Namespaced routes in rails 1 - ruby-on-rails

I'm working with some legacy code with the last version of rails 1. Upgrading to a later version of rails isn't possible and as such map.namespace is a private method in this version, otherwise I'd be using it.
I have several resources which I have nested into a submenu for simplicity. Their controllers are all within the folder of that submenu and all inherit that controller's namespace.
I'm having major issues with the sub elements. I have mapped the resources and applied a prefix to them in the form of
map.resources :subitem, :path_prefix => "/sub_menu"
but I'm having major issues with using RESTFul methods on these nested objects. The paths that I'm trying to use, say to delete an object simply do not work.
I'm refactoring this current code base so there's very little I can do, and there's very little I can find on rails 1 routes online. It's a bit of a finicky issue but maybe I've missed something simple about routes.
Do I simply need to manually write out every particular route I can use and then manually prefix it, or is there a simpler way?
My other routes work fine, I simply cannot use the
subitem_path(subitem) :method => :delete
function in my link to

So I figured out how to layout the routes appropriately so I could access the paths
Add to the above route the specified controller you're using and the routes will map accordingly
map.resources :subitem, :path_prefix => "/sub_menu, :controller => "submenu/subitem"
Thanks!

Related

Insert routes programmatically in Rails

I'm creating an engine that needs to insert some routes into the application's router. For this particular gem, I'd rather not application's routes.rb if possible. Is there a way to insert routes at a particular location in the router via code? I'm looking for an API that does something like:
Rails.application.routes.insert("resources :foos", :before => "some string in routes.rb")
If I create a config/routes.rb inside the engine and define some routes, this kind of works. Rails is smart enough to mix the engine's routes into the application's routes, but it tacks them on at the end of the route list. I need them to appear at the beginning so the engine's routes take priority.
I'm aware that I can namespace the routes by mounting the engine in the application's routes.rb, but this creates a routing structure that I don't really want. I want the engine's routes to look they are a part of the application by defining some routes in the actual application.
I have a workaround which is to add the following to the application's routes.rb.
Rails.application.routes.draw do
MyEngine.setup_routes(self)
#...other routes below
end
MyEngine.setup_routes looks like
def self.setup_routes(map)
map.get 'a_path', :to => 'a_controller#a_path'
end
This at least allows me to control the point where the routes get defined in the application's route list, but the user still has to manually update his routes.rb (or I have to build an installer that does it). It seems like there should be a way to tell rails to tack some routes onto the start of the route list...

Namespace in rails routing

First, why do we need to namespace controllers?
The example on rails guides shows
namespace :admin do
resources :post, :comments
end
In this case, we have paths such as GET /admin/posts. Is it identical to GET /posts? Or is GET /posts kept as original while another GET /admin/posts is added as extra?
Does rails create any other stuff for namespace?
You dont have to use namespaces if you dont want to, but it's there to make your life easier, specially in big applications with a lot of controllers. And no, routes arent duplicated if you namespace them, unless you specify the route again in your route file without the namespace, but that doesnt make much sense.

What's the "Rails Way" to create route helpers to other hosts?

My app consists of two rails servers with mostly different concerns sitting behind a reverse proxy. Let's call them Server1 and Server2. Occasionally, Server1 needs to render a link to a url on Server2. Is there a good way to use Rails route helpers for this? Specifically in Rails 2? I came up with this:
ActionController::Routing::Routes.draw do |map|
# other routes omitted
map.with_options(:host => 'server2.example.com') do |add|
# create a named route for 'http://server2.example.com/thingies'
add.server2_thingies '/thingies', :controller => 'fake'
# create a named route for 'http://server2.example.com/thingies/combobulate'
add.enhance_server2_thingies '/thingies/combobulate', :controller => 'fake'
# create a named route for 'http://server2.example.com/mabobs/combobulate'
add.enhance_server2_mabobs '/mabobs/combobulate', :controller => 'fake'
# etc..
end
end
So then I can use server2_thingies_url and such in my views. This works, but it makes me feel like a bad person because there is no FakeController and I certainly have no intention of routing requests to it. I considered making helper methods and placing them in app/controllers/application_controller.rb, but a colleague made the argument that it is best to keep all route helpers in routes.rb, so things like rake routes will be accurate, for instance. Is there a better way?
I think I'd make a counter-argument to your colleague: if you're having to dirty up routes.rb with a "FakeController", then your rake routes is still not going to be accurate. In fact, I'd say that this is exactly the kind of thing that a helper was meant to help: it's taking logic that belongs purely in the view (link generation), and removing it from your view templates. Helpers are also easier to maintain and tweak.
Another advantage to the helper style is that if and when it's time to upgrade to Rails 3.x, the less hackish your routes.rb file is, the happier you will be - and that I can attest to from experience. :)
I'd just pass in :host to your foo_path or foo_url calls.
Seems cleaner than messing around with a FakeController

Using Ruby on Rails link_to to link to controller action

I'd just started toying around with Ruby on Rails and had come across an issue with linking to another action in a controller from a particular view. I am almost certain it's an issue (or lack of code) in my routes.rb file, but I think I'm misunderstanding exactly how this file works & what I have to do. I've got a solution but pretty sure it's not the "best way" to do it.
I have one controller called home with two actions, index (which is the default) and newbill. Inside index.html.erb I have:
<h1>Home View</h1>
<%= link_to "new", :controller => "home", :action => "newbill" %>
However I was getting a routing error:
No route matches {:controller=>"home", :action=>"newbill"}
Doing rake routes gives me the following:
root / {:controller=>"home", :action=>"index"}
I then (following some Googling) added this code to routes.rb
match 'home/newbill' => 'home#newbill', :as => :newbill
And then in my index.html.erb I've got this:
<%= link_to "Name", newbill_path %>
And now this works as expected. My questions however are:
Why does this work? What exactly is going on behind the scenes?
Surely this is not the best way to do it? Adding another match 'home/newbill'... for every controller / action I want to link to seems a rubbish way of doing things.
I really like Ruby, but struggling a bit with this aspect of Rails...routing in general is messing up my head a bit I think!
Any help is much appreciated :D
Thanks,
Jack
I guess the first time your code didn't work because your home controller is defined as a resource.
If you define a controller as a resource in routes.rb file it will support only 7 standard methods (according to REST architecture):
index
new
create
show
edit
update
destroy
If you need any more custom routes you should add them manually, say in your case 'newbill', may go as:
resources :home do
collection do
get :newbill
end
end
But as per my understanding, your newbill method should go to bills controllers new, method not in the home controller.
You are right, Rails routes are little bit confusing (at least for me), but once you understand you can do lots of cool stuff.
Read here for the Rails official routes documentation:
http://guides.rubyonrails.org/routing.html.
You should check out the Rails Routing guide. A read through will help you understand what is going on behind the scenes.
This works becuase rails filters every request through the router looking for a match. This enables you to define custom routes such as domain.com/post when the path is actually blog#post. Prior to rails 3, a catch-all route was the last route in the routes file. This allowed you to define a controller and action and it would just work. I'm on my iPad and not near any projects, so I can't verify it, but I think that route is still there in rails 3.1, it just needs to be umcommented.

Ruby on rails path helpers

I know this is a minor issue, but why, if you use scaffolding in RoR, can you use lines like 'new_model name here_path' in link tags, but without using scaffolding, I get a NameError? For example, I have a simple address book app that uses basic CRUD operations. I am a RoR beginner but wanted to build an app without scaffolding and these kind of things don't seem to work. I compared my config/routes.rb and app/helpers/* with those in a scaffolded app and they are no different. What am I missing?
One way to check your routes and paths is to run:
rake routes
It outputs all your routes and paths.
Scaffolding sets up resource routes in the routes.rb file. The resource routes are what give you the path and url helpers. When you don't use scaffolding the routes aren't added, you must do it by hand.
Resource Routes can be added like so:
map.resources :models
where :models is the plural name of one of your models.

Resources