How does rails differentiate between users/:id and users/new routes? - ruby-on-rails

Question

I am trying to understand how does rails know the difference between the two routes
GET /users/:id
and
GET /users/new
when we type
resources :users
I tried to trace and understand the resources method in rails source code but I didn't understand it completely.
def resources(*resources, &block)
options = resources.extract_options!.dup
if apply_common_behavior_for(:resources, resources, options, &block)
return self
end
with_scope_level(:resources) do
options = apply_action_options options
resource_scope(Resource.new(resources.pop, api_only?, #scope[:shallow], options)) do
yield if block_given?
concerns(options[:concerns]) if options[:concerns]
collection do
get :index if parent_resource.actions.include?(:index)
post :create if parent_resource.actions.include?(:create)
end
new do
get :new
end if parent_resource.actions.include?(:new)
set_member_mappings_for_resource
end
end
self
end
Is the below piece of code does that?
new do
get :new
end if parent_resource.actions.include?(:new)
if yes, can you explain?
Also if I tried to write another route with the same GET users/new format it redirects to GET users/:id, so how can I write another route like GET users/whatever without considering whatever as the :id?
Below is example of routes.rb
Example 1:
get '/feedbacks/:id' => 'feedbacks#show'
get '/feedbacks/count' => 'feedbacks#count'
feedbacks/count redirects to /feedbacks/:id
Example 2:
resources :feedbacks
get '/feedbacks/count' => 'feedbacks#count'
feedbacks/count redirects to /feedbacks/:id

Answer 1

Rails does not differentiate, it's just searches for the first record, which is suitable for the conditions
That's why new is generated earlier than id
Example 1:
get '/feedbacks/count' => 'feedbacks#count'
get '/feedbacks/:id' => 'feedbacks#show'
Example 2:
resources :feedbacks do
member do
get '/feedbacks/count' => 'feedbacks#count'
end
end
You can read about it here

Answer 2

Rails routes are just basically regular expressions on steriods. They match a path expression, a HTTP method and any additional constraints.
Routes have priority in the order they are defined. When the router matches a request in the routes collection it stops searching farther which is why the routes at the top always win.
If you look at the output from the resources macro your can see that the routes are ordered to take this into consideration:
Prefix Verb URI Pattern Controller#Action
things GET /things(.:format) things#index
POST /things(.:format) things#create
new_thing GET /things/new(.:format) things#new
edit_thing GET /things/:id/edit(.:format) things#edit
thing GET /things/:id(.:format) things#show
PATCH /things/:id(.:format) things#update
PUT /things/:id(.:format) things#update
DELETE /things/:id(.:format) things#destroy
The GET /things/new route must be declared before GET /things/:id. Otherwise it would be matched by the things#show route and give an ActiveRecord:: RecordNotFound error as the controller would attempt to find Thing id = "new".
users/:id, so how can I write another route like GET users/whatever
without considering whatever as the :id?
Use the collection option:
resources :things do
get :foo, on: :collection
end
# same but with block syntax
resources :things do
collection do
get :foo
get :bar
end
end
You can also add additional member routes (prefixed with an id):
resources :trips do
patch :cancel
end
Note that Rails defaults to on: :member so you don't need to explicitly set it.
Since resources yields at the top (yield if block_given?) these routes will have the correct priority.
See:
http://guides.rubyonrails.org/routing.html#adding-more-restful-actions