Remote developement - cannot start the Language server

Hey,

I am trying to use the remote development tools. I have a start_julia.sh as julia.executablePath script that looks like :

#!/bin/bash
module load <modules that need to be loaded before julia can run on this particular cluster>
julia "$@"

The command Julia: Start REPL works fine and launches my script correctly. I am also using the Julia: persistent session: Enabled function, so that I get back my REPL when I restart my client / reopen the remote workspace, which also works correctly.

The only issue is that Starting Julia Language server... never ends and the Language server never starts. The Output: Julia language server looks like :

[Error - 15:21:05] Connection to server is erroring. Shutting down server.
start_julia.sh: line 2: module: command not found
start_julia.sh: line 5: julia: command not found

What should I do ?

Edit: this seems to be highly related to Remote development using VSCode tip - #9 by pfitzseb, although I did not found a solution there…

No, that’s a different issue. Not sure why module and julia work in one context but not another.

What is the difference between the two julia invocations ? (the one for the REPL and the one from the language server) ?

There shouldn’t be any difference relevant to this. I’d recommend checking PATH in the script though.

how ?

By dumping PATH into a file or stderr in your startup script

Checking the paths before issuing the module command gives me exactly the same thing in both cases.


Ok I found the solution. From the Language server run, it looks like the module command is not available. So I needed to load it (by finding it in my architecture first), exactly like linux - module: command not found - Stack Overflow.

The fact that it was not needed for the REPL is weird though.

Anyway, this is related to my architecture and not really the julia vscode extension.