The IR called them react-message and django-message, so a FastAPI channel had
to declare a DjangoMessage. They are client-message and server-message now,
and the direction words hold wherever a channel is declared: Params /
ClientMessage / ServerMessage, with mizan-core deriving <Pascal>Params and
friends so no backend names a type itself. Django's ReactChannel and
FastAPI's ReactChannel are both Channel.
mizan-fastapi never registered a channels extension, so build_ir() emitted no
channel at all and every payload type was invisible to codegen. It registers
one now. RegistryExtension is an ABC requiring all(), which is what the IR
reads — an extension that cannot enumerate its registrations no longer exists.
The gate that should have caught the rename could not: tests/afi registered no
channel because mizan-rust had no channel registry to register one in, so a
five-package rename of the wire contract passed byte-parity without a channel
byte crossing it. mizan-rust grows ChannelSlotKind, a CHANNELS slice, a
#[mizan::channel] macro, and KDL emission whose wire_to_pascal matches Python's
split; the AFI fixture now carries a channel with every slot and one with a
single slot, so all three backends prove the contract byte for byte.
MizanChannel held three Option<String> beside three has_*() predicates and
unwrapped them with defaults; it holds an ordered slot vector, so an absent
slot is absent rather than defaulted. The channels target emitted a React
hooks file that a stage1-only consumer could not compile — react emits that
now. The codegen's parity tests byte-compared emitted source against baselines
without ever compiling it: they compile the generated crate and run its tests,
import the generated Python package and call every method, and typecheck each
TypeScript target against a consumer.
Also fixed at source: app_visitor printed its import diagnostic to stdout, the
stream export_mizan_ir writes KDL to, so a failed import silently corrupted the
IR; the apps root was hardcoded to "apps"; _default_literal crashed build_ir on
any non-JSON-serializable field default; Django and mizan-core derived Pascal
names two different ways, disagreeing on every dotted channel name.
ir.py builds a document and renders templates/ir/document.kdl.j2 rather than
appending KDL strings with hand-tracked indentation, and named types resolve to
a fixed point — a model reachable only through a union branch was referenced by
a ref that no type block ever defined.
The rest is the write-gate's own classifiers run over the standing tree:
relative imports, silent swallows, Protocol contracts that should be ABCs,
emitters hand-rendering target source, catch-all arms over closed enums, and
comments narrating the project rather than the code.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The Mizan codegen substrate moves off JavaScript template-literal emission
onto a compiled Rust binary that consumes the same OpenAPI + x-mizan-* IR
the JS substrate consumed. Three structural wins fall out of one move:
1. Moat closes. The codegen logic (how `affects` becomes auto-invalidation,
how named contexts collapse onto bundled fetches, how the registry-to-
Provider mapping is shaped) ships compiled instead of as source bytes
in every consumer's node_modules.
2. Pattern F (lines.push append-walls) becomes structurally unauthorable.
The emit substrate is askama templates in templates/<target>/*.j2 —
actual target-language files with {{ ... }} substitution markers,
syntax-highlighted natively, type-checked against the render context
structs at compile time. The Rust emit modules build typed render
contexts and call .render(); no string-builder surface exists.
3. OpenAPI `default`-bearing fields now emit as non-optional in TS / Python
/ Rust — the server always populates them, so consumer code reads them
without nullable checks. Surfaced by Blazr's typecheck on regeneration.
Layout:
frontends/mizan-rust/ — Rust port of @mizan/base; #[cfg(feature="pyo3")]
exposes PyMizanClient for the Python target.
protocol/mizan-codegen/ — codegen binary source + askama templates.
protocol/mizan-generate/ — npm-package shim. bin/launcher.mjs dispatches
to the platform-appropriate prebuilt binary.
Old generator/ JS tree deleted.
tests/rust/ — wire-parity drivers. drive_kernel exercises
raw client.call() / fetch_context(); drive_emitted
exercises the typed crate the codegen emits.
tests/afi/afi_codegen_app.py — codegen entrypoint module (imports + registers).
backends/mizan-fastapi/.../schema.py — adds outputNullable so the Rust
codegen can wrap T | None responses in Option<T>.
Verification:
- 20 mizan-codegen tests green (IR deserialization, byte-equivalent
parity vs JS baseline for stage1/rust/python/react/vue/svelte,
structural test for channels).
- tests/rust/run_wire_parity.py — 12/12 probes green via the Rust binary
driving the FastAPI fixture end-to-end.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
