Control port API list reflector clients

[armelvil]

I have found myself in need of some state tracking regarding root ports.

For my use case, I was running a systemd unit with a reflector and updating the reflector clients by terminating,regenerating the systemd unit with a new command line and restarting the reflector process whenever the system needed to add or remove a client from the video stream. The issue with this approach is that it caused a disruption to the video streaming to existing clients every time I needed to dynamically add and remove a client.

So, I was quite happy when I found the control port API existed, and I could use this to dynamically add clients without having to restart the reflector process. So far so good.

The issue I subsequently ran into, is that it's hard to tell which client IP/hostname relates to the root port. Either it requires a wrapper with some kind of external state tracking, or I needed a way to interrogate the API to see what currently exists.

I've generated a fork of UG and attempted to add such functionality by modifying hd_rum_translator.cpp and control_socket.cpp.

This enables a "list-ports" or "query-ports" command which can be submitted via nc
echo "list-ports" | nc 127.0.0.1 6161

This intermittently provides an output of the currently subscribed reflector clients, and in the stdout in the terminal/journald output:
200 OK Root ports: [0] 192.168.108.31:5004 (192.168.108.31:5004) - forwarding [1] 192.168.108.33:5004 (192.168.108.33:5004) - forwarding [2] 192.168.108.34:5004 (192.168.108.34:5004) - forwarding [3] 192.168.108.35:5004 (192.168.108.35:5004) - forwarding
I suspect the repeat message suppression in debug.h is why it will only output the ports list after an addition/removal.

Commit here:
armelvil@42a6fcb

Another thing I am noticing is that it's possible to add the same IP/client multiple times, and it's possible to effectively DDoS target clients by adding multiple entries. To remedy this, I've tried adding a check for an existing client/port which should output:
Output port 192.168.108.35:5004 already exists.

Commit here:
armelvil@ec37ad6

disclosure; because this isn't really my wheelhouse, I've been leaning on LLM's to assist me in understanding the codebase, apologies in advance for the things I have undoubtedly missed and/or failed to take into account.

[armelvil]

I've chipped away at this a little more, and believe I have a functional implementation.

Screencast.From.2026-01-14.12-22-17.mp4

[armelvil]

Opened a PR here if the modifications are acceptable: #489