Stackoverflow error when installing

[OlshaMB]

[OlshaMB]

deps list:

"dependencies": {
    "@editorjs/editorjs": "^2.27.2",
    "@types/node": "20.4.5",
    "@types/react": "18.2.17",
    "@types/react-dom": "18.2.7",
    "autoprefixer": "10.4.14",
    "eslint": "8.46.0",
    "eslint-config-next": "13.4.12",
    "next": "13.4.12",
    "postcss": "8.4.27",
    "react": "18.2.0",
    "react-dom": "18.2.0",
    "tailwindcss": "3.3.3",
    "typescript": "5.1.6"
  }

nextjs 13 deps + editorjs

[anonrig]

Thanks for the bug report. I'll try to add better diagnostics!

[OlshaMB]

@anonrig Do you have any idea why the stackoverflow would happend when tokio futures are used?

[OlshaMB]

They are effectively threads???

[anonrig]

I'm seeing a different error in my content-addressable-store branch:

➜  bug-report git:(content-addressable-store) ✗ ../target/debug/pacquet install
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: FileSystem(Os { code: 21, kind: IsADirectory, message: "Is a directory" })', crates/package_manager/src/commands/add.rs:95:10
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[anonrig]

I released v0.1.2. Can you check if the issue persists? https://www.npmjs.com/package/pacquet

[steveklabnik]

They are effectively threads???

Yes and no. But in this sense, they're the same: there's an amount of stack that they get, and they can overflow it. In theory, Rust + Tokio should be able to analyze the exact amount of stack needed, and do that, but with recursion, stuff gets a bit more wonky, and can lead to overflow, just like with threads.