Hi HN,

Recently I wasted several hours wrangling my dev environment only to find out that the browser frontend was talking to the wrong version of the backend. This got me thinking--why on earth are we still using simple numbers to describe which process to connect to? Why not use names instead? I thought of all the times a program wouldn't start because of port conflicts. The more I thought about it, the crazier it seemed.

Modern operating systems already offer no-conflict ports: if you make your TCP server listen on port 0, the OS assigns you a random available port. But that only solves half the problem.

I built the other half: PortZero. It's a GPLv3 program that watches for processes (and docker containers!) with a special PZ_TUNNEL environment variable:

  PZ_TUNNEL=myapp-{branch}.portzero.local:80 npm run dev

1. Create a virtual NIC (if it hasn't already)

2. Create a new virtual IP address

3. Create a DNS record that substitutes things like {branch} based on the working directory of the process, and points at that virtual IP address

4. Start listening on that virtual IP address on the port of your choice (e.g. port 80 for http, port 443 for https, port 5432 for postgresql)

5. Forward any TCP connections to that virtual IP address / virtual port to the random, OS-assigned port that your process or container is actually listening on

The result is you don't have to think about ports anymore, you just have to think about subdomains. You can have multiple services available on port 80 without conflicts, as long as they have different portzero.local subdomains.

It does some other cool stuff like:

- Enable HTTPS on local HTTP services by creating a local CA and registering it on your machine

- Enable cloud tunnels so you can access your apps on other devices (paid feature)

Links:

https://portzero.net/docs/

https://github.com/PortZeroNetwork/portzero

> why on earth are we still using simple numbers to describe which process to connect to? Why not use names instead.

Not super popular, but: https://en.wikipedia.org/wiki/SRV_record

For those interested in the history of alternatives to port numbers: Chaosnet[1] in LispM/ITS times used names for describing which service was being requested. On the TCP side, a spec for requesting services by name has existed since 1988: TCPMUX[2], using a short initial handshake after connecting to port 1. Just never caught on AFAIK.

[1] https://chaosnet.net/protocol#application_layer [2] https://datatracker.ietf.org/doc/html/rfc1078

> Enable cloud tunnels so you can access your apps on other devices (paid feature)

It's not much extra work to do this technique (tunnel to publicly accessible domain name from locally or privately hosted setup) with cloudflare tunnels. It's also completely free, all you need is a domain name managed by cloudflare (~10 bucks a year).

You just throw up a cloudflared container (can be part of your local docker compose stack), it will serve whatever you like from that local system at the public domain name - no port forwarding, no IP configuration, just works, over private VPN tunnel.

I use this all the time for personal projects:

> https://developers.cloudflare.com/tunnel/

Unlimited cloud tunnels for 49 dollars a month is not that great - you can have up to 1000 tunnels on the free cloudflare plan, which I'm realistically never going to max out. Once you've done it once, it's a handful of lines to paste into any future compose stack with your Cloudflare API key to use again, it's incredibly simple to use, and will do the same automatic HTTPS setup.

Wild to vibe code something in a week then charge $50/month for it. I understand servers cost money but lol there’s no way the costs are that high. Also insane the amount of code generated here.

> Recently I wasted several hours wrangling my dev environment only to find out that the browser frontend was talking to the wrong version of the backend.

I have never faced this problem even though I hard-code the port. It's because I run the server program through VS Code's debugger. If there's already a session, it would warn me.

I get the problem well. I have built a networking runtime and a Python networking library that extensively used servers as part of testing. If you bind to a specific port and reuse it, you can end up with false failures due to:

- old / zombie server processes

- sockets that weren't properly cleaned up

- TIME_WAITs

my solution wasn't anything as clever as OPs though. Basically, I ended up conceiving of 127/8 as a rotating ring. Where you can pass offsets and easily end up on an unused bind tup every time. 127/8 gives you a huge number of IPs to play with. You just make address handling part of your test launcher and it sets a new IP each time.

The issue is it doesn't really work for IPv6 (v6 localhost is only one address!) Though global-scope v6 interface blocks and manually assigned link-local scope, you could accomplish much the same thing.

Why not have your program just listen on port 80 on an address under 127.0.0.0/8?

You will need to run

  sysctl net.ipv4.ip_unprivileged_port_start=80

I have never really needed something like this. The cloud aspect is a bit different, but Tailscale already does that better at a way lower price.

>Recently I wasted several hours wrangling my dev environment only to find out that the browser frontend was talking to the wrong version of the backend.

For 50 dollars a month you can rent my WebNote application, where you can write down the ports you are using for each of your development applications. With this amazing technology you can compare your environment issues against your documented port choices eliminating port conflicts forever.

I'm so confused: why does it say this thread was created 19 hours ago when this thread was posted like days ago? It says the same thing for my own comments here that were also posted days ago? What is going on? Have I missed something with how HN works?