JavaScript

HTTP Debugging with Node and http-console

    Ian Oxley
    Share

    http-console is a Node module that gives you a command-line interface for executing HTTP commands. It’s great for debugging and seeing exactly what is going on with your HTTP requests, regardless of whether they’re made against a web server, web service, or even a database server.

    Installation

    To use http-console you’ll need to have Node installed. If you haven’t got it installed either head over to http://nodejs.org and download the installer for your operating system, or head over to the Node wiki if you’d prefer to install via a package manager.

    Next, install http-console using npm:

    
$> npm install http-console2 -g

    A couple of things to note:

    • We’re actually installing http-console2, and not http-console. http-console2 is a fork of http-console, but includes a fix for a bug caused by require.paths being deprecated in newer versions of Node. It’s published to npm as http-console2, but once installed you still run it as http-console.

    • We’re installing http-console2 with the -g global switch. This means you can call http-console from anywhere, as it’s installed in a location in your $PATH:

      
$> type http-console
http-console is /usr/local/bin/http-console

    To start using http-console we just pass it the URL and port of whatever it is we want to connect to, and start issuing HTTP commands.

    Speaking HTTP

    Let’s connect to a server and issue some commands. We’ll keep things simple to start with and issue some GET requests to a web server. I’ll assume that, as you’re reading this, you’re a web developer. And, as you’re a web developer, you’ve probably got a web server running on http://localhost. Tell http-console to connect to it by typing in the following:

    $> http-console http://localhost
> http-console 0.6.1
> Welcome, enter .help if you're lost.
> Connecting to localhost on port 80.

    Now you’re connected you can start issuing commands. Type in GET / at the prompt:

    http://localhost:80/> GET /
HTTP/1.1 200 OK
Server: nginx/1.0.11
Date: Wed, 04 Jan 2012 08:40:04 GMT
Content-Type: text/html
Content-Length: 151
Last-Modified: Mon, 04 Oct 2004 15:04:06 GMT
Connection: keep-alive
Accept-Ranges: bytes

<html>
<head>
<title>Welcome to nginx!</title>
</head>
<body bgcolor="white" text="black">
<center><h1>Welcome to nginx!</h1></center>
</body>
</html>

    We get back the full HTTP response, including the HTTP headers, and the HTML itself. You can exit http-console by typing .q

    Let’s try another command. Recently I wrote about the express web framework for Node, we created a page to display the ten most recent tweets mentioning Sitepoint. I wonder what would happen if we use http-console to query Twitter’s Search API for similar tweets?

    $> http-console http://search.twitter.com
> http-console 0.6.1
> Welcome, enter .help if you're lost.
> Connecting to search.twitter.com on port 80.

    Now issue a GET request for /search.json?q=sitepoint&rpp=10:

    http://search.twitter.com:80/> GET /search.json?q=sitepoint&rpp=10
HTTP/1.1 200 OK
Cache-Control: max-age=15, must-revalidate, max-age=300
Expires: Fri, 17 Feb 2012 22:04:02 GMT
Content-Type: application/json;charset=utf-8
Content-Length: 7749
Vary: Accept-Encoding
Date: Fri, 17 Feb 2012 21:59:02 GMT
X-Varnish: 2065334673
Age: 0
Via: 1.1 varnish
Server: tfe

{
    page: 1,
    since_id: 0,
    max_id_str: '170628259464216576',
    refresh_url: '?since_id=170628259464216576&q=sitepoint',
    completed_in: 0.107,
    results: [
        {
            to_user_id_str: null,
            to_user_name: null,
            id: 170628259464216580,
            iso_language_code: 'en',
            ...

    Again, we get back the HTTP headers, but this time we get the body of the HTTP response as JSON (the full JSON is omitted to save space).

    But we’re not restricted to connecting to web servers and web services with http-console. We can also use it to connect to database servers that offer RESTful API’s, such as CouchDB. (If you don’t have CouchDB installed, the easiest way to get up and running is to clone https://github.com/iriscouch/build-couchdb and following the instructions in README.md).

    Assuming CouchDB is running (if you installed via build-couchdb starting CouchDB should be as simple as running . ~/path/to/build-couchdb/build/env.sh, then couchdb), connect http-console to it like so:

    $> http-console http://127.0.0.1:5984
> http-console 0.6.1
> Welcome, enter .help if you're lost.
> Connecting to 127.0.0.1 on port 5984.

    We can then issue commands against the database. Let’s get a list of all the databases:

    http://127.0.0.1:5984/> GET /_all_dbs
HTTP/1.1 200 OK
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Date: Wed, 04 Jan 2012 08:26:18 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 25
Cache-Control: must-revalidate

[ '_replicator', '_users' ]

    How about creating a new database?

    http://127.0.0.1:5984/> PUT /foodb
... 
HTTP/1.1 201 Created
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Location: http://127.0.0.1/foodb
Date: Wed, 04 Jan 2012 09:19:05 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 12
Cache-Control: must-revalidate

{ ok: true }

    Re-issue the GET /_all_dbs command, and we’ll see our new database listed:

    http://127.0.0.1:5984/> GET /_all_dbs
HTTP/1.1 200 OK
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Date: Wed, 04 Jan 2012 09:19:18 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 33
Cache-Control: must-revalidate

[ '_replicator', '_users', 'foodb' ]

    Now let’s add a document to the foodb database. We’ll need to set the Content-Type header to application/json, which is easily done by issuing the .j command (to see all available commands type .help at the http-console prompt):

    http://127.0.0.1:5984/> .j
http://127.0.0.1:5984/> POST /foodb
... { "name":"foo", "body":"bar" }
HTTP/1.1 201 Created
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Location: http://127.0.0.1/foodb/d4a833a173e9d22594b426fd300010a9
Date: Wed, 04 Jan 2012 09:36:30 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 95
Cache-Control: must-revalidate

{
    ok: true,
    id: 'd4a833a173e9d22594b426fd300010a9',
    rev: '1-de4f3804f6f3d2d3a393bec924951e5a'
}

    We can issue HEAD requests to get info about documents, DELETE requests to delete documents, and DELETE requests to delete databases:

    http://127.0.0.1:5984/> HEAD /foodb/d4a833a173e9d22594b426fd300010a9
HTTP/1.1 200 OK
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Etag: "1-de4f3804f6f3d2d3a393bec924951e5a"
Date: Wed, 04 Jan 2012 09:36:51 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 113
Cache-Control: must-revalidate

http://127.0.0.1:5984/> DELETE /foodb/d4a833a173e9d22594b426fd300010a9?rev=1-de4f3804f6f3d2d3a393bec924951e5a
HTTP/1.1 200 OK
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Etag: "2-3ac7397737175948b7a3a6b7e95d2949"
Date: Wed, 04 Jan 2012 09:40:14 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 95
Cache-Control: must-revalidate

{
    ok: true,
    id: 'd4a833a173e9d22594b426fd300010a9',
    rev: '2-3ac7397737175948b7a3a6b7e95d2949'
}

http://127.0.0.1:5984/> DELETE /foodb
HTTP/1.1 200 OK
Server: CouchDB/1.1.1 (Erlang OTP/R15B)
Date: Wed, 04 Jan 2012 09:41:49 GMT
Content-Type: text/plain;charset=utf-8
Content-Length: 12
Cache-Control: must-revalidate

{ ok: true }

    So that was a quick look at using http-console to make and inspect HTTP requests. We made a simple GET requests to a web server, made an API call to Twitter’s Search API, and issued commands to a CouchDB server. Granted YMMV, but hopefully you’ll find it a useful addition to your web development tool belt.