In the previous article, we learned the very basics of networking and how to write a simple TCP server in Ruby.
Here, we'll see a simple implementation of an HTTP server in Ruby.
Dynamic response
As of our example, the TCP server responds with a static message, Hey, client!
.
What if we wanted a dynamic response, for instance:
- client A requests, server responds "Hey, A!"
- client B requests, server responds "Hey, B!"
...and so on.
Let's change our test:
tcp_test.rb
require 'socket'
require 'test/unit'
class ServerTest< Test::Unit::TestCase
def test_client_a
server = TCPSocket.open('localhost', 4242)
request = 'HELLO from client=A'
server.puts(request)
response = server.gets
assert_equal "Hey, A!\n", response
server.close
end
end
And, as for the server implementation, we'll change a little bit to either:
- loop forever. then for each iteration, waits for a client connection and executes the request-response logic
- look up for a match in the request. If request matches our lookup pattern (regex), server responds accordingly
tcp_server.rb
require 'socket'
socket = TCPServer.new(4242)
loop do
client = socket.accept
request = client.gets
if result = request.match(/^HELLO from client=(.*?)$/)
client_id = result[1]
response = "HEY, #{client_id}!"
client.puts(response)
end
client.close
end
- server accepts a client connection
- server checks if request message matches the regex
- in case of match, server responds to the client
Tests passing, time to move on.
Upgrading to HTTP
HTTP is an application layer and protocol used to establish a set of rules for Web applications. HTTP is, essentially, the foundation of Web.
In order to make our client-server to support HTTP, we should request and respond messages using the HTTP format.
HTTP Request
The format of an HTTP request message should be as follows:
{verb} {path} HTTP/1.1\r\n
{request_header_1}\r\n
{request_header_2}\r\n
{request_header_n}\r\n
\r\n
{request_body}
Notes:
- request headers and request body are optional
- carriage return (
\r\n
) are mandatory
Here's the example of an HTTP request message without headers:
GET /customers/42 HTTP/1.1
With headers:
GET /customers/42 HTTP/1.1
Host: myhost.com
Accept: text/html
With headers and body:
POST /customers HTTP/1.1
Content-Type: application/x-www-form-urlencoded
Content-Length: 38
email=test@acme.com&password=pass123
HTTP Response
The format of an HTTP response message isn't much different from the request
HTTP/1.1 {status}\r\n
{response_header_1}\r\n
{response_header_2}\r\n
{response_header_n}\r\n
\r\n
{response_body}
Notes:
- response headers and response body are optional
- carriage return (
\r\n
) are mandatory
Example of an HTTP response message without headers:
HTTP/1.1 204 NoContent
With headers:
HTTP/1.1 201 Created
Location: http://myhost.com/customers/42
With headers and body:
HTTP/1.1 200 OK
Content-Type: text/html
<h1>Ola</h1>
Testing
Let's write a test to send an HTTP request message through the TCP socket at the port 80
. It could be on any other port, e.g 4242
, but let's use port 80 which is the standard for HTTP applications.
http_server_test.rb
require 'socket'
require 'test/unit'
class ServerTest < Test::Unit::TestCase
def test_http_customer_42
server = TCPSocket.open('localhost', 80)
request = """
GET /customers/42 HTTP/1.1\r\n
Accept: plain/text\r\n
\r\n
"""
server.puts(request)
response = ''
while line = server.gets
response += line
end
expected = """
HTTP/1.1 200\r\n
\r\n
Hey, 42!\n
"""
assert_equal expected, response
server.close
end
end
Then, the server implementation as follows:
http_server.rb
require 'socket'
socket = TCPServer.new(80)
loop do
client = socket.accept
first_line = client.gets
verb, path, _ = first_line.split
if verb == 'GET'
if result = path.match(/^\/customers\/(.*?)$/)
client_id = result[1]
response = "HTTP/1.1 200\r\n\r\nHey, #{client_id}!"
client.puts(response)
end
end
client.close
end
socket.close
Test should pass!
HTTP is stateless by default
It's strictly important for us, as developers, to understand one of the main traits of a web server, its stateless nature.
Such knowledge will save many hours of debugging in the future.
Here are the flow of an HTTP server:
- accepts a new client connection
- reads the client request
- performs some logic based on the client request
- sends a response to the client
- closes the client connection
- repeat the process
Note that this loop is important, since the server will forget anything that has been exchanged with a specific client.
HTTP is stateless by default. There are ways to make a persistent stateful HTTP connection, but that's a topic for another blogpost.
HTTP Clients
Until now we've been using TDD, so our test plays the role of the "HTTP client". However there are lots of HTTP clients available.
The Web browser is one of them.
Web browser
Start the server, open your web browser and visit http://localhost/customers/42
. The webpage should display the message in just plain text:
Hey, 42!
Also, check the developer console network of your web browser. You should see something similar to: I'm using Safari in this example.
More HTTP clients:
curl
ubuntu:~ $ curl -v http://localhost/customers/42
* Trying 127.0.0.1:80...
* TCP_NODELAY set
* Connected to localhost (127.0.0.1) port 80 (#0)
> GET /customers/42 HTTP/1.1
> Host: localhost
> User-Agent: curl/7.68.0
> Accept: */*
>
* Mark bundle as not supporting multiuse
< HTTP/1.1 200
* no chunk, no close, no size. Assume close to signal end
<
Hey, 42!
* Closing connection 0
Postman
Summary
In this blogpost we've seen what's HTTP and how to implement a simple HTTP server in Ruby. Next post will focus on HTML, CSS and Javascript, which altogether shape the World Wide Web.