This code was needed to work around http://redmine.ruby-lang.org/issues/show/2494. Since that bug is now closed in Ruby, we can remove it.

Instead of marking raw text in templates as safe, and then putting them through String#<< which checks if the String is safe, use safe_concat, which uses the original (internal) String#<< and leaves the safe flag as is. Results in a significant performance improvement.

The real use case it to make all of the following act the same:

  root 'store#index', :as => 'store'
  match '/' => 'store#index', :as => 'store'
  match '/', :to => 'store#index', :as => 'store'

The test case provided deviates from this in order to demonstrate all three
forms in a single set of test routes.
Signed-off-by: NJeremy Kemper <jeremy@bitsweat.net>

Conflicts:
	railties/lib/rails/application.rb

Isolation tests intentionally avoid loading any state (because they're often testing things that have their own load path semantics that should not be polluted), so rack/test is not yet  on the load path. Moving require "rack/test" into the setup means and after boot_rails means that it'll be required after the laod path has been altered to add in the Rails vendor/gems

Routing: fix that route shorthand shouldn't ignore other options. Raise if :as option is given to root method since its name is always 'root'

Shift more responsibility from application class to its singleton instance. Treat instantiation and boot as separate steps. Use app.config rather than app.configuration.

Make /rails/info/properties work again. Also, the mocked up tests were passing so we added a test that actually tested this functionality.

Isolation tests intentionally avoid loading any state (because they're often testing things that have their own load path semantics that should not be polluted), so rack/test is not yet  on the load path. Moving require "rack/test" into the setup means and after boot_rails means that it'll be required after the laod path has been altered to add in the Rails vendor/gems

According to RFC 4627, only the following Unicode code points are
allowed unescaped in JSON:

  unescaped = %x20-21 / %x23-5B / %x5D-10FFFF

However, ActiveSupport::JSON did not escape the range %x00-1f.  This caused
parse errors when trying to decode the resulting output.

[#3345 state:committed]
Signed-off-by: NJeremy Kemper <jeremy@bitsweat.net>

Fixes "Rack-mount boot time is slow as shit" [#3567 state:resolved]