Remove the trailiing slash from a base_path'd router's root URL

[pyrrho]

Currently, given a Dioxus.toml with base_path = "base/path", navigating to e.g. localhost:8080/base/path will result in the router pushing /base/path/ to the web history API, and therefore rendering localhost:8080/base/path/ in the browser address bar. This PR attempts to remove the trailing slash from the root url of base_path'd routers without breaking the functionality of apps that do not specify a base_path.

This isn't a functional bug, but it's a peeve of mine, and I wanted to figure out why this was happening. Additionally, this behavior is inconsistent with

• axum routers that are hosting dioxus router apps; calling .nest("/base/path", MyDioxusMethodRouter) will result in browsers rendering e.g. localhost:8080/base/path/ as the root of the dioxus app, but might cause direct navigations to localhost:8080/base/path/ to 404. (NB. this is probably an issue with nested fallbacks in axum, but it was very frustrating for me while I was figuring that out.)
• subroutes in the same dioxus router; adding e.g. #[route("/not/home/")] (note the trailing slash) will let both localhost:8080/base/path/not/home (note the not trailing slash) and localhost:8080/base/path/not/home/ to resolve to localhost:8080/base/path/not/home (again, note the lack of trailing slash).

I'm opening this as a draft because I am not at all confident in the implementation. The fundamental change is displaying the root URL of routers as "" rather than "/". This might be a bad idea. As you can see from the modified tests, this change requires special-case'ing Links that direct to the root url for non-base_path'd apps. Seeing href="" in an <a> is... kinda weird. Maybe a bit of a red flag. WebHistory::full_path does correctly resolve that to "/" before pushing to the history API, but I've not tested this behavior with liveview or native apps, nor have I tested this change with routers that use the #[layout(...)] or #[nest(...)] features.

I'm planning on building some axum experiments with this change in place, so I may update the PR as I find issues. If there's no additional movement on this PR from myself or the dioxus team in, idk, a few weeks I'll probably close this PR to keep it from taking up space.

[jkelleyrtp]

Closing in favor of #2756 but thank you for the PR!

We will borrow some of the research here in the PR and add you as an author.