Incorrect path generation if some controller actions are under path prefix: expected behavior or bug?

rubyonrails-talk
1

Imagine I have a simple scaffolded model Order, controller OrderControllers with all the default views etc.

Then I modify routes to have all actions except index and show be available under /admin subdirectory:

  resources :orders, :only => [:index, :show]   scope '/admin' do     resources :orders, :except => [:index, :show]   end

And after that on my index and show pages the following helper

  link_to 'Show', @order or   link_to 'Show', order

generate links that look like: 'http://localhost:3000/admin/orders/1’ which is obviously incorrect and leads to Routing Error.

I know how to work around this (I'm using link_to 'Show', {action => 'show', :id => @order}) just wondering this is expected behavior for some reason or a bug?

2

Oh, I found the reason.

When routes are defined and helpers are generated in the 'define_hash_access' method, the hash access function is named the same for show and update actions: hash_for_order_path. So the update action's path helper overwrites the show's one.

I still wonder if it is a bug or a feature...

I can assume it may not RESTful what I did - routing show and update to different URLs. Maybe according to REST is expected to be GET and PUT to the same url.

But if specifying {action => 'show', :id => @order} produces correct path, then it seems it is not prohibited to route show and update to different urls.

Vitalii