agentleFS
SageV1SageV1/makepad-skills/skills/robius-app-architecture/SKILL.md

robius-app-architecture

Skill1 starsChanged 10 months agoMITView on GitHub

CRITICAL: Use for Robius app architecture patterns. Triggers on: Tokio, async, submit_async_request, 异步, 架构, SignalToUI, Cx::post_action, worker task, app structure, MatchEvent, handle_startup

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when:

    line 20
  2. Distinctive:

    Separation of Concerns: Keep UI logic on the main thread, async operations in Tokio runtime

    line 311
  3. Distinctive:

    Request/Response Pattern: Use typed enums for requests and actions

    line 312
  4. Very specific:

    Lock-Free Updates: Use crossbeam::SegQueue for high-frequency background updates

    line 313
  5. Very specific:

    SignalToUI: Always call SignalToUI::set_ui_signal() after enqueueing updates

    line 314
  6. Distinctive:

    Cx::post_action(): Use for async task results that need action handling

    line 315

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Robius App Architecture Skill

Best practices for structuring Makepad applications based on the Robrix and Moly codebases - production applications built with Makepad and Robius framework.

Source codebases:

  • Robrix: Matrix chat client - complex sync/async with background subscriptions
  • Moly: AI chat application - cross-platform (native + WASM) with streaming APIs

Triggers

Use this skill when:

  • Building a Makepad application with async backend integration
  • Designing sync/async communication patterns in Makepad
  • Structuring a Robius-style application
  • Keywords: robrix, robius, makepad app structure, async makepad, tokio makepad

Production Patterns

For production-ready async patterns, see the _base/ directory:

PatternDescription
08-async-loadingAsync data loading with loading states
09-streaming-resultsIncremental results with SignalToUI
13-tokio-integrationFull tokio runtime integration

Core Architecture Pattern

┌─────────────────────────────────────────────────────────────┐
│                     UI Thread (Makepad)                     │
│  ┌─────────┐     ┌──────────┐     ┌──────────────────────┐ │
│  │   App   │────▶│ WidgetRef │────▶│ Widget Tree (View)  │ │
│  │ State   │     │    ui     │     │ Scope::with_data()  │ │
│  └────┬────┘     └──────────┘     └──────────────────────┘ │
│       │                                                     │
│       │ submit_async_request()                              │
│       ▼                                                     │
│  ┌─────────────────┐          ┌─────────────────────────┐  │
│  │ REQUEST_SENDER  │─────────▶│  Crossbeam SegQueue     │  │
│  │ (MPSC Channel)  │          │  (Lock-free Updates)    │  │
│  └─────────────────┘          └─────────────────────────┘  │
└───────────────────────────────────┬─────────────────────────┘
                                    │
                    SignalToUI::set_ui_signal()
                                    │
┌───────────────────────────────────┴─────────────────────────┐
│                   Tokio Runtime (Async)                      │
│  ┌──────────────────────────────────────────────────────┐   │
│  │           worker_task (Request Handler)               │   │
│  │  - Receives Request from UI                           │   │
│  │  - Spawns async tasks per request                     │   │
│  │  - Posts actions back via Cx::post_action()           │   │
│  └──────────────────────────────────────────────────────┘   │
│  ┌──────────────────────────────────────────────────────┐   │
│  │         Per-Item Subscriber Tasks                     │   │
│  │  - Listens to external data stream                    │   │
│  │  - Sends Update via crossbeam channel                 │   │
│  │  - Calls SignalToUI::set_ui_signal() to wake UI       │   │
│  └──────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────┘

App Structure

Top-Level App Definition

use makepad_widgets::*;

live_design! {
    use link::theme::*;
    use link::widgets::*;

    App = {{App}} {
        ui: <Root>{
            main_window = <Window> {
                window: {inner_size: vec2(1280, 800), title: "MyApp"},
                body = {
                    // Main content here
                }
            }
        }
    }
}

app_main!(App);

#[derive(Live)]
pub struct App {
    #[live] ui: WidgetRef,
    #[rust] app_state: AppState,
}

impl LiveRegister for App {
    fn live_register(cx: &mut Cx) {
        // Order matters: register base widgets first
        makepad_widgets::live_design(cx);
        // Then shared/common widgets
        crate::shared::live_design(cx);
        // Then feature modules
        crate::home::live_design(cx);
    }
}

impl LiveHook for App {
    fn after_new_from_doc(&mut self, cx: &mut Cx) {
        // One-time initialization after widget tree is created
    }
}

AppMain Implementation

impl AppMain for App {
    fn handle_event(&mut self, cx: &mut Cx, event: &Event) {
        // Forward to MatchEvent trait
        self.match_event(cx, event);

        // Pass AppState through widget tree via Scope
        let scope = &mut Scope::with_data(&mut self.app_state);
        self.ui.handle_event(cx, event, scope);
    }
}

Tokio Runtime Integration

Static Runtime Initialization

use std::sync::Mutex;
use tokio::sync::mpsc::{UnboundedReceiver, UnboundedSender};

static TOKIO_RUNTIME: Mutex<Option<tokio::runtime::Runtime>> = Mutex::new(None);
static REQUEST_SENDER: Mutex<Option<UnboundedSender<AppRequest>>> = Mutex::new(None);

pub fn start_async_runtime() -> Result<tokio::runtime::Handle> {
    let (request_sender, request_receiver) = tokio::sync::mpsc::unbounded_channel();

    let rt_handle = TOKIO_RUNTIME.lock().unwrap()
        .get_or_insert_with(|| {
            tokio::runtime::Runtime::new()
                .expect("Failed to create Tokio runtime")
        })
        .handle()
        .clone();

    // Store sender for UI thread to use
    *REQUEST_SENDER.lock().unwrap() = Some(request_sender);

    // Spawn the main worker task
    rt_handle.spawn(worker_task(request_receiver));

    Ok(rt_handle)
}

Request Submission Pattern

pub enum AppRequest {
    FetchData { id: String },
    SendMessage { content: String },
    // ... other request types
}

/// Submit a request from UI thread to async runtime
pub fn submit_async_request(req: AppRequest) {
    if let Some(sender) = REQUEST_SENDER.lock().unwrap().as_ref() {
        sender.send(req)
            .expect("BUG: worker task receiver has died!");
    }
}

Worker Task Pattern

async fn worker_task(mut request_receiver: UnboundedReceiver<AppRequest>) -> Result<()> {
    while let Some(request) = request_receiver.recv().await {
        match request {
            AppRequest::FetchData { id } => {
                // Spawn a new task for each request
                let _task = tokio::spawn(async move {
                    let result = fetch_data(&id).await;
                    // Post result back to UI thread
                    Cx::post_action(DataFetchedAction { id, result });
                });
            }
            AppRequest::SendMessage { content } => {
                let _task = tokio::spawn(async move {
                    match send_message(&content).await {
                        Ok(()) => Cx::post_action(MessageSentAction::Success),
                        Err(e) => Cx::post_action(MessageSentAction::Failed(e)),
                    }
                });
            }
        }
    }
    Ok(())
}

Lock-Free Update Queue Pattern

For high-frequency updates from background tasks:

use crossbeam_queue::SegQueue;
use makepad_widgets::SignalToUI;

pub enum DataUpdate {
    NewItem { item: Item },
    ItemChanged { id: String, changes: Changes },
    Status { message: String },
}

static PENDING_UPDATES: SegQueue<DataUpdate> = SegQueue::new();

/// Called from background async tasks
pub fn enqueue_update(update: DataUpdate) {
    PENDING_UPDATES.push(update);
    SignalToUI::set_ui_signal();  // Wake UI thread
}

// In widget's handle_event:
impl Widget for MyWidget {
    fn handle_event(&mut self, cx: &mut Cx, event: &Event, scope: &mut Scope) {
        // Poll for updates on Signal events
        if let Event::Signal = event {
            while let Some(update) = PENDING_UPDATES.pop() {
                match update {
                    DataUpdate::NewItem { item } => {
                        self.items.push(item);
                        self.redraw(cx);
                    }
                    // ... handle other updates
                }
            }
        }
    }
}

Startup Sequence

impl MatchEvent for App {
    fn handle_startup(&mut self, cx: &mut Cx) {
        // 1. Initialize logging
        let _ = tracing_subscriber::fmt::try_init();

        // 2. Initialize app data directory
        let _app_data_dir = crate::app_data_dir();

        // 3. Load persisted state
        if let Err(e) = persistence::load_window_state(
            self.ui.window(ids!(main_window)), cx
        ) {
            error!("Failed to load window state: {}", e);
        }

        // 4. Update UI based on loaded state
        self.update_ui_visibility(cx);

        // 5. Start async runtime
        let _rt_handle = crate::start_async_runtime().unwrap();
    }
}

Shutdown Sequence

impl AppMain for App {
    fn handle_event(&mut self, cx: &mut Cx, event: &Event) {
        if let Event::Shutdown = event {
            // Save window geometry
            let window_ref = self.ui.window(ids!(main_window));
            if let Err(e) = persistence::save_window_state(window_ref, cx) {
                error!("Failed to save window state: {e}");
            }

            // Save app state
            if let Some(user_id) = current_user_id() {
                if let Err(e) = persistence::save_app_state(
                    self.app_state.clone(), user_id
                ) {
                    error!("Failed to save app state: {e}");
                }
            }
        }
        // ... rest of event handling
    }
}

Best Practices

  1. Separation of Concerns: Keep UI logic on the main thread, async operations in Tokio runtime
  2. Request/Response Pattern: Use typed enums for requests and actions
  3. Lock-Free Updates: Use crossbeam::SegQueue for high-frequency background updates
  4. SignalToUI: Always call SignalToUI::set_ui_signal() after enqueueing updates
  5. Cx::post_action(): Use for async task results that need action handling
  6. Scope::with_data(): Pass shared state through widget tree
  7. Module Registration Order: Register base widgets before dependent modules in live_register()

Reference Files

  • references/tokio-integration.md - Detailed Tokio runtime patterns (Robrix)
  • references/channel-patterns.md - Channel communication patterns (Robrix)
  • references/moly-async-patterns.md - Cross-platform async patterns (Moly)
    • PlatformSend trait for native/WASM compatibility
    • UiRunner for async defer operations
    • AbortOnDropHandle for task cancellation
    • ThreadToken for non-Send types on WASM
    • spawn() platform-agnostic function

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.