identify.js

var urlmod = require('url');
var http = require('http');
var https = require('https');
var gm = require('gm');
var TIMEOUT = 10000;

function identify(url, callback) {
    var parsed = urlmod.parse(url);
    var get;
    console.log(parsed.protocol);
    if (parsed.protocol == 'https:') {
        console.log("USING HTTPS");
        get = https.get;
    } else {
        console.log("USING HTTP");
        get = http.get;
    }
    var req = get(parsed, function(res) {
        var image = gm(res);
        image.identify(callback);
    });
    req.on('error', function(err) {
        console.error(err);
        callback(err);
    });
    req.setTimeout(TIMEOUT, function() {
        req.abort();
        console.log("timeout");
        callback({name: "Timeout"});
    });
}

if (!module.parent) {
    var httpURL = "http://www.google.co.uk/images/srpr/logo3w.png";
    var httpsURL = "https://www.google.co.uk/images/srpr/logo3w.png";

    identify(httpURL, function(err, features) {
        console.log(httpURL, err, features);
        console.log();
    });
    identify(httpsURL, function(err, features) {
        console.log(httpsURL, err, features);
        console.log();
    });
}

On version 0.8.7 the http case works and the https case fails. Both those urls point to the same image.

Ok so the problem is actually that https.get doesn't take a url string and you have to send it the parsed url (separate host and path). See nodejs/node-v0.x-archive#3882