The cursed way to use a neovim-IDE on the CLOUD
It was a stupid proof-of-concept attempt but it fucking works and I kinda like it.
- So, I have my all-mighty-neovim-config, featuring a set of fully configured out-of-box IDE like perks.
- And I have my thinkpad x200, whose HW is not quite capable for most modern IDE features.
- Also I a pretty powerful VPS which is an overkill for what I use it for .
So I installed my schnitzel.nvim on the VPS, SSH (well I use Mosh but you get the idea) to attach to a tmux session on the server. Then I launched the neovim on the VPS. This is what it looks like:
Also this is the command I used to attach to the remote session:
alias vr="mosh email@example.com -- tmux attach -t 0 -d"
Why not using vim’s builtin remote server support?
This still requires some degree of setup and configurations on the client side: at least you need to have your (n)vim installed and configured and some helper programs like neovim-remote.
This setup, on the other hand, requires ZERO dependency on the client side except for a SSH client. So basically I can use my favorite editor setup on a fairly powerful machine from any terminal machine.
Elephant in the room, how does it actually feel? lag?
Surprisingly it’s almost as smooth as on my local machine. The
around 20~30 ms latency but I don’t even feel it while typing in insert mode.
And the completion menu pops up as I type without any problem.
I notice that when exiting from insert mode to normal mode, the UI (i.e. status line) response is a little behind. I doubt if this is related to the status line plugin that I use. Nevertheless, this latency is only visual, for example when I press ESC, immediately followed by a command, that command is executed correctly in time regardless the visual latency of the status line UI.
 4 Cores of E5-2660 and 8 GBs of RAM. This machine hosts a Pleroma instance and a Matrix instance plus several static sites.