Tracking Issue for is_symlink

[maxwase]

Feature gate: #![feature(is_symlink)]

Public API

// for `Path` and `Metadata`
pub fn is_symlink(&self) -> bool

Steps / History

• Implementation: Path methods — symlinks improvement  #85747
• Final commenting period (FCP)
• Stabilization PR

Unresolved Questions

None

[marcospb19]

My thoughts on the unresolved questions:

Does exists() or try_exists() method work correctly? How do we implement these for symlinks?

Every function in std::fs is supposed to follow links, except for the ones who contain link name in it:

1. fs::metadata_symlink
2. fs::hard_link
3. fs::read_link

Assuming this pattern is consistent in all library stable functions, I'd say exists() and try_exists() are implemented correctly, they fail if there is a broken link at the path.

If we had exists_symlink() and try_exists_symlink(), I'd assume they both return true for broken links.

---

Linked issues: fs::Metadata::is_dir() and is_file() should be deprecated #76487

I created that issue when I was confused why there wasn't a is_symlink available for fs::Metadata, as seen by the feedback given, people prefer those methods to not be deprecated, as they are useful shorthands to metadata's filetype methods, I agree, and I believe that Path::is_symlink and Metadata::is_symlink would be just as useful.