Navigate and a href fail to pass on query parameters

[stopbystudent]

Describe the bug

In my application, I have a URL that takes query parameters. They are extracted by the router using JavaScript which works well. Now if I call my app with /?filter=text I expect it to receive the query parameter. That indeed works if called directly from the browser's URL bar (typing it manually). Coming from sycamore, however, this has two interesting variations:

• When calling sycamore_router::navigate("/?filter=text") I end up on my 404 page because apparently the router tries to resolve the full page by name instead of trying to resolve / and passing on all query parameters as is. The address bar of the browser, however, contains the correct address /?filter=text; manually reloading the page indeed gets me to the right page with all query parameters present.
• When using a(href = "/?filter=text") { "Link" } I end up on the page for / but with all query parameters cleared, i.e. the filter parameter did not make it to the routed endpoint.

Losing the query parameters while navigating or having to manually reload is a bug, in my opinion.

To Reproduce

pub fn get_query_parameter(key: &str) -> Option<String> {
    web_sys::UrlSearchParams::new_with_str(&web_sys::window()?.location().search().ok()?).ok()?.get(key)
}

pub enum Action {
    View(ViewParameters),
}

#[derive(Clone, Debug, Default)]
pub struct ViewParameters {
    pub filter: Option<String>,
}

impl TryFromSegments for Action {
    fn try_from_segments(param: &[&str]) -> Option<Self> {
        match param {
            [] => Some(Action::View(ViewParameters {
                filter: get_query_parameter("filter"),
            })),
            _ => None,
        }
    }
}

Expected behavior

I expect sycamore-router to be query-parameter-agnostic, i.e. even while #130 is unresolved I expect sycamore-router to pass on query parameters as is, neither ignoring them (the a(href case) nor leading to 404 pages (the navigate case).

Furthermore, navigate should not behave differently from what the href interceptor does on a tags in this regard.

Environment

• Sycamore: 0.8.2
• Browser: various chromium-based browsers
• OS: linux

[lukechu10]

Yup definitely a bug. I think it's because we try to parse the URL but don't succeed because the parser has no knowledge of query params yet. A simple fix should just be to ignore everything after a ? character in the URL parser.