Panic with nested islands

[boyswan]

Currently, rendering an island directly in another panics (ex: OuterFailing). The only way around this is to render via children (ex: OuterWorking).

This can be problematic as it uses up the child prop, and the slot solution to this is incompatible with islands. As a workaround I have various "island wrappers" which wrap a component in an empty island component, but it can get quite messy.

Perhaps what I'm attempting is an anti-pattern and nesting islands should be best avoided?

use leptos::*;

#[island]
pub fn CommonIsland() -> impl IntoView {
    let val = RwSignal::new(0);
    view! {
        <div>
            {move || format!("CommonIsland value is {}", val.get())}
            <button on:click=move|_| val.update(|x| {*x += 1})>Click</button>
        </div>

    }
}

#[island]
pub fn OuterWorking(children: Children) -> impl IntoView {
    let val = RwSignal::new(0);
    view! {
        <>
            <div>
                {move || format!("outer value is {}", val.get())}
                <button on:click=move|_| val.update(|x| {*x += 1})>Click</button>
            </div>
            {children()}
        </>

    }
}

#[island]
pub fn OuterFailing() -> impl IntoView {
    let val = RwSignal::new(0);
    view! {
        <>
            <div>
                {move || format!("outer value is {}", val.get())}
                <button on:click=move|_| val.update(|x| {*x += 1})>Click</button>
            </div>
            <CommonIsland/>
        </>

    }
}
#[component]
pub fn Example() -> impl IntoView {
    view! {
        <OuterFailing/>

        <OuterWorking>
            <CommonIsland/>
        </OuterWorking>

        <CommonIsland/>
    }
}

[gbj]

Perfectly reasonable bug report. It's not likely I'll have the time to investigate it further in the near future (say, before mid-April) but would appreciate any help you can give in understanding the cause.

[gbj]

(Oh I should add: It is likely this will be fixed in 0.7, which is where I'm spending most of my energy, as the hydration works completely differently. I will tag this with 0.7 to make sure it is closed in that version if not before.)

[boyswan]

Great! In the meantime I'll do a little digging and see if I can find anything. Will update on here if so.

[boyswan]

So I have found the offending unwrap:

leptos/leptos_dom/src/components/dyn_child.rs

Line 348 in fda4dba

.unwrap()

Handling this removes the panic and the island works as expected, however hydration warnings remain. My knowledge of the inner works of Leptos are limited to the last hour, so I can only offer so much until I learn more!