Skip to content

affinescript-deno-test

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History
 
 

affinescript-deno-test

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
example
example
 
 
lib
lib
 
 
.gitignore
.gitignore
 
 
README.adoc
README.adoc
 
 
cli.ts
cli.ts
 
 
deno.json
deno.json
 
 
deno.lock
deno.lock
 
 
mod.ts
mod.ts
 
 

README.adoc

affinescript-deno-test

Table of Contents

A Deno test-framework integration for AffineScript. Compiles .affine test files to WebAssembly via the standard affinescript compiler, loads them through the existing @hyperpolymath/affine-js bridge, and registers each as a Deno.test() case so deno test reports pass/fail in the usual way.

Part of the developer-ecosystem monorepo. Sibling tools in the AffineScript ecosystem live alongside this component: affinescript/ (the compiler — actually mirrored from upstream), affinescriptiser/, affinescript-vite/, rattlescript/.

Status

v0.2.0. Multi-test-per-file harness. The "one test per file" MVP constraint was resolved same-day via an upstream compiler change (see Constraints and planned follow-ups).

Quick start

# Compile + smoke-test the bundled example:
deno task smoke:test
# → hello ... ok (1 passed, 0 failed)

Writing a test

Each test case is a top-level pub fn test_<name>() → Bool. Return true to pass, false to fail. One .affine file may contain as many test cases as you like. Non-pub helpers stay internal to the module. Filename ends in _test.affine or .test.affine.

// example/hello_test.affine
// SPDX-License-Identifier: CC-BY-SA-4.0

pub fn test_addition() -> Bool {
  let result = 2 + 3;
  result == 5
}

pub fn test_identity() -> Bool {
  let x = 42;
  x == 42
}

fn helper_not_exported() -> Int {
  99
}

pub fn test_uses_helper() -> Bool {
  helper_not_exported() == 99
}

The harness reports each test as <filestem> / <casename> — e.g. the file above produces hello / addition, hello / identity, hello / uses_helper under deno test.

Running tests

The public API is runAll(root) in mod.ts:

// driver.ts
import { runAll } from "@hyperpolymath/affinescript-deno-test";

await runAll("./tests");

Invoke with deno test:

deno test --allow-read --allow-run --allow-env driver.ts

The runner:

  1. Walks root for _test.affine / .test.affine files

  2. Compiles each via affinescript compile …​ -o …​.wasm

  3. Loads the WASM (with a minimal WASI fd_write stub — AffineScript codegen imports this unconditionally)

  4. Registers a Deno.test() case named after the file (stripped of the _test suffix) that invokes main and checks the Bool return

CLI (smoke-check)

cli.ts is a thin smoke-check that prints discovery + compile results without running assertions. Useful for CI pipelines that want to fail fast on compile errors before invoking deno test.

deno run --allow-read --allow-run --allow-env cli.ts ./tests
# Discovered 3 test file(s):
#   ✓ tests/foo_test.affine → tests/foo_test.wasm
#   ✓ tests/bar_test.affine → tests/bar_test.wasm
#   ...

Configuration

  • AFFINESCRIPT_BIN env var — absolute path to the affinescript compiler. Default: /var/mnt/eclipse/repos/developer-ecosystem/nextgen-languages/affinescript/_build/install/default/bin/affinescript (the local dev build). Override when the compiler is installed elsewhere or on $PATH.

Constraints and planned follow-ups

Two constraints were resolved same-day (one for v0.2.0, plus the three codegen bugs below). Two remain, both resolvable with small upstream changes.

RESOLVED (v0.2.0): one test per file

The AffineScript codegen previously hardcoded an exportable-name allowlist to ["main"; "init_state"; "step_state"; "get_state"; "mission_active"] with no pub fn / @export keyword — so only those five names appeared in the WASM export list, no matter the source.

The fix (AffineScript commit ce324fa) makes both codegen.ml and codegen_gc.ml honour the AST’s fd_vis = Public field in addition to the game-hook allowlist. The allowlist is retained verbatim for backward compatibility (pre-pub programs and the IDApTIK CharacterSelect bridge continue to work unchanged).

The harness now uses the expected multi-test convention: any top-level pub fn test_<name>() → Bool becomes a separate Deno.test() case.

WASI fd_write imported unconditionally

Every AffineScript WASM output imports wasi_snapshot_preview1.fd_write even when the program uses no IO. The harness provides a minimal no-op stub so tests instantiate cleanly.

Planned follow-up: upstream compiler change so pure programs (no effect IO) skip the WASI import entirely. Until then, the stub is correct and harmless.

@hyperpolymath/affine-js only supplies env imports

The existing bridge’s AffineModule.fromBytes merges custom imports under the env module key only. For WASM modules that import from a non-env module (e.g. wasi_snapshot_preview1), the harness bypasses AffineModule and uses raw WebAssembly.instantiate plus its own WASI stub. The Bool return is unmarshalled by hand (AffineScript compiles Bool → i32, 0/1).

Planned follow-up: extend affine-js to accept a namespacedImports: Record<string, Record<string, ImportValue>> option, then unify both paths behind AffineModule.

RESOLVED (2026-04-19): codegen bugs blocking richer test idioms

Both bugs discovered while writing the double-track-browser lifecycle test pilot were fixed upstream in developer-ecosystem/nextgen-languages/ affinescript/lib/codegen.ml the same day. Regression coverage lives in example/codegen_regression_test.affine and the full double-track-browser extension_lifecycle_test.affine (10/10 green).

  • Enum-in-match stack imbalance — PatCon-with-args pushed the tag-test boolean onto the stack twice (LocalTee + a trailing LocalGet on the match-result local), so any arm body that produced an i32 blew up WASM validation with "expected 1 elements on the stack for fallthru, found 2". Fixed by removing the save/restore pair; the field-binding code between them is stack-neutral by construction.

  • Bare zero-arity variant as an expression — Initialised (no parens) fell through to the ExprVar lookup and failed with UnboundVariable unless written as State::Initialised. Fixed by falling back to ctx.variant_tags when lookup_local misses, mirroring the existing ExprCall-with-variant-tag branch.

  • Non-first struct-field read from parameter/call-result — the per-variable field_layouts map was only populated for let-bound ExprRecord literals, so every other binding path defaulted offsets to 0 and .field_1_or_later read the tag byte. Fixed by registering struct layouts globally from TopType(TyStruct) and propagating them to (a) function parameters via p_ty, (b) call-result lets via a new fn_ret_structs map, (c) let-bindings with an explicit type annotation, and (d) let-bindings whose RHS is another tracked variable.

The harness now supports idiomatic enum-plus-match application-state test suites without fallback tagged-struct workarounds.

Why AffineScript rather than ReScript?

ReScript is the estate’s default TypeScript replacement and has a mature Deno integration (see fireflag). The AffineScript path exists because AffineScript has stronger semantics for:

  • affine / quantitative types (resource lifecycle tracking)

  • algebraic effects (explicit side-effect boundaries)

  • row polymorphism (extensible records)

Test suites that want to demonstrate these properties — e.g. "this resource is consumed exactly once", "this protocol sequence is honoured" — are a natural fit for AffineScript. Plain behavioural tests can continue in ReScript; there’s no migration pressure either direction.

The trade-off today: AffineScript’s compiler matures while ReScript’s is mature, so this harness carries the MVP constraints above. As the compiler gains pub fn / effect-conditional WASI imports / better bridge ergonomics, the harness sheds them.

License

MPL-2.0. See the repository root LICENSE file. Legal fallback: MPL-2.0 (automatic, per the hyperpolymath License Policy Rule 2 in ~/.claude/CLAUDE.md).