René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"fs: rmSync() reports UV_UNKNOWN with an empty code for unmapped errno values","articleBody":"### Version\nv26.7.0\n\n### Platform\nLinux x64\n\n### Subsystem\nfs / src/node_file.cc\n\n### What steps will reproduce the bug?\n\n```js\nconst fs = require('node:fs');\nfs.mkdirSync('a/b', { recursive: true });\ntry {\n  fs.rmSync('a/.', { recursive: true, force: true });\n} catch (err) {\n  console.log('code   :', JSON.stringify(err.code));\n  console.log('errno  :', err.errno);\n  console.log('syscall:', err.syscall);\n  console.log('message:', err.message);\n}\n```\n\nOutput:\n\n```\ncode   : \"\"\nerrno  : -4094\nsyscall: \"rm\"\nmessage: , Unknown error: Invalid argument 'a/.'\n```\n\n### What is the expected behavior? Why is that the expected behavior?\n\n`err.code` should be `'EINVAL'` and `err.errno` should be `-22`. The underlying\nfailure is a plain `EINVAL` from `rmdir(2)`, which POSIX requires for a path whose\nfinal component is `.`.\n\nEvery other `fs` error carries a usable `code`, and `err.code === 'EINVAL'` is the\ndocumented way to branch on `fs` failures, so an empty string breaks ordinary error\nhandling. The message is also malformed: it begins with `, ` because the empty code\nis interpolated into it.\n\n### What do you see instead?\n\n`errno` is `-4094` (`UV_UNKNOWN`) and `code` is the empty string.\n\n### Additional information\n\nThis is not specific to `EINVAL`. `RmSync()` in `src/node_file.cc` translates the\n`std::error_code` from `std::filesystem::remove_all()` with a hardcoded chain that\ncovers four values:\n\n* `operation_not_permitted` -> `EPERM`\n* `directory_not_empty` -> `ENOTEMPTY`\n* `not_a_directory` -> `ENOTDIR`\n* `permission_denied` -> `EACCES` (`EPERM` on Windows)\n\nAnything else falls through to the tail of the function:\n\n```c++\nstd::string message = \"Unknown error: \" + error.message();\nreturn env->ThrowErrnoException(\n    UV_UNKNOWN, \"rm\", message.c_str(), path_c_str);\n```\n\nSo the mapping is an allowlist rather than a translation, and any errno outside\nthose four surfaces as `UV_UNKNOWN` with an empty `code`. `EINVAL` is simply the\none that is easiest to trigger from user code.\n\nThe async form is unaffected, because `fs.rm()` and `fsPromises.rm()` use the JS\n`rimraf` and report `EINVAL` correctly.\n\nNoting for context that the dot segment handling in #61958 is being addressed\nseparately at the JS layer. That change stops `a/.` from reaching this code path,\nbut it does not fix the mapping, which stays reachable for other errnos.\n","author":{"url":"https://github.com/AmarWaqar-TSKLI","@type":"Person","name":"AmarWaqar-TSKLI"},"datePublished":"2026-09-07T18:40:59Z","url":"https://github.com/nodejs/node/issues/65884"}