nixd crashes repeatedly with SIGPIPE

[ian-h-chamberlain]

Describe the bug

nixd crashes repeatedly until VSCode client decides not to restart it:

I[20:48:39.631] 60519: unhandled notification $/cancelRequest
I[20:48:39.631] 60519: <-- textDocument/inlayHint(6)
I[20:48:39.631] 60519: --> call attrset/attrpathInfo(3)
[Error - 8:48:39 PM] The Nix server crashed 5 times in the last 3 minutes. The server will not be restarted. See the output for more information.
[Error - 8:48:39 PM] Server process exited with signal SIGPIPE.

Logs (Required)

1-Nix.log

Configuration

I think this is the default from the flake:

{
"security.workspace.trust.enabled": false,
"nix.enableLanguageServer": true,
"nix.serverPath": "nixd",
}

To Reproduce
Steps to reproduce the behavior:

1. Clone https://github.com/ian-h-chamberlain/dotfiles
2. nix develop github:nix-community/nixd#vscodium
3. codium-test
4. Open .config/flake.nix and .config/nix-darwin/configuration.nix (maybe also .config/home-manager/home.nix it seems)
5. Try to do some hovers, go-to-definition, etc.

Expected behavior
Not to crash

Screenshots
Not sure this is necessary but can provide if it helps.

Additional context
None I can think of. I can try to come up with a smaller example too if it's too complicated to figure out with this one, I just notice it pretty frequently in my dotfile so thought it might be easiest to point to that as an example.

[inclyc]

I followed the steps to reproduce the behavior. However on my machine it works perfectly and does not crash. Would you like to provide some additional context?

[ian-h-chamberlain]

Sorry for the delay in responding; other info that might be helpful that I can think of offhand

• nixd, version: 2.3.2 (I can try 2.4.0 too now)
• macOS 14.6.1 (arm64)

My system has these, not sure if codium-test would be different:

• VSCode version: 1.94.0
• Nix IDE extension v0.3.5

Maybe I can try wrapping the nixd invocation in a debugger so I can see where it's crashing? It seems like certain files are more problematic than others, as I can sometimes run nixd without issues - but when I filed this issue I was able to reproduce the crash pretty consistently. I'll post back here if I come up with any additional info.