mongo-uri

I parse mongo URIs; especially the tricky repl set ones.

Dear NodeJS MongoDB Drivers

I want to use the mongodb connection string... and not just the simple mongodb://localhost.

As it turns out, as soon as you start working with real... non-localhost/side-project deployments, you need to connect to replication sets. Luckily the defined connection string format accomodates such configurations!

Sadly, you do not understand these connection strings, or have unacceptable bugs with your parsing (e.g. uris are uri encoded).

So, I made a uri parser for you.

Please find bugs in my code, and help me make it the best mongo connection string parser in history.

AND THEN USE IT!

Heads up

I refuse to make assumptions about defaults as part of the parser. So, if a port goes unspecified, I will say that the port is null (because it is), and it is your job to realize that the default port is 27015.

Usage

var MongoURI = require('mongo-uri')
  , uriString
  , uri
  ;

// let's deal with a complicated (yet real-life) uri
uriString = "mongodb://%40u%2Fs%3Fe%3Ar:p%40a%2Fs%3Fs%[email protected],example2.com,example3.com:27018/?readPreference=secondary&w=2&wtimeoutMS=2000&readPreferenceTags=dc:ny,rack:1&readPreferenceTags=dc:ny&readPreferenceTags=";

try {
  uri = MongoURI.parse(uriString);
} catch (err) {
  // handle this correctly, kthxbye
}

console.log(uri);
/*
 * username: "@u/s?e:r" // yay! it got decoded like it was supposed to!
 * password: "p@a/s?s:"
 * hosts: ["example1.com", "example2.com", "example3.com"] // wow, multiple hosts!
 * ports: [null, null, 27018] // remember, I don't assume
 * database: null // what did I just say?
 * options: {
 *   readPreference: "secondary",
 *   w: "2",
 *   timeoutMS: "2000",
 *   readPreferenceTags: ["dc:ny,rack:1", "dc:ny", ""] // yeah, this is correct
 * }
 */

Source

Ok, so I made a Tweet (which is as good as a promise) that my next npm module would be in literate coffeescript. So, I had fun documenting my code, but I do not want the .litcoffee to be transpiled upon npm install of everbody using this code.

GitHub should store the raw source (without the transpiled lib/*.js contents), but npm should store the ready-to-use no-transpiling-needed javascript!

So, I created a grunt task to do my publishing. Instead of using npm publish, I use grunt publish which cleans my directory, compiles my .litcoffee files, and then publishes.

So, don't hate me for using .litcoffee. You're getting raw javascript goodness when you npm install me.

License

MIT