http self-request

test.ts

// demo of web server and client request.  Lines 17 or 18 can be used and it works.
import http = require('http');

class Test {
  request: http.ClientRequest;

  private Connect() {
    console.log("connecting");
    this.request = http.request({
      hostname: "localhost",
      port: 22440,
      path: "/",
      method: "post",
      headers: {
        "Transfer-Encoding": "chunked"
      }
    //}, (r) => this.OnResponse(r));  // this works
  }, this.OnResponse);                // and this works

    this.request.write("");
    this.request.end();
  }

  private OnResponse(response: http.IncomingMessage) {
      console.log("STATUS: " + response.statusCode);
      console.log("HEADERS: " + JSON.stringify(response.headers));
      response.setEncoding("utf8");
      response.on('data', (chunk: any) => {
        console.log("BODY: " + chunk);
      });
  }

  constructor() {
    this.Connect();
  }
}


(() => {

  http.createServer(function (req, res) {
    res.writeHead(200, {'Content-Type': 'text/plain'});
    res.end('Hello World\n');
  }).listen(22440, 'localhost');

  console.log('Server running at http://localhost:22440/');

  var testConn = new Test();

})();

OK I see what you're saying.

Change the definition of OnResponse to arrow-style.

private OnResponse = (response: http.IncomingMessage) => {

instead of

private OnResponse(response: http.IncomingMessage) {

TypeScript (and ES6) lambdas automatically use a lexically-scoped this instead of the "it depends on how you called it" behavior of functions (or class members) declared in the traditional way.

If you look at the emit, you'll see that this gets captured as _this behind-the-scenes.

Let me know if this doesn't work for you.

I initially didn't like that very much, but after doing something sad and silly, i thought about itmore, and have changed my mind, and like this very very much.

This doesn't break readability of the code at all, and actually implicitly marks a given function as an intended callback.

Thanks!