Convert route params array into object

This PR converts the route params array into an object and moves the
error generation code into its own object as well. This is a refactoring
of internal code to make it easier to change and simplier for internals
to interface with. We have two new objects:

1) `RouteWithParams`

Previously `#generate` returned an array of parameterized parts for a
route. We have now turned this into an object with a `path` and a
`params methods` to be able to access the parts we need.

2) `MissingRoute`

`#generate` was also responsible for constructing the message for the
error. We've moved this code into the new `MissingRoute` object so it
does that work instead.

This change makes these methods reusable throughout the code and easier
for future refactoring we plan to do.
Co-authored-by: NAaron Patterson <aaron.patterson@gmail.com>

Silence Rake task backtraces

Silence Rake task backtraces, similar to Rails::BacktraceCleaner.
Application lines are preserved, but all other lines are hidden.

This also affects unrecognized tasks, causing the backtrace to be hidden
entirely.  Closes #39524.
Co-authored-by: NPetrik <petrik@deheus.net>

Follow up to #39409.

It is too pessimistic, we can check the attribute whether aggregated or
not by `reflect_on_aggregation`.

```ruby
ActiveRecord::Schema.define do
  create_table :posts, force: true do |t|
  end

  create_table :comments, force: true do |t|
    t.belongs_to :post
  end
end

class Post < ActiveRecord::Base
  has_many :comments
end

class Comment < ActiveRecord::Base
  belongs_to :post
end

post = Post.create!

Benchmark.ips do |x|
  x.report("find_by post") { Comment.find_by(post: post) }
end
```

Before:

```
Warming up --------------------------------------
        find_by post   395.000  i/100ms
Calculating -------------------------------------
        find_by post      4.026k (± 3.1%) i/s -     20.145k in   5.008059s
```

After:

```
Warming up --------------------------------------
        find_by post     1.183k i/100ms
Calculating -------------------------------------
        find_by post     11.859k (± 2.5%) i/s -     60.333k in   5.090685s
```

This reverts commit c1b8833e.

Revert "Lock benchmark-ips version < 2.8"

This reverts commit 62bfa50f.

LoadError is fixed in benchmark-ips 2.8.2.

https://github.com/evanphx/benchmark-ips/pull/101

Delegating to just one line method is to not be worth it.
Avoiding the delegation makes `read_attribute` about 15% faster.

```ruby
ActiveRecord::Schema.define do
  create_table :users, force: true do |t|
    t.string :name
  end
end

class User < ActiveRecord::Base
  def fast_read_attribute(attr_name, &block)
    name = attr_name.to_s
    name = self.class.attribute_aliases[name] || name

    name = @primary_key if name == "id" && @primary_key
    @attributes.fetch_value(name, &block)
  end
end

user = User.create!(name: "user name")

Benchmark.ips do |x|
  x.report("read_attribute('id')") { user.read_attribute('id') }
  x.report("read_attribute('name')") { user.read_attribute('name') }
  x.report("fast_read_attribute('id')") { user.fast_read_attribute('id') }
  x.report("fast_read_attribute('name')") { user.fast_read_attribute('name') }
end
```

```
Warming up --------------------------------------
read_attribute('id')   165.744k i/100ms
read_attribute('name')
                       162.229k i/100ms
fast_read_attribute('id')
                       192.543k i/100ms
fast_read_attribute('name')
                       191.209k i/100ms
Calculating -------------------------------------
read_attribute('id')      1.648M (± 1.7%) i/s -      8.287M in   5.030170s
read_attribute('name')
                          1.636M (± 3.9%) i/s -      8.274M in   5.065356s
fast_read_attribute('id')
                          1.918M (± 1.8%) i/s -      9.627M in   5.021271s
fast_read_attribute('name')
                          1.928M (± 0.9%) i/s -      9.752M in   5.058820s
```

Expand on specs for foreign keys with to_table and add make doc changes

Remove assumption that a primary config exists

Fix schema cache load and corresponding test

Expand on specs of self join and add make doc changes to suggest users to also do  foreign_key when doing self join

This test was incorrect. `primary` was winning for the schema cache load
but when you boot an application it's actually the first configuration
that wins (in a multi db app).

The test didn't catch this because I forgot to add a migrations_paths to
the configuration.

We updated the schema cache loader railtie as well because any
application that didn't have a `primary` config would not be able to use
the schema cache. Originally we thought we'd enforce a `primary`
configuration but no longer feel that's correct. It's simpler to say
that the first wins in a 3-tier rather than implementing a solution to
require `primary` and / or allow aliases.
Co-authored-by: NJohn Crepezzi <john.crepezzi@gmail.com>
Co-authored-by: NJohn Hawthorn <john@hawthorn.email>

Applications may not have a primary configuration so we should not
assume there is one. In both these cases we can get the right connection
without that.

For the databases.rake file we want to re-establish a connection for the
environment we're in. The first config defined under an environment for
a multi-db app will win. This is already the case on application boot so
we should be consistent.

For the info.rb file we already have a connection so we can lookup the
adapter from the connection's db_config. If a primary hadn't existed
this would have thrown an exception.

Followup to https://github.com/rails/rails/pull/39535 which removed the
assumption there was a primary config from the schema cache load
railtie.
Co-authored-by: NJohn Crepezzi <john.crepezzi@gmail.com>

This reverts commit dd1aa738.

Since this ivar is updated in a synchronize block just below,
and has to be protected from other threads
Co-authored-by: NRyuta Kamizono <kamipo@gmail.com>

[ci skip]

This reverts commit 5e0ef9ca.

@num_waiting has to be synchronized with `poll` method
Co-authored-by: NRyuta Kamizono <kamipo@gmail.com>

This reverts commit ea791e53.

reason: these ivars are modified within other synchronize blocks,
        and so need to be blocked from other threads for ensuring therad safety.
Co-authored-by: NRyuta Kamizono <kamipo@gmail.com>

Link to current version of API docs [ci skip]

remove redundant require: follow #39142

It is a super edge case but the alias should be resolved before
`clear_attribute_changes` since `clear_attribute_changes` is not aware
of attribute aliases.

Redundant `to_s` has a few overhead. Especially private methods are not
intend to be passed user input directly so it should be passed always
string.

Removing redundant `to_s` makes attribute methods about 10% faster.

```ruby
ActiveRecord::Schema.define do
  create_table :users, force: true do |t|
  end
end

class User < ActiveRecord::Base
  def fast_read_attribute(attr_name, &block)
    @attributes.fetch_value(attr_name, &block)
  end
end

user = User.create!

Benchmark.ips do |x|
  x.report("user._read_attribute('id')") { user._read_attribute("id") }
  x.report("user.fast_read_attribute('id')") { user.fast_read_attribute("id") }
end
```

```
Warming up --------------------------------------
user._read_attribute('id')
                       272.151k i/100ms
user.fast_read_attribute('id')
                       283.518k i/100ms
Calculating -------------------------------------
user._read_attribute('id')
                          2.699M (± 1.3%) i/s -     13.608M in   5.042846s
user.fast_read_attribute('id')
                          2.988M (± 1.2%) i/s -     15.026M in   5.029056s
```

Promote `clear_attribute_change` as attribute methods

Do not mark attributes as changed when the type is case insensitive a…

Do not mark Postgresql MAC address and UUID attributes as changed when the assigned value only varies by case.

Fix Rubocop warnings.

Fix merge conflict error.

Fix merge conflict error.

For now, `increment` with aliased attribute does work, but `increment!`
with aliased attribute does not work, due to `clear_attribute_change` is
not aware of attribute aliases.

We sometimes partially updates specific attributes in dirties, at that
time it relies on `clear_attribute_change` to clear partially updated
attribute dirties. If `clear_attribute_change` is not attribute method
unlike others, we need to resolve attribute aliases manually only for
`clear_attribute_change`, it is a little inconvinient for me.

From another point of view, we have `restore_attributes`,
`restore_attribute!`, `clear_attribute_changes`, and
`clear_attribute_change`. Despite almost similar features
`restore_attribute!` is an attribute method but `clear_attribute_change`
is not.

Given the above, I'd like to promote `clear_attribute_change` as
attribute methods to fix issues caused by the inconsisteny.

Make custom inspect methods more consistent

Fix issue with expression index in insert_all

Based on the docs which state unique_by option of insert_all can use the
index name if desired, one would expect the method to work normally and
use the `unique_by` option to determine duplicates.

However, there's an issue where the insert_all expects a Set instead
of the string representing the index expression it is given. This causes
an error. Returning the string expression instead of attempting to
format it works perfectly though.

Replace 'Stubs out' with 'Generates' in generator USAGE's [ci skip]

Alias direct_descendants as subclasses

Generators generate things, but what is meant by 'Stubbing out' might
confuse beginners and non-native English speakers.
While generated tests are stubs that should have an implementation, a
generated model is a valid model that doesn't require any changes.

This alias overrides Class#subclasses.  Thus calls to `subclasses` get a
free performance boost when the class extends DescendantsTracker.

Support attribute aliases for `type_for_attribute`