This adds a new method to request and response so we don't need to
violate the law of demeter.

We are changing `Request` and `Response` so that they always have a
`cookie_jar`

This is a continuation on work to combine integration and controller
test code bases in Rails.

We want to get rid of the `Live::Response` so we are consolidating methods
from `Live::Response` and `Response` by merging them together.

This adds an `#empty` method to the request so we don't need to
hard-code the empty array each time we call an empty
`ActionDispatch::Request`.

The work here is a continuation on combining controller and integration
test code bases into one.

The previous spelling seemed a bit too generous with the whitespace, and
looked out of place when amongst others.

- If the assert_response is checking for any non-redirect response like
  :success and actual response is :redirect then, the error message displayed
  was -

     Expected response to be a <success>, but was <302>

- This commit adds the redirect path to the error message of
  assert_response if the response is :redirect.
  So above message is changed to -

     Expected response to be a <success>, but was a redirect to <http://test.host/posts/lol>

* Integrate to raise `ArgumentError`
* Detailed error message when `path` is not defined
* Add a test case, invalid rack app is passed

Make it clear we use `ActionDispatch::Routing::Endpoint`

This code was added for migration from Rails 3.1 to upper,
now we are developing Rails 5.

Closes #21230 by following the indication of @rafaelfranca:

> I think the output change would be simpler.
> What is really important to show is the class of the middleware, so we should change the output to show that.

When a request is made with AJAX and an error occurs, Rails will render
a text-template for the exception instead of the HTML error page
(#11960).

The `.text.erb` variant of the `_source` template is currently missing,
causing HTML to be rendered in the response. This commit adds the text
template.

To keep the page scannable we only only show the first three source
extracts.

Related to #14745.

Before:

```
~/testing-exceptions  ᐅ curl 'http://localhost:3000/' -H
'X-Requested-With: XMLHttpRequest'
RuntimeError in PostsController#index

    <div class="source " id="frame-source-0">
      <div class="info">
        Extracted source (around line <strong>#3</strong>):
      </div>
      <div class="data">
        <table cellpadding="0" cellspacing="0" class="lines">
          <tr>
```

After:

```
~/testing-exceptions  ᐅ curl 'http://localhost:3000/' -H
'X-Requested-With: XMLHttpRequest'
RuntimeError in PostsController#index

Extracted source (around line #3):

*3     raise
```

It used to behave like this:

    url_for(controller: 'x', action: 'y', q: {})
    # -> "/x/y?"

We previously avoided empty query strings in most cases by removing
nil values, then checking whether params was empty.  But as you can
see above, even non-empty params can yield an empty query string.  So
I changed the code to just directly check whether the query string
ended up empty.

(To make everything more consistent, the "removing nil values"
functionality should probably move to ActionPack's Hash#to_query, the
place where empty hashes and arrays get removed.  However, this would
change a lot more behavior.)

This reverts commit 9f93a5ef.

rather than an action name and *args.  The *args were not being used in regular
applications outside tests.  This causes a backwards compatibility
issue, but reduces array allocations for most users.

`dispatch` sets the request and response on the controller for us
automatically, so the test harness doesn't need to know the internals of
how request / response is set.

Conflicts:
	actionpack/lib/action_controller/test_case.rb

`skip_filter`, `skip_action_callback` may both are deprecated in Rails 5.1 so waring msg should be specific.

For ActionController::Base we write the cookies in a middleware if it
was not yet committed no matter if the response was committed or not. [1]

For ActionController::Live we write the cookies before the response is
committed. [2]

We already mimic ActionController::Live in
ActionController::TestCase but we don't mimic the ActionController::Base
behavior because we were checking if the response was committed before
writing the cookies.

Now we are matching the behavior of the middleware and writing the
cookies if it was not written before.

[1]: https://github.com/rails/rails/blob/80c6b901d4d87cee610ab0a438ff6e3c6bf118d1/actionpack/lib/action_dispatch/middleware/cookies.rb#L599-L604
[2]: https://github.com/rails/rails/blob/80c6b901d4d87cee610ab0a438ff6e3c6bf118d1/actionpack/lib/action_controller/metal/live.rb#L218-L223

This is leftover from when `render nothing: true` rendered blank string.

This commit follows up of ea9bc06c.
To check `@response.content_type.to_s` is ended with `"xml"`,
to use `\z` is sufficient.

`Forwardable` has been used in the past

parse RSS/ATOM responses as XML, not HTML

When `require 'active_support/rails'`, 'active_support/deprecation'
is automatically loaded.

Prior to this change, given a route:

    # config/routes.rb
    get ':a' => "foo#bar"

If one pointed to http://example.com/%BE (param `a` has invalid encoding),
a `BadRequest` would be raised with the following non-informative message:

    ActionController::BadRequest

From now on the message displayed is:

    Invalid parameter encoding: hi => "\xBE"

Fixes #21923.

When an application has multiple root entries with different
constraints, the current solution is to use `get '/'`. Example:

**Currently I have to do:**
```ruby
get '/', to: 'portfolio#show', constraints: ->(req) { Hostname.portfolio_site?(req.host) }
get '/', to: 'blog#show',      constraints: ->(req) { Hostname.blog_site?(req.host) }
root 'landing#show'
```

**But I would like to do:**
```ruby
root 'portfolio#show', constraints: ->(req) { Hostname.portfolio_site?(req.host) }
root 'blog#show',      constraints: ->(req) { Hostname.blog_site?(req.host) }
root 'landing#show'
```

Other URL matchers such as `get`, `post`, etc, already allows this, so I
think it's fair that `root` also allow it since it's just a shortcut for
a `get` internally.

Following commit https://github.com/rails/docrails/commit/495722a95687e25114ae75608dd3107ac5d6611b

When generating the url for a mounted engine through its proxy, the path should be the sum of three parts:

1. Any `SCRIPT_NAME` request header or the value of `ActionDispatch::Routing::RouteSet#relative_url_root`.
2. A prefix (the engine's mounted path).
3. The path of the named route inside the engine.

Since commit https://github.com/rails/rails/commit/44ff0313c121f528a68b3bd21d6c7a96f313e3d3, this has been broken. Step 2 has been changed to:

2. A prefix (the value of `ActionDispatch::Routing::RouteSet#relative_url_root` + the engine's mounted path).

The value of `ActionDispatch::Routing::RouteSet#relative_url_root` is taken into account in step 1 of the route generation and should be ignored when generating the mounted engine's prefix in step 2.

This commit fixes the regression by having `ActionDispatch::Routing::RouteSet#url_for` check `options[:relative_url_root]` before falling back to `ActionDispatch::Routing::RouteSet#relative_url_root`. The prefix generating code then sets `options[:relative_url_root]` to an empty string. This empty string is used instead of `ActionDispatch::Routing::RouteSet#relative_url_root` and avoids the duplicate `relative_url_root` value in the final result.

This resolves #20920 and resolves #21459

Use the methods rack provides so we don't have to worry about the exact
header key.

Rack implements the Etag header manipulation methods, so we can use
those instead of ours.