fix(tauri): Webview::navigate unnecessarily borrows &mut self

[WSH032]

close #12430

---

Sorry for opening this PR without permission.

#12430 is blocking me from implementing the Python bindings for Webview::navigate. I plan to release pytauri v0.2 soon, so I hope pytauri can catch up with this PR.

Is there any specific reason why tauri::Webview needs to do this? If not, can we change it to only require an immutable borrow &self?

Not sure if my thoughts are correct, feel free to close this PR if I am wrong.

[github-actions]

Package Changes Through 8df2800

There are 4 changes which include tauri-cli with patch, @tauri-apps/cli with patch, tauri-bundler with patch, tauri with minor

Planned Package Versions

The following package releases are the planned based on the context of changes in this pull request.

package	current	next
tauri-bundler	2.2.2	2.2.3
tauri	2.2.3	2.3.0
@tauri-apps/cli	2.2.5	2.2.6
tauri-cli	2.2.5	2.2.6

Add another change file through the GitHub UI by following this link.

---

Read about change files or the docs at github.com/jbolda/covector

[sftse]

This would break the public API and need a major version update to respect semver.

[FabianLars]

Does this really need a major version? What would a user's code look like that works with &mut self but not &self ? I can't think of anything 🤔 From my testing this doesn't even emit check/clippy warnings which would be totally fine.

Edit: In comparison switching from &self to &mut self would indeed be a clear breaking change.

In context of #12465 and this:

i wouldn't mind a minor instead of patch though.

[WSH032]

i wouldn't mind a minor instead of patch though.

TBH, if strictly following semver, I agree more with releasing it as a minor.

I chose patch just for selfish reasons, as I worry that the tauri v2.3 release cycle might be too long, causing pytauri v0.2 to miss this PR.

Anyway, whether tauri chooses patch or minor, I am okay with it. Feel free to push to this PR.

[FabianLars]

I chose patch just for selfish reasons, as I worry that the tauri v2.3 release cycle might be too long, causing pytauri v0.2 to miss this PR.

I plan to release 2.3 for the new webview2-com and windows version relatively soon - (and yes, that's actually a breaking change but we can't have it locked for the whole lifetime of v2). I just want to get a few last patches in before that.

[sftse]

I withdrew my objection because I tested the cases I had in mind and they turned out to be bogus, do you have an example how this is indeed a breaking change?

fn foo<T>(_: &T) {}
fn bar<T>(t: &mut T) {
    foo(t)
}

Edit: Seems to compile, which I've been reading as proof that this is backwards compatible.

[FabianLars]

do you have an example how this is indeed a breaking change?

as i said, no i don't :D I just can't think of any breakage for &mut self -> &self because you can call &self methods on mutable objects.

&mut self -> &self however is pretty obvious as you need let mut instead of let for &mut self to work.