Tracking Issue for PathBuf::add_extension and Path::with_added_extension

[tisonkun]

Feature gate: #![feature(path_add_extension)]

This is a tracking issue for

• PathBuf::add_extension
• Path::with_added_extension

Public API

impl PathBuf {
    pub fn add_extension<S: AsRef<OsStr>>(&mut self, extension: S) -> bool { ... )
}

impl Path {
    pub fn with_added_extension<S: AsRef<OsStr>>(&self, extension: S) -> PathBuf { ... }
}

Steps / History

• ACP: impl PathBuf::add_extension and Path::with_added_extension libs-team#368
• Implementation: impl PathBuf::add_extension and Path::with_added_extension #123600
• Final comment period (FCP)
• Stabilization PR

Implementation history

• impl PathBuf::add_extension and Path::with_added_extension #123600
• Validate extension in PathBuf::add_extension #140163

Unresolved Questions

None yet.

Footnotes

1. https://std-dev-guide.rust-lang.org/feature-lifecycle/stabilization.html ↩

[MordechaiHadad]

This would be really awesome honestly, just encountered an issue where this would have made my life easier.

[dilawar]

Hoping this becomes available in stable soon. Such a nice QoL api.

[anatawa12]

This new API is awesome!

But I think It would be better to rename (deprecate old and add new name) existing API with_extension with with_replaced_extension and set_extension with replace_extension along with this feature.
I feel existing function with_extension and set_extension is confusing if it will replace or append extension.

Actually I mistakenly used with_extension as with_added_extension and we can see some such case in rust form like this.

[dvdsk]

But I think It would be better to rename (deprecate old and add new name) existing API with_extension with with_replaced_extension and set_extension with replace_extension along with this feature.

I agree with the rename however we can only do that with a new edition (since with_extension is stable) right? And that might take a few more years. That should not block the stabilization of the added API.

[anatawa12]

I don't think edition change is required since we don't have to actually rename the function.
We can introduce with_replaced_extension in some version, and deprecate with_extension in following or later version.

I think there are some cases that deprecating without edition change like deprecate _ms functions with Duration parameter variants.