//! Example: Antipatterns //! --------------------- //! //! This example shows what *not* to do and provides a reason why a given pattern is considered an "AntiPattern". Most //! anti-patterns are considered wrong due performance reasons or violate the "rules" of Dioxus. These rules are //! borrowed from other successful UI frameworks, and Dioxus is more focused on providing a familiar, ergonomic interface //! rather than building new harder-to-misuse patterns. //! //! In this list we showcase: //! - Not adding keys for iterators //! - Heavily nested fragments //! - Understanding ordering of set_state //! - Naming conventions //! - Rules of hooks //! //! Feel free to file a PR or Issue if you run into another antipattern that you think users of Dioxus should know about. use dioxus::prelude::*; /// Antipattern: Iterators without keys /// ----------------------------------- /// /// This is considered an anti-pattern for performance reasons. Dioxus will diff your current and old layout and must /// take a slower path if it can't correlate old elements with new elements. Lists are particularly susceptible to the /// "slow" path, so you're strongly encouraged to provide a unique, stable ID between renders. Additionally, providing /// the *wrong* keys is even worse - props might be assigned to the wrong components! Keys should be: /// - Unique /// - Stable /// - Predictable /// /// Dioxus will log an error in the console if it detects that your iterator does not properly generate keys #[derive(PartialEq, Props)] struct NoKeysProps { data: std::collections::HashMap, } fn AntipatternNoKeys(cx: Scope) -> Element { // WRONG: Make sure to add keys! cx.render(rsx! { ul { cx.props.data.iter().map(|(k, v)| rsx!(li { "List item: {v}" })) } }); // RIGHT: Like this: cx.render(rsx! { ul { cx.props.data.iter().map(|(k, v)| rsx!(li { key: "{k}", "List item: {v}" })) } }) } /// Antipattern: Deeply nested fragments /// ------------------------------------ /// /// This particular antipattern is not necessarily an antipattern in other frameworks but does has a performance impact /// in Dioxus apps. Fragments don't mount a physical element to the DOM immediately, so Dioxus must recurse into its /// children to find a physical DOM node. This process is called "normalization". Other frameworks perform an agressive /// mutative normalization while Dioxus keeps your VNodes immutable. This means that deeply nested fragments make Dioxus /// perform unnecessary work. Prefer one or two levels of fragments / nested components until presenting a true DOM element. /// /// Only Component and Fragment nodes are susceptible to this issue. Dioxus mitigates this with components by providing /// an API for registering shared state without the ContextProvider pattern. fn AntipatternNestedFragments(cx: Scope<()>) -> Element { // Try to avoid heavily nesting fragments cx.render(rsx!( Fragment { Fragment { Fragment { Fragment { Fragment { div { "Finally have a real node!" } } } } } } )) } /// Antipattern: Using state after it's been updated /// ----------------------------------------------- /// /// This is an antipattern in other frameworks, but less so in Dioxus. However, it's important to highlight that use_state /// does *not* work the same way as it does in React. Rust provides explicit guards against mutating shared data - a huge /// problem in JavaScript land. With Rust and Dioxus, it's nearly impossible to misuse `use_state` - you simply can't /// accidentally modify the state you've received! /// /// However, calling set_state will *not* update the current version of state in the component. This should be easy to /// recognize from the function signature, but Dioxus will not update the "live" version of state. Calling `set_state` /// merely places a new value in the queue and schedules the component for a future update. fn AntipatternRelyingOnSetState(cx: Scope) -> Element { let (state, set_state) = use_state(&cx, || "Hello world").classic(); set_state("New state"); // This will return false! `state` will *still* be "Hello world" assert!(state == &"New state"); todo!() } /// Antipattern: Capitalization /// --------------------------- /// /// This antipattern is enforced to retain parity with other frameworks and provide useful IDE feedback, but is less /// critical than other potential misuses. In short: /// - Only raw elements may start with a lowercase character /// - All components must start with an uppercase character /// /// i.e.: the following component will be rejected when attempted to be used in the rsx! macro static antipattern_component: Component = |cx| todo!(); /// Antipattern: Misusing hooks /// --------------------------- /// /// This pattern is an unfortunate one where Dioxus replicates the same behavior as other frameworks. Dioxus supports /// "hooks" - i.e. "memory cells" that allow a value to be stored between renders. This allows other hooks to tap into /// a component's "memory" without explicitly adding all of its data to a struct definition. In Dioxus, hooks are allocated /// with a bump arena and then immediately sealed. /// /// This means that hooks may not be misused: /// - Called out of order /// - Called in a conditional /// - Called in loops or callbacks /// /// For the most part, Rust helps with rule #3 but does not save you from misusing rule #1 or #2. Dioxus will panic /// if hooks do not downcast to the same data between renders. This is validated by TypeId and, eventually, a custom key. #[derive(PartialEq, Props)] struct MisuedHooksProps { should_render_state: bool, } fn AntipatternMisusedHooks(cx: Scope) -> Element { if props.should_render_state { // do not place a hook in the conditional! // prefer to move it out of the conditional let (state, set_state) = use_state(&cx, || "hello world").classic(); cx.render(rsx!(div { "{state}" })) } else { cx.render(rsx!(div { "Not rendering state" })) } } /// Antipattern: Downcasting refs and panicking /// ------------------------------------------ /// /// Occasionally it's useful to get the ref of an element to handle it directly. Elements support downcasting to /// Dioxus' virtual element types as well as their true native counterparts. Downcasting to Dioxus' virtual elements /// will never panic, but downcasting to native elements will fail if on an unsupported platform. We recommend avoiding /// publishing hooks and components that deeply rely on controlling elements using their native `ref`, preferring to /// use their Dioxus Virtual Element counterpart instead. /// This particular code *will panic* due to the unwrap. Try to avoid these types of patterns. /// --------------------------------- /// TODO: Get this to compile properly /// let div_ref = use_node_ref(&cx); /// /// cx.render(rsx!{ /// div { ref: div_ref, class: "custom class", /// button { "click me to see my parent's class" /// onclick: move |_| if let Some(div_ref) = div_ref { /// panic!("Div class is {}", div_ref.to_native::().unwrap().class()) /// } /// } /// } /// }) static _example: Component = |cx| todo!(); /// Antipattern: publishing components and hooks with all features enabled /// ---------------------------------------------------------------------- /// /// The `dioxus` crate combines a bunch of useful utilities together (like the rsx! and html! macros, hooks, and more). /// However, when publishing your custom hook or component, we highly advise using only the `core` feature on the dioxus /// crate. This makes your crate compile faster, makes it more stable, and avoids bringing in incompatible libraries that /// might make it not compile on unsupported platforms. /// /// We don't have a code snippet for this, but just prefer to use this line: /// dioxus = { version = "*", features = ["core"]} /// instead of this one: /// dioxus = { version = "*", features = ["web", "desktop", "full"]} /// in your Cargo.toml /// /// This will only include the `core` dioxus crate which is relatively slim and fast to compile and avoids target-specific /// libraries. static __example: Component = |cx| todo!(); fn Example(cx: Scope) -> Element { cx.render(rsx! { AntipatternNoKeys { data: std::collections::HashMap::new() } AntipatternNestedFragments {} }) }