Not sure about mpv, but I use a gstreamer pipeline to tender cameras to the raw KMS terminal display. It works much more reliably (I.e. Predictable loading times and no stutter) compared to loading a DE first. Noting that it was on a low power RPI.
I'm not too sure about the overs can issue, I would have thought the options you found were the right ones.
Another option for display along the same lines might be btop. People like to get fancy with status monitors. Either way, I think yours is looking great.
I would think there is a kernel parameter to either disable a display output or the driver used by the laptop display.
I believe "agetty" is the program that gives you a terminal on a display. Thile general term for what you are interacting with is a getty. You can see an example of setting up a generic automatic login in section 2.3.1 of that link.
The easy way out to run a command on start is to run bash, then have the command in the bottom of your .bashrc file.
the slightly better way to do this is to create a new user that will be just for this purpose (like a service account).
the better way would be to run the command straight through the terminal, instead of starting a shell in interactive mode. This would be replacing ${TERM} with something like /bin/sh -c '/bin/htop'.
I'm not sure how you could get a specific tty on a specific monitor, but I would expect there is a way to do it through the kernel parameters. This probably isn't needed as you don't plan on having another screen anyway. You could just use tty1 and be done with it.
the other thing you may want to do is to set "quiet" mode in the kernel parameters, as the system may print status messages onto the tty be default.
The people here talking about a display server (x11/Wayland) are missing your point I think. If you put a display server on this "box" then it will become a normal server, not a headless one. At that point, you may as well run a full VM and have the output go wherever you want, etc. I'm not sure what the equivalent is in the docker world, but I'm pretty sure that's not what you're asking for.
Is it possible for you to get SSH running on this "server"? If so, you may be able to set it up with an SSH client on the host PC (the laptop?) that is full screen on the CRT/HDMI output?
Maybe I'm misunderstanding as well though. Any headless server I've used in Linux will still give you a TTY on the display. Do you not get that? Someone else mentioned Getty, which is likely the service that is managing that. You should be able to configure Getty to give you a specific tty (e.g. tty9) on a specific output, then configure it either to autologin or to run a script on that tty.
I'll admit, I haven't looked at the code. I would stand by my comment of the unsafe block being a start point.
Countering that however, what is the difference to just debugging effectively? Not sure. I suppose it's down to the people that identified it and fixed it at the end of the day to say if there was any benefit.
I think the other takeaway here is that it was found in a section marked "unsafe". At the very least, that's a useful tool for the Devs to isolate potential problem areas. Comparing that to a pure C codebase where the problem could be anywhere.
Sure, one could learn a new program, new UI, new plugins... Or you could use something that follows your conventions and do what you actually want to do; edit photos (or edit text without knowing the cryptic string to save it).
For me, a full transition never really worked. I've needed up using bash for my login shell to keep general compatibility, then GUI terminals, etc, all get started running fish. Most of the scripts I write still use bash, but the thing I interact with 90% of the time is fish and it's out of the box features cover everything I could want for those times.
There's also tools like "bass" that help for situations where you need to do "source xx.bash" to get dev environments running.
This setup to me is pretty much painless, and doesn't require any upkeep. All wins.
Other comments are on the right track. You likely need to get your music tagged correctly. From memory, you want the Artist tag to be A feat. B, then the Album Artist tag to be A.
MusicBrainz Picard is great. Especially if your music comes from formally released albums. My issue was having lots of indie playlists/albums. I solved this by having Various Artists labelled in the Album Artist tag (I think).
The messaging bubble issue may seem silly, but could I not open the gates to more unified messaging in general? Like WhatsApp to Signal to iMessage? I would hope there is a greater scope than just the bubble.
Not sure about mpv, but I use a gstreamer pipeline to tender cameras to the raw KMS terminal display. It works much more reliably (I.e. Predictable loading times and no stutter) compared to loading a DE first. Noting that it was on a low power RPI.