I'm trying to expose a JSON as a POST request, where I'm trying to append the base-url with another value.
How could I get the base url value?
I tried using:
var root = RED.settings.httpNodeRoot;
but then it returned only /, where as I'm expecting something like http://localhost:1880.
Is it possible to get the base url by using any node-red api? Any help could be appreciated.
It's not directly available, but you can assemble it from some of the available parts.
Have a look at the subscribe function in the Wemo nodes
Basically you can get the port and the path from the RED.settings object, but the IP address very much depends on the machine you are running on. By default Node-RED binds to 0.0.0.0 (which is short hand for all available IP addresses).
If you are running on NodeJS newer than 0.12.x then you can get hold of the IP address of the default route which is normally a fair guess. For NodeJS 0.10.x you pretty much just have to guess.
var ipAddr;
//device.ip
var interfaces = os.networkInterfaces();
var interfaceNames = Object.keys(interfaces);
for (var name in interfaceNames) {
if (interfaceNames.hasOwnProperty(name)) {
var addrs = interfaces[interfaceNames[name]];
for (var add in addrs) {
if (addrs[add].netmask) {
//node 0.12 or better
if (!addrs[add].internal && addrs[add].family == 'IPv4') {
if (ip.isEqual(ip.mask(addrs[add].address,addrs[add].netmask),ip.mask(device.ip,addrs[add].netmask))) {
ipAddr = addrs[add].address;
break;
}
}
} else {
//node 0.10 not great but best we can do
if (!addrs[add].internal && addrs[add].family == 'IPv4') {
ipAddr = addrs[add].address;
break;
}
}
}
if (ipAddr) {
break;
}
}
}
var callback_url = 'http://' + ipAddr + ':' + settings.uiPort;
if (settings.httpAdminRoot) {
callback_url += settings.httpAdminRoot;
}
Looking at this code reminds me I have to add a fix for if HTTPS has been enabled....
Related
I am trying to create a local server accessible via local network in my electron app.
I want to allow incoming access to a specific port.
import server from "./server";
import os from "os";
//TODO determine the local network ip i.e LAN IP
let networkIP = '';
const networkInterfaces = os.networkInterfaces();
//console.log(networkInterfaces);
if (networkInterfaces.hasOwnProperty("lo")) {
delete networkInterfaces.lo;
}
let noConnectedNetwork = true;
let networkAddresses = [];
Object.keys(networkInterfaces).forEach((key) => {
let networkInterface = networkInterfaces[key];
for (let i = 0; i < networkInterface.length; i++) {
let nInterface = networkInterface[i];
if (!nInterface.internal && nInterface.family === "IPv4") {
networkAddresses.push(nInterface.address);
noConnectedNetwork = false;
}
}
});
if (networkAddresses.length > 0) {
networkIP = networkAddresses[0];
} else {
//Start a Wi-Fi hotspot if Wi-Fi hardware is available and get the assigned local IP.
}
//Here I want to enable access to port 6795 for the IPs stored in the networkIP
const eServer = {
start: () => {
server.start(networkIP);
}
};
I want the firewall rule to be cross-platform i.e, works on linux(ufw/iptables), mac and windows.
Also, how do I check for wifi hardware and start a hotspot??
I'm building a simple, STARTTLS capable POP3 Proxy in Node.JS and I'm having quite a hard time.
The proxy serves as a front-end for many back-end servers, so it must load their certificates dynamically, depending on the Client's connection.
I'm trying to use the SNICallback, which brings me the servername the client uses, but I can't set the right certificate after this, because I need one certificate before I have this call, when I create the secure context.
The code is as bellow:
// Load libraries
var net = require('net');
var tls = require('tls');
var fs = require('fs');
// Load certificates (created with openssl)
var certs = [];
for (var i = 1; i <= 8; i++) {
var hostName = 'localhost' + i;
certs[hostName] = {
key : fs.readFileSync('./private-key.pem'),
cert : fs.readFileSync('./public-cert' + i + '.pem'),
}
}
var server = net.createServer(function(socket) {
socket.write('+OK localhost POP3 Proxy Ready\r\n');
socket.on('data', function(data) {
if (data == "STLS\r\n") {
socket.write("+OK begin TLS negotiation\r\n");
upgradeSocket(socket);
} else if (data == "QUIT\r\n") {
socket.write("+OK Logging out.\r\n");
socket.end();
} else {
socket.write("-ERR unknown command.\r\n");
}
});
}).listen(10110);
and upgradeSocket() is as follows:
function upgradeSocket(socket) {
// I need this 'details' or handshake will fail with a message:
// SSL routines:ssl3_get_client_hello:no shared cipher
var details = {
key : fs.readFileSync('./private-key.pem'),
cert : fs.readFileSync('./public-cert1.pem'),
}
var options = {
isServer : true,
server : server,
SNICallback : function(serverName) {
return tls.createSecureContext(certs[serverName]);
},
}
sslcontext = tls.createSecureContext(details);
pair = tls.createSecurePair(sslcontext, true, false, false, options);
pair.encrypted.pipe(socket);
socket.pipe(pair.encrypted);
pair.fd = socket.fd;
pair.on("secure", function() {
console.log("TLS connection secured");
});
}
It handshakes correctly but the certificate I use is the static one in 'details', not the one I get in the SNICallback.
To test it I'm running the server and using gnutls-cli as a Client:
~$ gnutls-cli -V -s -p 10110 --crlf --insecure -d 5 localhost3
STLS
^D (Control+D)
The above command is supposed to get me the 'localhost3' certificate but it's getting the 'localhost1' because it's defined in 'details' var;
There are just too many examples throughout the internet with HTTPS or for TLS Clients, which it's a lot different from what I have here, and even for Servers as well but they're not using SNI. Any help will be appreciated.
Thanks in advance.
The answer is quite simple using tls.TLSSocket, though there is a gotcha with the listeners.
You have to remove all the listeners from the regular net.Socket you have, instantiate a new tls.TLSSocket using your net.Socket and put the listeners back on the tls.TLSSocket.
To achieve this easily, use a wrapper like Haraka's tls_socket pluggableStream over the regular net.Socket and replace the "upgrade"
function to something like:
pluggableStream.prototype.upgrade = function(options) {
var self = this;
var socket = self;
var netSocket = self.targetsocket;
socket.clean();
var secureContext = tls.createSecureContext(options)
var tlsSocket = new tls.TLSSocket(netSocket, {
// ...
secureContext : secureContext,
SNICallback : options.SNICallback
// ...
});
self.attach(tlsSocket);
}
and your options object would have the SNICallback defined as:
var options {
// ...
SNICallback : function(serverName, callback){
callback(null, tls.createSecureContext(getCertificateFor(serverName));
// ...
}
}
There is a mechanism in node to get the remote IP by calling req.connection.remoteAddress.
I tried the below code to get the remote IP of the client but the result is undefined.
var ipAddr = req.headers["x-forwarded-for"];
if (ipAddr){
var list = ipAddr.split(",");
ipAddr = list[list.length-1];
} else {
ipAddr = req.connection.remoteAddress;
}
console.log("req.connection.remoteAddress:" + ipAddr);
Does anyone know if I can get the remoteIP in Parse.com cloud code?
I hope you can help me (us). I'm working on a API project which have two databases :
Production DB : api.myapp.fr
Testing DB : test.api.myapp.fr
Theses two databases are writable by the user.
When a user call our API, he can set the authorization header whichever he needs. For example :
Authorization: s_0
Will perform operations on api.myapp.fr and
Authorization: s_t_0
Will perform operations on test.api.myapp.fr .
My question is : How can I do that with sails ?
Actually, I have a policie which check if the user is using a production key or a testing key, and I override the default models with the one for testings purposes, like this :
if (!is_production) {
req.session.isProd = false;
req.session.logs.environment = "test";
User = UserTest;
Payment = PaymentTest;
PayzenStatus = PayzenStatusTest;
Transaction = TransactionTest;
Card = CardTest;
Doc = DocTest;
}
But you can see the problem if a user makes a test request and then a production request, the models are still the tests ones...
I use my models in services and policies, therefor I can't do
req.models = {};
// If not in production, use the test models
if (!is_production) {
req.session.isProd = false;
req.session.logs.environment = "test";
req.models.User = UserTest;
req.models.Payment = PaymentTest;
req.models.PayzenStatus = PayzenStatusTest;
req.models.Transaction = TransactionTest;
req.models.Card = CardTest;
req.models.Doc = DocTest;
}
// Otherwise use the production models
else {
req.models.User = User;
req.models.Payment = Payment;
req.models.PayzenStatus = PayzenStatus;
req.models.Transaction = Transaction;
req.models.Card = Card;
req.models.Doc = Doc;
}
If you have any idea on how ton achieve this (whatever the way, we can still perform deep changes in our code), I would be really happy to ear it.
Thanks
Two different ways of doing this.
First you could set an environment variable on your production host and check that environment variable to see if you are running in prod. If you are running in prod then use the URI to the production database.
Secodonly, which is probably the better way of doing this creating a config.js file that allows you to read environment variables. What I do for all my apps is set environment variables for connection info to databases and API keys. When running locally/testing I have some defaults in my app but when they environment variables are set they are read and used. So set the environment variable in production to point to your production databases.
The config.js file Im posting below contains references to VCAP, which assumes you are running on Cloud Foundry.
config.js
var VCAP_SERVICES = process.env["VCAP_SERVICES"],
vcapServices;
if (VCAP_SERVICES) {
vcapServices = JSON.parse(VCAP_SERVICES);
}
function getEnv(propName, defaultValue) {
if (process.env[propName]) {
return process.env[propName];
} else {
return defaultValue;
}
}
module.exports = function() {
return {
getEnv : getEnv,
couchDbURL: function() {
// Default to a local couch installation for development
if (VCAP_SERVICES) {
return vcapServices["cloudantNoSQLDB"][0].credentials.url;
}
else {
return "http://localhost:5984";
}
},
couchDbName: function() {
return getEnv("COUCHDB_NAME", "mydb");
}
};
};
app.js
var config = require("./config")();
console.log(config.couchDbURL());
I'm building a Request/Acknowledge/Poll style REST service with NServiceBus underneath to manage queue processing. I want to give the client a URI to poll for updates.
Therefore I want to return a location header element in my web service as part of the acknowledgement. I can see that it is possible to do this:
return new HttpResult(response, HttpStatusCode.Accepted)
{
Location = base.Request.AbsoluteUri.CombineWith(response.Reference)
}
But for a Url such as: http://localhost:54567/approvals/?message=test, which creates a new message (I know I should probably just use a POST), the location will be returned as: http://localhost:54567/approvals/?message=test/8f0ab1c1a2ca46f8a98b75330fd3ac5c.
The ServiceStack request doesn't expose the Uri fragments, only the AbsouteUri. This means that I need to access the original request. I want this to work regardless of whether this is running in IIS or in a self hosted process. The closest I can come up with is the following, but it seems very clunky:
var reference = Guid.NewGuid().ToString("N");
var response = new ApprovalResponse { Reference = reference };
var httpRequest = ((System.Web.HttpRequest)base.Request.OriginalRequest).Url;
var baseUri = new Uri(String.Concat(httpRequest.Scheme, Uri.SchemeDelimiter, httpRequest.Host, ":", httpRequest.Port));
var uri = new Uri(baseUri, string.Format("/approvals/{0}", reference));
return new HttpResult(response, HttpStatusCode.Accepted)
{
Location = uri.ToString()
};
This now returns: http://localhost:55847/approvals/8f0ab1c1a2ca46f8a98b75330fd3ac5c
Any suggestions? Does this work regardless of how ServiceStack is hosted? I'm a little scared of the System.Web.HttpRequest casting in a self hosted process. Is this code safe?
Reverse Routing
If you're trying to build urls for ServiceStack services you can use the RequestDto.ToGetUrl() and RequestDto.ToAbsoluteUri() to build relative and absolute urls as seen in this earlier question on Reverse Routing. e.g:
[Route("/reqstars/search", "GET")]
[Route("/reqstars/aged/{Age}")]
public class SearchReqstars : IReturn<ReqstarsResponse>
{
public int? Age { get; set; }
}
var relativeUrl = new SearchReqstars { Age = 20 }.ToUrl("GET");
var absoluteUrl = HostContext.Config.WebHostUrl.CombineWith(relativeUrl);
relativeUrl.Print(); //= /reqstars/aged/20
absoluteUrl.Print(); //= http://www.myhost.com/reqstars/aged/20
For creating Urls for other 3rd Party APIs look at the Http Utils wiki for example extension methods that can help, e.g:
var url ="http://api.twitter.com/user_timeline.json?screen_name={0}".Fmt(name);
if (sinceId != null)
url = url.AddQueryParam("since_id", sinceId);
if (maxId != null)
url = url.AddQueryParam("max_id", maxId);
var tweets = url.GetJsonFromUrl()
.FromJson<List<Tweet>>();
You can also use the QueryStringSerializer to serialize a number of different collection types, e.g:
//Typed POCO
var url = "http://example.org/login?" + QueryStringSerializer.SerializeToString(
new Login { Username="mythz", Password="password" });
//Anonymous type
var url = "http://example.org/login?" + QueryStringSerializer.SerializeToString(
new { Username="mythz", Password="password" });
//string Dictionary
var url = "http://example.org/login?" + QueryStringSerializer.SerializeToString(
new Dictionary<string,string> {{"Username","mythz"}, {"Password","password"}});
You can also serialize the built-in NameValueCollection.ToFormUrlEncoded() extension, e.g:
var url = "http://example.org/login?" + new NameValueCollection {
{"Username","mythz"}, {"Password","password"} }.ToFormUrlEncoded();