Hi! I'm trying to figure out the recommended way to use terminal-code together with Herdr for remote development.
My desired setup is:
Local machine
├── Herdr client / TUI
└── terminal-code client / frontend
│ SSH
▼
Remote machine
├── Herdr server / terminals
├── project files
└── terminal-code / code-server backend
In other words, I want all project-related servers/processes to run on the remote machine, while the interactive clients/rendering stay on my local machine.
Both tools seem to support this independently:
- With terminal-code, I can run
tode --ssh <host> locally, so the backend runs remotely while the frontend stays local.
- With Herdr, I can use
herdr --remote <host> and keep the Herdr client local while the actual terminals/processes live on the remote machine.
The part I'm unsure about is how to combine them.
Once I'm inside a terminal created by the remote Herdr session, that shell is already running on the remote machine. If I run:
from that terminal, tode itself is launched remotely, which means the terminal-code client/frontend also ends up running on the remote side instead of locally.
I also noticed that this repository already includes a Herdr plugin, but from what I understand, open-split invokes tode --split right from the current Herdr pane, so I assume the same issue would apply when Herdr itself is attached to a remote machine.
Is there currently a supported/recommended way to achieve the topology above?
Ideally, I'd like to be able to trigger tode from a remote Herdr pane, but have it:
- start or connect to the terminal-code backend on that remote machine, and
- open the terminal-code frontend on my local machine.
For example, is there some mechanism for the remote tode command / Herdr plugin to communicate back to a local terminal-code client, similar to how tode --ssh works when initiated locally?
If this isn't currently supported, would this be something that could potentially be supported by the Herdr plugin?
Thanks!
Hi! I'm trying to figure out the recommended way to use terminal-code together with Herdr for remote development.
My desired setup is:
In other words, I want all project-related servers/processes to run on the remote machine, while the interactive clients/rendering stay on my local machine.
Both tools seem to support this independently:
tode --ssh <host>locally, so the backend runs remotely while the frontend stays local.herdr --remote <host>and keep the Herdr client local while the actual terminals/processes live on the remote machine.The part I'm unsure about is how to combine them.
Once I'm inside a terminal created by the remote Herdr session, that shell is already running on the remote machine. If I run:
from that terminal,
todeitself is launched remotely, which means the terminal-code client/frontend also ends up running on the remote side instead of locally.I also noticed that this repository already includes a Herdr plugin, but from what I understand,
open-splitinvokestode --split rightfrom the current Herdr pane, so I assume the same issue would apply when Herdr itself is attached to a remote machine.Is there currently a supported/recommended way to achieve the topology above?
Ideally, I'd like to be able to trigger
todefrom a remote Herdr pane, but have it:For example, is there some mechanism for the remote
todecommand / Herdr plugin to communicate back to a local terminal-code client, similar to howtode --sshworks when initiated locally?If this isn't currently supported, would this be something that could potentially be supported by the Herdr plugin?
Thanks!