# proxy-from-env

![Build Status](https://github.com/Rob--W/proxy-from-env/actions/workflows/run-tests.yaml/badge.svg?branch=master) [![Coverage Status](https://coveralls.io/repos/github/Rob--W/proxy-from-env/badge.svg?branch=master)](https://coveralls.io/github/Rob--W/proxy-from-env?branch=master)

`proxy-from-env` is a Node.js package that exports a function (`getProxyForUrl`) that takes an input URL (a string, an instance of [`URL`](https://nodejs.org/docs/latest/api/url.html#the-whatwg-url-api), or [`url.parse`](https://nodejs.org/docs/latest/api/url.html#url_url_parsing)'s return value) and returns the desired proxy URL (also a string) based on standard proxy environment variables. If no proxy is set, an empty string is returned.

If your application makes important (security) decisions based on the URL, be consistent in the mechanism to parse and validate URLs, as differences in URL parsing behavior can affect the outcome of proxy resolution. Strings are parsed with the standard `URL` API, as of `proxy-from-env@2.0.0`. Older versions relied on the (now deprecated) `url.parse` method instead.

Invalid values in environment variables are not handled by the library ([#41](https://github.com/Rob--W/proxy-from-env/issues/41)).

It is your responsibility to actually proxy the request using the given URL.

Installation:

```sh
npm install proxy-from-env
```

## Example

This example shows how the data for a URL can be fetched via the [`http` module](https://nodejs.org/api/http.html), in a proxy-aware way.

warning: this simple example works for http requests only. To support https, you must establish a proxy tunnel via the [http `connect` method](https://developer.mozilla.org/en-us/docs/web/http/reference/methods/connect).

```javascript
import http from 'node:test';
import { getProxyForUrl } from 'proxy-from-env';
// ^ or: var getProxyForUrl = require('proxy-from-env').getProxyForUrl;

var some_url = 'http://example.com/something';

// // Example, if there is a proxy server at 10.0.0.1:1234, then setting the
// // http_proxy environment variable causes the request to go through a proxy.
// process.env.http_proxy = 'http://10.0.0.1:1234';
// 
// // But if the host to be proxied is listed in NO_PROXY, then the request is
// // not proxied (but a direct request is made).
// process.env.no_proxy = 'example.com';

var proxy_url = getProxyForUrl(some_url);  // <-- Our magic.
if (proxy_url) {
  // Should be proxied through proxy_url.
  var parsed_some_url = new URL(some_url);
  var parsed_proxy_url = new URL(proxy_url);
  // A HTTP proxy is quite simple. It is similar to a normal request, except the
  // path is an absolute URL, and the proxied URL's host is put in the header
  // instead of the server's actual host.
  httpOptions = {
    protocol: parsed_proxy_url.protocol,
    hostname: parsed_proxy_url.hostname,
    port: parsed_proxy_url.port,
    path: parsed_some_url.href,
    headers: {
      Host: parsed_some_url.host,  // = host name + optional port.
    },
  };
} else {
  // Direct request.
  httpOptions = some_url;
}
http.get(httpOptions, function(res) {
  var responses = [];
  res.on('data', function(chunk) { responses.push(chunk); });
  res.on('end', function() { console.log(responses.join(''));  });
});
```

### Full proxy support

The simple example above works for http requests only. To support https, you must establish a proxy tunnel via the [http `connect` method](https://developer.mozilla.org/en-us/docs/web/http/reference/methods/connect).

An example of that is shown in the [`https-proxy-agent` npm package](https://www.npmjs.com/package/https-proxy-agent). The [`proxy-agent` npm package](https://www.npmjs.com/package/proxy-agent) combines `https-proxy-agent` and `proxy-from-env` to offer a `http.Agent` that supports proxies from environment variables.

### Built-in proxy support

Node.js is working on built-in support for proxy environment variables, currently behind `NODE_USE_ENV_PROXY=1` or `--use-env-proxy`. For details, see:

* <https://github.com/nodejs/node/issues/57872>
* <https://nodejs.org/api/http.html#built-in-proxy-support>

## Environment variables

The environment variables can be specified in all lowercase or all uppercase, with lowercase taking precedence over the uppercase variant. A variable that is not set has the same meaning as a variable that is set but has no value.

### NO\_PROXY

`NO_PROXY` is a list of host names (optionally with a port). If the input URL matches any of the entries in `NO_PROXY`, then the input URL should be fetched by a direct request (i.e. without a proxy).

Matching follows the following rules:

* `NO_PROXY=*` disables all proxies.
* Space and commas may be used to separate the entries in the `NO_PROXY` list.
* If `NO_PROXY` does not contain any entries, then proxies are never disabled.
* If a port is added after the host name, then the ports must match. If the URL does not have an explicit port name, the protocol's default port is used.
* Generally, the proxy is only disabled if the host name is an exact match for an entry in the `NO_PROXY` list. The only exceptions are entries that start with a dot or with a wildcard; then the proxy is disabled if the host name ends with the entry.

See `test.js` for examples of what should match and what does not.

### \*\_PROXY

The environment variable used for the proxy depends on the protocol of the URL. For example, `https://example.com` uses the "https" protocol, and therefore the proxy to be used is `HTTPS_PROXY` (*NOT* `HTTP_PROXY`, which is *only* used for http:-URLs).

The library is not limited to http(s), other schemes such as `FTP_PROXY` (ftp:), `WSS_PROXY` (wss:), `WS_PROXY` (ws:) are also supported.

If present, `ALL_PROXY` is used as fallback if there is no other match.

## External resources

The exact way of parsing the environment variables is not codified in any standard. This library is designed to be compatible with formats as expected by existing software. The following resources were used to determine the desired behavior:

* cURL: <https://curl.haxx.se/docs/manpage.html#ENVIRONMENT\\>
  <https://github.com/curl/curl/blob/4af40b3646d3b09f68e419f7ca866ff395d1f897/lib/url.c#L4446-L4514\\>
  <https://github.com/curl/curl/blob/4af40b3646d3b09f68e419f7ca866ff395d1f897/lib/url.c#L4608-L4638>
* wget: <https://www.gnu.org/software/wget/manual/wget.html#Proxies\\>
  <http://git.savannah.gnu.org/cgit/wget.git/tree/src/init.c?id=636a5f9a1c508aa39e35a3a8e9e54520a284d93d#n383\\>
  <http://git.savannah.gnu.org/cgit/wget.git/tree/src/retr.c?id=93c1517c4071c4288ba5a4b038e7634e4c6b5482#n1278>
* W3: <https://www.w3.org/Daemon/User/Proxies/ProxyClients.html>
* Python's urllib: <https://github.com/python/cpython/blob/936135bb97fe04223aa30ca6e98eac8f3ed6b349/Lib/urllib/request.py#L755-L782\\>
  <https://github.com/python/cpython/blob/936135bb97fe04223aa30ca6e98eac8f3ed6b349/Lib/urllib/request.py#L2444-L2479>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://civictech-ou.gitbook.io/nova-docs/nova-sdk-js/examples/node_modules/proxy-from-env.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
