1# llimphi-jaula
2
3The **app-side** shell: runs the Llimphi loop inside a WASM sandbox and emits an
4[`EscenaWire`](../llimphi-wire-escena/README.md) instead of rasterizing.
5
6It is the twin of [`llimphi-wawa`](../llimphi-wawa/README.md), on the other side
7of the boundary: that one receives and executes, this one composes and ships.
8
9Phase W4 of [`03_ukupacha/wawa/PLAN-LLIMPHI.md`](../../../03_ukupacha/wawa/PLAN-LLIMPHI.md).
10
11```
12View<Msg> → mount → layout → paint → EscenaWire → postcard → …
13…………………………… WASM boundary ……………………………
14                                   … → native shaping → rasterize
15```
16
17## The thing to understand
18
19Both shells run the **same `llimphi-compositor`** — same mount, same taffy
20layout, same 4,000-line `render.rs`. The only difference is the `Lienzo` alias
21and, with it, the `Typesetter`.
22
23That is the W4 result worth saying out loud: **there is no app compositor and
24kernel compositor.** A widget written once works on Linux, inside the kernel, and
25inside a sandbox.
26
27## Two things this one does that the kernel shell doesn't
28
29### It asks for measurements
30
31taffy cannot size a label without knowing how big it is, and measuring means
32shaping — precisely what must not happen inside the sandbox: the interpreter's
3330.3x would eat 72% of the frame (W4.1). So the measurer is an **argument**:
34
35```rust
36let mut jaula = Jaula::new(320.0, 72.0, |p: &Peticion| {
37    // On wawa: a `sys_medir_texto` crossing to parley on the native side.
38    // In a test: a real parley `Typesetter`.
39    medir_del_otro_lado(p)
40});
41let montado = jaula.fotograma(vista());
42let bytes = postcard::to_allocvec(jaula.escena())?;
43```
44
45With no measurer it does not return zero — a silent zero would collapse the whole
46layout and the screen would come out blank while "working". It estimates coarsely
47**and counts**: a non-zero `jaula.medidas_estimadas()` means that layout is
48approximate.
49
50On wawa that measurer already exists — it is [`medidor::wawa`](src/medidor.rs),
51wrapping the `sys_medir_texto` syscall:
52
53```rust
54let mut jaula = Jaula::new(ancho, alto, llimphi_jaula::medidor::wawa);
55```
56
57`sys_medir_texto` is **not permission-gated**: measuring text is not a privilege,
58it is arithmetic over a font, and every app that draws needs it. What it does
59carry is a quota, because the kernel pays for the shaping outside the app's
60`fuel`. Off wasm the measurer returns `None` rather than an invented number.
61
62### It keeps the tree
63
64The `Mounted` never crosses the boundary, so hit-testing resolves here: the
65kernel sends a coordinate and the app answers what was hit.
66
67```rust
68if let Some(msg) = jaula.click(&montado, x, y) {
69    modelo = App::update(modelo, msg);
70}
71```
72
73The opposite of the kernel shell, which **cannot** hit-test precisely because the
74tree stayed here.
75
76## Why it lives outside the workspace
77
78It enables `llimphi-text/wire`, a backend mutually exclusive with `vello` (what
79the rest of the repo uses by default). Cargo's feature unification would turn on
80both and trip the `compile_error!` `llimphi-text` carries to catch exactly that —
81the same trap `llimphi-wawa` already hit.
82
83Both being outside isn't friction: it's the design. They are two processes on two
84sides of a boundary.
85
86```bash
87cargo test  --manifest-path 02_ruway/llimphi/llimphi-jaula/Cargo.toml \
88  --no-default-features --features std
89cargo check --manifest-path 02_ruway/llimphi/llimphi-jaula/Cargo.toml \
90  --target wasm32-unknown-unknown --no-default-features --features libm
91```
92
93## The fixture
94
95Since encoder and replayer cannot be compiled together, the test seam is the same
96as the production one: **bytes**.
97
98```bash
99cargo run --example emitir_fixture --no-default-features --features std
100```
101
102It composes a view with the real compositor and drops the bytes in
103`../llimphi-wawa/pruebas/fixture_jaula.postcard`, where the
104`la_escena_que_emitio_el_compositor_llega_a_pixeles` test **and the kernel's boot
105path** pick them up. If anyone changes the wire format on one side only, those
106two will say so.