- In-tai.py TUI framework (panels, compositor, z-layers, mouse, focus, VT screen, pty sub-terminals, widgets, animation) plus /demo, --demo, --demo-shot, tui.md docs, and research/tui-framework-design archive - Focus safety: Tab/Shift+Tab always cycle, auto-focus prefers non-terminals, dialogs restore previous focus on close - Native Consts/Install/Sandbox/Opencode/SysInfo classes in tai.py; module functions delegate so live patches apply; fixes 19 failures - Retargeted 4 test patches to the real tai.* seams (assertions unchanged) - Deleted tai_pkg/; tai.py is standalone stdlib-only (bare copy imports) - AGENTS.md/README.md updated to the standalone architecture - test_tai.py 320/320 OK, test_seal.py 16/16 OK
3.0 KiB
3.0 KiB
Events and Messages - Textual (official guide)
Metadata
| Field | Value |
|---|---|
| Institute | Molodetz Institute — Research & Retrieval Division |
| Subject | tui-framework-design |
| Document type | Source Dossier |
| Tier | 1 |
| Source title | Events and Messages - Textual |
| Source URL | https://textual.textualize.io/guide/events/ |
| Source publisher / author | Textualize |
| Source type | vendor docs |
| Source date (published) | Unknown — not stated |
| Date researched | 2026-10-10 |
| Researched by | tai (agent), model: Muse Spark, on behalf of retoor |
| Research method | direct page fetch |
| Confidence | High |
| Status | Active |
Summary
Textual's official guide to its event system: every App and Widget owns a message queue drained by its own asyncio Task; events are framework-reserved messages; handlers attach via on_ naming or the @on decorator; messages with bubble=True propagate to the parent after processing; stop(), prevent_default(), and the prevent() context manager control propagation and default handling.
Key findings
- Every App and Widget has its own message queue plus an asyncio Task started at mount that dispatches queued messages to handlers in order.
- Handler naming scheme:
on_+ snake_case namespace (outer widget class, if the message is nested) + snake_case message class name; each message class exposeshandler_name. - Messages carry a
bubbleattribute; when true the message is sent to the widget's parent after processing, up to App (the DOM root, where bubbling ends). stop()on a message halts bubbling;prevent_default()stops base-class handlers from running (Textual otherwise calls base-class handlers automatically, e.g.Paddle.on_keythenStatic.on_keythenWidget.on_key).- Custom messages are Message subclasses, conventionally nested inside the sending widget (creates the handler namespace, avoids name clashes); sent via
post_message(), commonly to self so they bubble. prevent(MessageType)context manager temporarily suppresses a message type while updating children programmatically.
Pros / strengths
- Authoritative (vendor's own docs) with runnable code for every concept; ordering and bubbling guarantees stated plainly.
- The bubble/stop/prevent trio is a small, complete vocabulary for event flow control.
Cons / limitations / caveats
- asyncio-centric (per-widget tasks); tai's UI loop is a synchronous select loop, so the queue-per-widget half does not transfer, only the dispatch/bubble half.
- No capture phase (root-to-target); bubbling only. No performance numbers.
Relevance to tai
- Direct template for the tai TUI event model: an Event base with a
bubbleflag,on_<name>handler dispatch walking the MRO,stop()halting the parent walk, modal dialogs consuming keys before they reach layers below. - tai will dispatch synchronously in its tick loop instead of per-widget tasks; ordering guarantee ("one event fully handled before the next") is kept trivially that way.
Changelog
| Date | Change |
|---|---|
| 2026-10-10 | Initial dossier created. |