In the real world, thanks to Rack::Builder, I've started up two sinatra apps and mapped one to "/api/v1" and the other to "/ui".
I'd like the app mapped to /ui to be able to take advantage of the routes in /api but,
becuase they're separate, the ui app can't perform calls on the api side of things.
Is there a way to call one app's route from another via rack, or should I just use Net::HTTP ?
Here's a simplified example of what I'm trying to do:
#/usr/bin/ruby
require 'sinatra'
class API < Sinatra::Base
get '/accounts/' do
'{"json":"account_data"}'
end
end
class UI < Sinatra::Base
get '/accounts/' do
# How do I get /api/accounts?
# call "/api/accounts" obviously does not work
# would use erb here to render accounts list in human readable form
end
end
rack = Rack::Builder.new
rack.map "/ui" do
run UI.new
end
rack.map "/api" do
run API.new
end
Rack::Server.start :app => rack
exit
Thanks very much!
If these apps are racked up in Rack together, they have no real connectivity. What Rack does call .call(env) on whatever app you have racked up and that responds with [status,env,body] You can have layers in the middle that act on and modify env. Your Rack is triggering #call on whatever app you mapped to respond to that path. The called app has no concept of what is running in Rack. Nor can it call another app unless there is a handle for it in env that was populated upstream.
But you have to consider why you decided to make these 2 different apps in the first place. I would imagine it is due to the fact that there could be multiple apps using the API. Or you enjoy the decoupled nature. So in that case you have to think if you would want some coupling through Rack at all. What if the API was on another server? Any internal link between the 2 through Rack would break the UI application. If you always know your API will be self contained on the same machine you could just as well make it a library and not do HTTP calls.
So even if it is possible to call routes from a co racked up app I would never do it. The whole reason for having API servers is to have a decoupled layer. And you should access that layer via HTTP so it would work if it were on same box or across the world.
Related
Given a Rack app that is not Rails,
builder.rb:
def app
Rack::Builder.new{
use Rack::Static, urls:static_paths, root:'public'
run ThaApp
}.to_app
end
How to inject a testing middleware using the spec_helper?
If you're using Builder (with use, run, etc.) it does not look like you can easily inject or remove middleware at runtime. Here's the code: https://github.com/rack/rack/blob/master/lib/rack/builder.rb
Notice that it builds the stack of middleware, and when you call run it instantiates the stack (called "#use") in a tree of middleware objects which each have a reference to the next one - see methods "use" and "to_app".
So: don't think Builder is designed to allow dynamically adding and subtracting middleware in the stack.
You could rebuild a new dynamic stack, or use multiple Rack apps with and without testing middleware, or do some backflips like Rails does to dynamically reconfigure the stack.
You could also add a testing middleware only in test mode, or one that can be deactivated easily so it becomes a pass-through middleware. Then your spec_helper would just set and clear the variable telling it to pass through.
Since I want to prepend to the middleware stack, solving for this particular use case was easy.
Given a app defined as above named "app", add the new middleware: use ...
def new_app
Rack::Builder.new do
use ...
use ...
run app
end.to_app
end
I'm experimenting, and I'm trying to launch dummy Sinatra application from RSpec and kill it when the spec is finished. Something like:
# spec/some_spec.rb
before(:all)
# launch sinatra dummy app
end
after (:all)
# kill sinatra dummy app
end
it 'should return list of whatever' do
expect(JSON.parse(make_request('0.0.0.0:4567/test.json')))
.to include('whatever')
end
I could use system("ruby test/dummy/dummy_app.rb"), but how can I kill that process only? Does anyone know how I can launch the Sinatra inside a test (or from another ruby script)? I know about WebMocks, but I want to see if I can manage to make my test work this way.
Look under RSpec on "Testing Sinatra with Rack::Test". I'd suggest you use that code as boilerplate to get started.
Just add this to your describe block:
def app
Sinatra::Application
end
I would suggest you read up RSpec.
Since you want to test an external system, by the looks of your comment, instead of system "curl whatewer.com", you can use Net::HTTP to make requests and then test against the response.
Have a look at "Testing an external API using RSpec's request specs".
As I'm writing request specs to ensure the features won't be broken I decided to rather write separate Cucumber features. The nice thing is that I can use Capybara, and thanks to Selenium Web Drive, I can launch a server before I run my tests.
So, I created a dummy Sinatra application (that will represent the external service to which the actual code I'm testing is doing requests (including a nasty system('curl whatever.com')).
All I have to do is stub out the methods passed to curl to use Capybara.current_session.server.host and Capybara.current_session.server.port.
Once I'm done with my re-factoring all I have to do is remove the Capybara server variables, and Selenium web drive from the cucumber/capybara configuration.
Tests after a brief change will be still working and will be valid.
Update
In the end I wrote it all with RSpec request tests, as doing it in Cucumber was little bit time consuming and I already spend too much time on this.
I mark these kind of request tests with RSpec tag and Before I lunch these I manually lunch simple Sinatra/Grape dummy API application to which the request are made. (Then I run RSpec tests with this tag)
So basically I end up with specs for functionality that uses net/http that uses WebMock and don't need a server, and request tests for which I need to run the server before I run the specs. So the original question remains, how to lunch a server before tests start
After I cover all the functionality I'm gonig to rewrite the curl to net/http however I'm going to keep those requests specs as I discovered they are nice idea when it comes to crazy API scenarios (like testing https + diggested authentication)
i use framwork Espresso with Enginery generator (Ruby gems). I create new project, and not understand how work this application. I will work with Espresso. Explain me struct Expresso Application, please.
I can run this apllication: rackup config.ru
I can edit controllers, but i not understand depending between ruby scripts in this project.
i run projects, but why this work it?
config.ru
require File.expand_path('../base/boot', __FILE__)
puts App.urlmap
run App
in project not /base/boot directory.
requiring '../base/boot' will actually load dependencies, controllers, models etc. and build the application.
The application are stored under App constant, so you can access it from different files:
https://github.com/espresso/enginery/blob/master/app/base/base/boot.rb#L9
puts App.urlmap will display all the routes to be served by app.
And run App will start your app.
You can also start app by ruby app.rb, then you do not need to pass server/port at startup.
Instead you'll set them in config/config.yml, like this:
development:
server: Thin
port: 5252
The config.ru file looks quite normal for a rack start-up file. You would start the application from the project folder, with a command like:
rackup -p 8080
The following line:
require File.expand_path('../base/boot', __FILE__)
will load in the ../base/boot file (similar to require_relative, but also works with older Ruby e.g. 1.8.7), which I would guess requires the dependencies where App is defined. The class or module App will implement a call method. To start the server, the rack host calls App.new (which is called due to run App) and then on each request it will call .call( env ) on the resulting object (the object doesn't have to be an App object, but in simpler frameworks it will be).
The variable env contains all the details of the request and the rack environment that can be inspected to fetch details of the current path, cookies, query params etc. Typically access to this data is abstracted through Sinatra and Espresso helper methods that you will use.
The Sinatra and Espresso helper methods look like they are doing magic declarations, but they are just normal methods. Usually they do some calculation and then stash a code block/lambda for rack to call later. Sinatra's get is like this . . . it isn't true declarative code. Instead, when the controller is parsed, it just takes the code block and tells the application object to call it (later) when the path matches.
I have a rails app called X and a rails mountable engine called Y.
App X has Devise and engine Y has CanCan which works with the hosts apps devise installation. In order to seperate the concerns of host app and mountable engine, the engine has a CanCan ability model as well as the host app.
My problem is the following, as per CanCan documentation you can change the default ability model like so https://github.com/ryanb/cancan/wiki/Changing-Defaults.
My problem is that if I setup the of default ability class in the mountable engines application controller like so:
module Y
class ApplicationController < ActionController::Base
def current_ability
#current_ability = Ability.new(current_user)
end
end
end
I get the following error:
uninitialized constant Ability
(On a side note, when I edit the file and reload the page without restarting the server it works fine)
If however I put the same method into the host app's application controller:
class ApplicationController < ActionController::Base
def current_ability
#current_ability = Y::Ability.new(current_user)
end
end
Everything works as expected (which also means that Devise and CanCan are setup correctly).
Anyone got any ideas of why this is happening? I might not understand mountable engines correctly however I always thought that they are isolated from the host app, thus when accessing a page defined by the mountable engine the host app's application controller shouldn't even be executed.
it is most likely that your engines application controller is not loaded at all. Rails engines behave bad if you have same class name in the main app. here is the issue if you want to look at. I would recommend using the explicit declaration in your engine. Use Y::Ability instead of Ability.
What framework do you recommand for writing simple web applications in ruby, between WebRick, Mongrel and Sinatra ?
I would like to answer in json to requests from a client. I would like to have my own code decoupled from the Http framework as much as possible.
Do you know any other framework ?
I wouldn't recommend using WEBrick, period. You would best be served by a Rack-compatible framework. You could write directly in Rack for speed, but it's really unnecessary since Sinatra is so much more pleasant and still very fast.
You may also want to check out Halcyon. I don't know if it's still maintained, but it's designed for writing APIs that respond in JSON.
WEBrick and Mongrel are servers, not frameworks for building web applications. As such, they have APIs that are lower level and tied to their own idiosyncrasies which makes them a bad place to start if you want to design your web application so that it can run on different servers.
I would look for a framework that builds on Rack, which is the standard base layer for building web apps and web frameworks in Ruby these days.
If you are making something really simple, learning Rack's interface by itself is a good place to start.
E.G., a Rack Application that parses json out of a post request's body and prints it back out prettified.
# in a file named config.ru
require 'json'
class JSONPrettyPrinterPrinter
def call env
request = Rack::Request.new env
if request.post?
object = JSON.parse request.body
[200, {}, [JSON.pretty_generate(object)]]
else
[200, {}, ["nothing to see here"]]
end
end
end
run JSONPrettyPrinterPrinter
you can run it by running rackup in the same dir as the file.
Or, if you want something a bit more high level, you can use sinatra, which looks like this
require 'sinatra'
post '/' do
object = JSON.parse request.body
JSON.pretty_generate(object)
end
Sinatra's README is a good introduction to it's features.