept numbers with a leading '-'.

Update Active Record's attribute query methods documentation to describe its full behaviour. [ci skip]

`ActiveRecord::Base#[]` has overhead that was introduced in 4.2. The
`foo["id"]` working with PKs other than ID isn't really a case that we
want to support publicly, but deprecating was painful enough that we
avoid it. `_read_attribute` was introduced as the faster alternative for
use internally. By using that, we can save a lot of overhead. We also
save some overhead by reading the attribute one fewer times in
`stale_state`.

Fixes #18151

If there is a method defined such as `find_and_do_stuff(id)`, which then
gets called on an association, we will perform statement caching and the
parent ID will not change on subsequent calls.

Fixes #18117

Calling `changed_attributes` will ultimately check if every mutable
attribute has changed in place. Since this gets called whenever an
attribute is assigned, it's extremely slow. Instead, we can avoid this
calculation until we actually need it.

Fixes #18029