Question

For my project i need to have custom urls that should not be easy to guess as it is with increasing integer ids of objects. To solve this problem i gave my objects a unique_id attribute which is a string and wrote the seven actions that would be generated by resources :objects myself, which looks like:

  get '/objects' => 'objects#index', as: :objects 
  post '/objects' => 'objects#create'
  get '/objects/new' => 'objects#new', as: :new_object
  get '/:unique_id/edit' => 'objects#edit', as: :edit_object
  get '/:unique_id' => 'objects#show', as: :object
  put '/:unique_id' => 'objects#update'
  delete '/:unique_id' => 'objects#destroy'

It all works out great, I simply had to tell the controller to find the Objects via Object.find_by_unique_id(params[:unique_id]) instead of Object.find(params[:id])
I can create, read and delete perfect, but i can not update.

When I open the edit_object_path it renders the form correctly, but when i click on “Update Object” it screws up.
From the WEBrick output i can see that it creates a PUT request, but it requests for the id of the object and not the unique_id. So for instance, instead of requesting PUT /osefushe7398hr9 it states PUT /26. The update method is not even called, but instead it asks to create a new object with unique_id of 26.

I really don’t know where to change that behavior or how to pass along the unique_id instead of the id.

Answer 1

Okay the solution was to add

def to_param
  unique_id
end

to the Object model without parameterize or the #{} syntax.
I actually found that out on accident trying to pass the unique_id to the form as mentioned in the second solution. It turns out that i did not have to change the form_for tag like it was suggested in solution two. The parameterize was the reason why the unique_id was made to lower case.

Now everything works as intended (and still is restful)