@genroc/eval-node 0.0.0-edge.e4a77190 → 0.0.0-edge.e50224ac

package/README.md

@@ -1,337 +1,145 @@
 # @genroc/eval-node
 
-Script tasks for Node. Two halves with different jobs:
+TypeScript script tasks for [genroc](https://genroc.org). The package provides two binaries:
 
-* **the bundler** (`genroc-import`) — an author-time resolver genctl runs on every apply and
-  every edit, turning `$import: ./x.ts` into a bundled, typechecked string
-* **the worker** (`genroc-eval-node`) — claims parked `external` script tasks off genroc's queue
-  and evaluates each in its own realm
+* **`genroc-import`**: the bundler. `genctl` runs it on every `apply` and `generate` to turn
+  `$import: ./script.ts` into one typechecked, self-contained module string.
+* **`genroc-eval-node`**: the worker. It claims script tasks from genroc's queue and runs each
+  one in a fresh worker thread.
 
-## Setting up a project
+Requires **Node 24+**. Full guide:
+[TypeScript evaluation](https://genroc.org/guides/process-definition/typescript-evaluation/).
 
-    npm i -D @genroc/eval-node
-
-Then register the resolver in a `.genroc` beside your definitions — a dotfile, so it does not
-read as another process definition. Discovery walks up from the source file, so nothing depends
-on the cwd:
-
-    resolvers:
-      import:
-        phase: code
-        ext: .ts
-        command: [npx, genroc-import]
-
-`genctl apply` and `genctl types` now resolve `$import` directives. Typechecking is the
-resolver's exit code, so a stored definition cannot hold code that failed to typecheck.
-
-## Building
-
-    pnpm run build      # tsc -p tsconfig.build.json -> dist/
-
-The published package is JavaScript, not TypeScript. **Node refuses to strip types for files
-under `node_modules`**, so shipping `.ts` works from a checkout and fails for every consumer —
-which is why `prepublishOnly` builds and CI installs the packed tarball rather than only running
-from source. The source keeps its `.ts` import extensions; `rewriteRelativeImportExtensions`
-turns them into `.js` on the way out.
-
-## Running the worker
-
-    docker run -e GENROC_SERVER=http://host:8448 ghcr.io/genroc/eval-node:preview
-
-or locally, if you already have Node:
-
-    GENROC_SERVER=http://localhost:8448 npx genroc-eval-node
-
-Add `GENROC_TOKEN` once the server has auth on; `worker` is the only permission it needs.
-
-**The bundler cannot be the image.** It runs per edit and must resolve your project's own
-`node_modules`, so a container start per invocation would break the editor loop.
-
-Design record: [specs/external-task-queue.md](../specs/external-task-queue.md) for the queue,
-[specs/script-tasks.md](../specs/script-tasks.md) for the runtime.
+## Quick start
 
-    GENROC_SERVER=http://localhost:8448 node eval-node/worker.ts
+    genctl init --eval-node
 
-Against a server started with `--auth token`, mint a scoped credential first:
+This scaffolds a project with everything set up: the package, a `.genroc` that registers the
+resolver, a `tsconfig.json`, a reusable `script-node` child process, and a `compose.yaml` that
+runs the worker.
 
-    TOKEN=$(genctl token create --perms worker --label evaluator -q)
-    GENROC_SERVER=http://localhost:8448 GENROC_TOKEN=$TOKEN node eval-node/worker.ts
+## Adding it to an existing project
 
-`worker` reaches the four queue verbs and `GET /api/objects/{ref}` — enough to claim, fetch an
-externalized input and answer, and nothing else. This is the credential most likely to end up on
-a machine you trust least, so it is worth scoping rather than reusing an admin token: a leaked
-one cannot read a definition, list an instance, or mint another token.
-
-A credential problem **exits the worker** rather than being retried. Polling through a 401 looks
-like a healthy worker that never picks anything up, which is the worst shape to debug.
-
-It needs **Node 24 or newer**: the sources are TypeScript and are run as-is, by Node's own
-type stripping.
-
-| env | default | |
-|---|---|---|
-| `GENROC_SERVER` | `http://localhost:8448` | where to claim from |
-| `GENROC_TOKEN` | *(none)* | credential, when the server runs with `--auth token`. Needs the **`worker`** permission and nothing more |
-| `WORKER_ID` | `evaluator-<pid>` | the claim holder; renewals are scoped to it |
-| `CONCURRENCY` | `4` | how many scripts run at once — **this worker's** decision |
-| `LEASE_MS` | `30000` | visibility timeout; renewed at a third of it while working |
-| `POLL_MS` | `250` | idle poll interval; a non-empty claim polls again immediately |
-| `PROCESS` / `TASK` | unset | claim only this process / task id |
-
-**The worker calls genroc, not the other way round.** That is the whole reason this is a
-queue worker rather than the HTTP sidecar it used to be:
-
-- **Concurrency is the worker's to set.** Under the old `fetch` shape genroc decided how many
-  scripts ran at once (`--max-concurrent`, default 200) and the evaluator accepted every one.
-  A backlog is now a queue, not 200 threads fighting over a core.
-- **No engine worker is held.** A `fetch` occupies one of genroc's advance slots for its whole
-  duration, so a slow evaluator starved unrelated tasks. An `external` task parks.
-- **The connection direction inverts.** `/eval` was unauthenticated arbitrary code execution
-  that genroc had to be able to reach. A worker only needs outbound access, so it can live
-  anywhere — behind NAT, in another trust zone.
-
-The cost of that inversion is that genroc cannot reach a running worker, so anything it needs
-to say has to ride the renewal the worker already makes. **Renewing is therefore mandatory,
-not an optimisation** — a worker that stops renewing is indistinguishable from one that died.
-The response answers per token:
-
-| list | what it means | what this worker does |
-|---|---|---|
-| `renewed` | still yours | carry on |
-| `lost` | already someone else's | stop; do **not** release — that would bump the new holder's claim |
-| `cancelled` | still yours, nobody wants it | abort the script and release the claim |
-
-`cancelled` is how an operator's `genctl cancel` reaches work already running: the evaluation
-is aborted through an `AbortSignal`, its realm is terminated, and no outcome is submitted —
-the process is stopping, so there is nothing left to answer. `renew_before_ms` on the claim
-says how long the worker may wait before renewing again, so the interval is a value it reads
-rather than one it guesses.
-
-## The task input
+    npm i -D @genroc/eval-node
 
-The `input` of the external task IS the evaluation request:
+Register the resolver in the `.genroc` beside your definitions:
 
-```jsonc
-{
-  "code": "export default (input) => ({ fee: input.amount * 0.1 });",  // required — a module
-  "input": { "amount": 250 },                                         // optional — its argument
-  "timeout_ms": 5000                                                  // optional — default 5000
-}
+```yaml
+resolvers:
+  - name: import
+    phase: typed
+    ext: [.ts]
+    command: [npx, genroc-import]
+    types:
+      Input: task.action.input.input
+      Output: task.action.result
 ```
 
-`code` is an **ES module**, and the evaluator imports it and calls its **default export** with
-`input`. The value it returns reaches genroc; the function may be async, and so may the module's
-top level. A module exporting anything else — or nothing — is a `compile_error`: there is
-nothing to call, and so is one that will not parse; a throw from its top level is a `threw`
-like any other. `import` of a node builtin resolves as it does anywhere else, and that is what
-a bundled builtin lands on.
-
-## The answer — the failure kind IS the error code
-
-genroc's `on_error` matches error codes, and an external task's `raises` declares the closed
-set a worker may send. So the classification lives in the **code**, where a rule can match it,
-rather than in a body a `switch` has to read.
-
-| outcome | the worker submits | |
-|---|---|---|
-| the script returned | `result: <the return value>` | `return;` sends nothing, which genroc reads as `null` |
-| the script faulted | `error: {code, message, data}` | `code` is the kind, below |
-| the **evaluator** faulted | *nothing* — it releases the claim | the task returns to the queue for another worker |
-
-The five kinds, every one of them **permanent** — a retry re-runs the same code on the same
-input and fails identically:
-
-`compile_error` · `threw` · `timeout` · `nonserializable` · `exited`
-
-`data` carries `{name, stack?}`: `name` is what a script sets to tell one refusal from another
-(`e.name = 'LimitExceeded'`), and `stack` is renumbered to the lines the author wrote and
-trimmed to 2 KiB.
-
-**The retryable class has no code on purpose.** A runner that faults releases its claim, which
-is how a queue spells "try this somewhere else" — it puts the task in front of a different
-worker instead of spending the definition's `on_error` budget on this one's bad day. There is
-no retry policy to write.
-
-## The genroc side
+Then reference a script from a task. The simplest way is to spread the `script-node` child that
+`genctl init --eval-node` generates:
 
 ```yaml
-- id: price
+- id: greet
   action:
-    type: external
+    type: child
+    <<: "$process: ./script-node.genroc.yaml"
     input:
-      code: |
-        export default function (input) {
-          if (input.amount > 100) {
-            const e = new Error('amount over the limit');
-            e.name = 'LimitExceeded';
-            throw e;
-          }
-          return { fee: input.amount * 0.1 };
-        }
-      input: "$: input"
-    result_schema: { type: object, properties: { fee: { type: number } }, required: [fee] }
-    raises:
-      threw:           { $ref: "#/$defs/script_error" }
-      timeout:         { $ref: "#/$defs/script_error" }
-      compile_error:   { $ref: "#/$defs/script_error" }
-      nonserializable: { $ref: "#/$defs/script_error" }
-      exited:          { $ref: "#/$defs/script_error" }
-  timeout: 30s
-  on_error:
-    - code: [threw]
-      goto: $script_failed
-    - code: [compile_error, nonserializable, exited]
-      panic: { code: script_broken, message: "the script did not run: ${error.code}" }
-  switch: [{ goto: end }]
-
-- id: script_failed
-  switch:
-    - case: 'last_error.data.name == "LimitExceeded"'
-      raise: { code: limit_exceeded, message: "the script rejected the amount" }
-    - raise: { code: script_failed, message: "the script failed" }
-```
-
-**A script cannot name a genroc error code.** `raise`/`panic` codes are literals, never
-expressions, so the mapping from a thrown error to an authored code is this `switch` — one
-task the definition owns, reading `error.data`. That is the whole error protocol: the worker
-classifies into a code, the definition names the outcome.
-
-Set the task `timeout` **above** `timeout_ms`. If the task's deadline fires first the code is
-`external.timeout`, which is in `errcode.Unknowable()` — permanently unretryable on an
-`only_once` task, and indistinguishable from no evaluator running at all.
-
-`only_once` is worth considering if your scripts have side effects: without it a worker that
-dies mid-script has its task re-claimed and the script runs again. With it the task is never
-handed out twice, and the instance gets a catchable `external.lost` instead.
-
-## `${` must be escaped as `$${` — when the code is inline
-
-An external task's `input` is a Shape, so `${…}` is genroc's interpolation marker and a JS
-template literal inside `code` is read by genroc rather than passed through. Write `` `<$${x}>` ``.
-A leading `$:` on the code string needs `$$:` for the same reason. See
-[specs/typed-values.md](../specs/typed-values.md). Moving the code into a `.ts` file
-removes this entirely — see the next section.
-
-## `import.ts` — the author-time half
-
-`import.ts` is the **code-phase resolver** genctl runs before a definition is applied. It
-never touches the queue and the worker never runs it; the two halves share this package only
-because they share a calling convention, which is exactly the coupling that breaks silently
-if they version apart.
-
-Register it in the project's `.genroc`:
-
-```yaml
-resolvers:
-  import: { phase: code, ext: .ts, command: [node, ../eval-node/import.ts] }
+      code: "$import: ./greet.ts"
+      input: { who: "$: input.who" }
+    result_schema:
+      type: object
+      properties: { greeting: { type: string } }
+      required: [greeting]
 ```
 
-then write the script as a module and name it from the definition:
+## Writing a script
 
-```yaml
-body:
-  code: "$import: ./fee.ts"
-  input: "$: input"
-```
+A script is an ES module. Its default export receives `input`, and its return value becomes the
+task result. The function may be async.
 
 ```ts
-import type { Input, Output } from "./fee.genroc";
+import type { Input, Output } from "./greet.genroc";
 
-export default async function (input: Input): Promise<Output> {
-  return { fee: input.amount * 0.1 };
+export default function (input: Input): Output {
+  return { greeting: `hello, ${input.who}` };
 }
 ```
 
-`genctl types -f process.yaml` writes `fee.genroc.d.ts` beside the script — named for the
-**script's path**, not the task, so renaming a task cannot break the import line. `Input` is
-the inferred type of what the definition passes; `Output` is what it declares
-(`responses.200`, or `result_schema` on a child). `genctl apply` regenerates them, runs
-`tsc --noEmit`, and bundles — so **a type error is a failed apply**, and a stored definition
-cannot hold code that failed to typecheck.
+* `genctl generate` writes `greet.genroc.d.ts` beside the script. `Input` and `Output` come from
+  the definition, so a mismatch fails the typecheck.
+* `genctl apply` typechecks and bundles. **A type error or a missing default export fails the
+  apply.**
+* Imported files and npm packages are inlined into the bundle, so a stored definition never
+  depends on what is installed later. Node builtins (`node:fs`, …) stay as imports and resolve
+  on the worker.
+* The generated config extends your nearest `tsconfig.json`. To use Node globals and
+  builtins, add `"types": ["node"]` to it.
 
-The bundle is one self-contained ES module whose default export is the author's own — nothing
-wraps or rewrites it, so what the editor checks and what the realm runs are the same module, and
-**a missing default export is a failed apply** rather than a fault at run time. Imports resolve
-through TypeScript under the same config the check ran with, so a `paths` alias that typechecks
-also bundles. They are inlined at build time, so the string a definition version stores is
-self-contained forever — with one exception: **node builtins stay as imports**, which the realm
-resolves. A package is frozen into the definition; `node:fs` is resolved by whatever runner
-executes it.
-
-### Your tsconfig, your types
+## Running the worker
 
-The generated project config `extends` **the nearest `tsconfig.json` above the script** —
-the one your editor already reads, so the two cannot disagree. Two scripts under two
-different configs are two `tsc` runs.
+```sh
+docker run -e GENROC_SERVER=http://host:8448 ghcr.io/genroc/eval-node:latest
+# or
+GENROC_SERVER=http://localhost:8448 npx genroc-eval-node
+```
 
-Of that config, three keys are the toolchain's and the rest are yours:
+The worker connects to genroc, not the other way round, so it only needs outbound access.
 
-| key | owner | why |
+| env | default | |
 |---|---|---|
-| `lib` | the toolchain | Describes the realm. A worker thread has no `document`, whatever a config claims. |
-| `include` | the toolchain | Forced to `[]`. A base `include` survives beside our `files` and would drag your whole tree in to be checked as scripts. |
-| `types` | **you** | How a script opts into the node globals. The realm has them, so refusing the declarations would only lie. With no tsconfig at all the default is `[]`. |
-
-```jsonc
-// tsconfig.json beside your scripts
-{ "compilerOptions": { "types": ["node"] } }  // now `import { appendFile } from "node:fs/promises"` typechecks
-```
-
-**This is what removes the `$${` escaping above** — a template literal in a `.ts` file is
-never read by genroc, because genctl doubles every `$` on splice.
-
-## The realm — one Worker per execution
-
-`evaluate()` starts a Worker (`realm.ts`), posts the code into it — where a module loader hook
-serves it to `import()`, so the engine numbers the script's frames from the author's own source —
-and races the reply against the budget; `terminate()` runs on every path. That thread is what the contract rests on, and it buys
-exactly three things the previous in-process evaluator could not:
-
-- **The budget is enforced, not merely reported.** A synchronous `while(true){}` never
-  yields, so no in-process timer can interrupt it — the old evaluator hung forever on that
-  input, and said so in this file. Killing the thread is the only bound. Measured on Node
-  24: `terminate()` stops a spinning worker, and the CPU it was burning goes with it.
-- **A fresh global object per execution.** One script cannot configure the next. It is also
-  why there is no compile cache any more: a cache inside a discarded realm can never be hit.
-- **The script's mistakes stay the script's.** An uncaught throw, and `process.exit()`, end
-  the realm and come back as a `422` — neither reaches the runner.
-
-It costs about **27ms per execution** end to end for a trivial script (Node 24, M-series
-laptop), a 200 KiB body about 30ms. Roughly 11ms of that is Node re-stripping `realm.ts`'s
-types on every realm — precompiling it to JavaScript measures 17ms and would buy that back, and
-is deliberately not done: a build artefact that goes stale against its source fails silently,
-and this file is the one where a wrong line number is invisible.
-
-A subprocess per execution measures ~55ms here, twice the thread rather than the ten times it
-cost on the previous runtime. It contains the two things a thread cannot (below), so the upgrade
-path is a live one at a price worth naming rather than a theoretical one.
-
-## What this is not
-
-- **Not deterministic.** A script reads the real clock and the real RNG, and a retry
-  re-executes — so attempt two can differ from attempt one. An earlier version injected a
-  pinned `Date` and a seeded `Math`; nothing could supply a stable `now` (the expression
-  environment has no clock), so the pin was the wall clock under another name, and the `ctx`
-  it needed was surface with nothing behind it. A value that must survive a retry belongs in
-  the definition, passed through `input`.
-- **Not a sandbox.** The realm isolates *execution*, not *authority*: a script gets the
-  worker's filesystem, network and environment, and any node builtin it imports. That is
-  deliberate — a script task is meant to do real work — but the trust boundary stays the
-  same-trust-domain one (your genroc, your worker host). It is not the multi-tenant story, and
-  nothing here should be mistaken for one. Pulling does move the boundary in one useful way:
-  the worker needs no inbound reachability, so nothing but the worker can ask it to run code.
-- **A thread does not contain memory or a native crash.** A worker shares the process
-  address space, so a script that exhausts memory takes the runner with it — and so does a
-  fault in the runtime itself, which is not hypothetical: the previous one segfaulted on
-  resume from laptop sleep, taking the in-flight evaluation with it. Containing that is what
-  the subprocess strategy is for — `eval.ts` keeps HTTP out precisely so it can be swapped
-  underneath.
-- **Concurrency is capped by this worker, not by genroc.** Each evaluation is a thread, and
-  `CONCURRENCY` is how many it will claim at once. Raising it past what the host can run turns
-  a queue back into threads fighting over a core.
-- **Not where imports and type checking happen.** The evaluator takes one self-contained
-  module and knows nothing about TypeScript; `import.ts` is what bundles an author's script and
-  its dependencies into that module, at author time. See above.
-- **`eval.ts` and `realm.ts` know nothing about genroc.** `worker.ts` is the entire
-  queue-facing half, which is what keeps the containment strategy swappable — and what lets
-  the realm's own properties be tested by calling `evaluate()` directly.
+| `GENROC_SERVER` | `http://localhost:8448` | the genroc server |
+| `GENROC_TOKEN` / `GENROC_TOKEN_FILE` | *(none)* | token, or a file holding it, when the server uses `--auth token` |
+| `PROCESS` / `TASK` | `script-node` / `eval_node` | claim only this process / task id; the defaults are script-node's. Set `PROCESS` for a renamed copy, or `""` to drop a filter; **with both `""`, the worker claims every external task.** |
+| `CONCURRENCY` | `4` | scripts run at once |
+| `LEASE_MS` | `30000` | claim lease, renewed every third of it |
+| `POLL_MS` | `250` | idle poll interval |
+| `WORKER_ID` | `evaluator-<pid>` | the claim holder's name |
+
+With auth enabled, create a token that has only the `worker` permission:
+
+    genctl token create --perms worker --label evaluator -q
+
+If the token is rejected, the worker exits instead of retrying.
+
+## Errors
+
+When a script fails, the worker reports one of five error codes for your `on_error` rules to
+match. Retrying does not help with any of them, because the same code fails the same way:
+
+| code | when |
+|---|---|
+| `threw` | the script threw or rejected |
+| `timeout` | it ran past the task's `timeout` (5000 ms when the task has none) |
+| `compile_error` | the module does not parse, or has no default-export function |
+| `nonserializable` | the return value cannot be serialized to JSON (a cycle, a `BigInt`) |
+| `exited` | the script called `process.exit()` |
+
+`error.data` is `{name, stack?}`. `name` is the error's class name, so throwing your own
+`class LimitExceeded extends Error {}` tells your failures apart; setting `e.name` overrides it,
+and a thrown non-`Error` is `Thrown`. The stack is mapped back to your source lines.
+
+If the worker itself faults, it releases the task for another worker and reports no error.
+
+The task's `timeout` is the script's budget, waiting in the queue included. The worker stops
+the script just before it and answers `timeout`, so `external.timeout` means no worker answered
+at all. With `script-node`, set it through the `timeout_ms` input.
+
+## Limits
+
+* **Not a sandbox.** Scripts have the worker host's filesystem, network and environment.
+  Run only code you trust.
+* **Not deterministic.** On a retry the script runs again with the real clock and random
+  numbers. Pass any value that must stay stable through `input`.
+* **Memory is not isolated.** A script that runs out of memory takes down the whole worker.
+* **Run-once is opt-in.** If a worker dies mid-script, the task is claimed again and the script
+  runs again. Mark the task `only_once` if your script has side effects.
+* **Inline code needs escaping.** When you write `code` directly in YAML rather than
+  `$import`, write `${` as `$${`, because genroc reads `${` as interpolation. Scripts in `.ts`
+  files need no escaping.
+
+## Links
+
+* [TypeScript evaluation guide](https://genroc.org/guides/process-definition/typescript-evaluation/)
+* [External tasks and the queue](https://genroc.org/guides/process-definition/external-tasks/)
+* [Error handling](https://genroc.org/guides/process-definition/error-handling/)
+* [Source on GitHub](https://github.com/genroc/genroc/tree/main/eval-node)

package/dist/import.js

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
-// The code-phase resolver: manifest on stdin, `{"code": [...]}` on stdout, diagnostic on stderr
-// with a non-zero exit (specs/source-resolution.md). Two modes share one `tsc` pass: "types"
-// writes declarations only, "build" typechecks and bundles.
+// The typed-phase resolver: manifest on stdin, `{"values": [...]}` on stdout, diagnostic on stderr
+// with a non-zero exit (specs/source-resolution.md). Two modes share one `tsc` pass: "generate"
+// writes declarations only, "resolve" typechecks and bundles.
 import { existsSync } from "node:fs";
 import { access, writeFile } from "node:fs/promises";
 import { builtinModules } from "node:module";
@@ -116,7 +116,7 @@ function declarations(at) {
         emitted.push(`export type ${identifier(next)} = ${body};`);
     }
     return [
-        "// Generated by genroc. Do not edit - regenerate with `genctl types`.",
+        "// Generated by genroc. Do not edit - regenerate with `genctl generate`.",
         `// ${where.name}  (${address(site.pointer)})`,
         "",
         ...emitted,
@@ -307,11 +307,14 @@ function address(pointer) {
         .join("")
         .replace(/^\./, "");
 }
-const located = manifest.processes.flatMap((where) => where.sites.map((site) => ({
-    site,
-    where,
-    file: resolve(where.dir, site.argument),
-})));
+const located = manifest.processes.flatMap((where) => where.sites.map((site) => {
+    const [script, ...rest] = site.args;
+    if (script === undefined || rest.length > 0) {
+        die(`${join(where.dir, where.file)}: ${address(site.pointer)}: $import takes one script, ` +
+            `and is given ${site.args.length} arguments`);
+    }
+    return { site, where, file: resolve(where.dir, script) };
+}));
 // The `code`-field contract is this resolver's, not genctl's, so it checks the directive landed
 // there: in a child call (the scaffold's shape) or an external task's input.
 for (const at of located) {
@@ -328,7 +331,7 @@ for (const at of located) {
 for (const at of located) {
     if (!existsSync(at.file)) {
         die(`${join(at.where.dir, at.where.file)}: ${address(at.site.pointer)}: ` +
-            `"$import: ${at.site.argument}" names no file (looked at ${at.file})`);
+            `"$import: ${at.site.args[0]}" names no file (looked at ${at.file})`);
     }
 }
 // One script at two sites with different input types is a refusal, not a union: the union
@@ -347,13 +350,12 @@ for (const at of located) {
 for (const at of byPath.values()) {
     await writeFile(typesPathFor(at.file), declarations(at));
 }
-if (manifest.mode === "types") {
-    console.log(JSON.stringify(manifest, null, 2));
+if (manifest.mode === "generate") {
     process.exit(0);
 }
 await typecheck([...byPath.values()]);
-const code = [];
+const values = [];
 for (const at of located) {
-    code.push(await bundle(at));
+    values.push(await bundle(at));
 }
-process.stdout.write(JSON.stringify({ code }));
+process.stdout.write(JSON.stringify({ values }));

package/dist/realm.js

@@ -36,7 +36,10 @@ function scriptStack(err) {
 }
 function describe(err, kind) {
     if (err instanceof Error) {
-        return { kind, name: err.name, message: err.message, stack: scriptStack(err) };
+        // A subclass inherits `name` from Error.prototype, so `class LimitExceeded extends Error {}`
+        // would report "Error"; a name the script set still wins.
+        const name = err.name === "Error" && err.constructor.name ? err.constructor.name : err.name;
+        return { kind, name, message: err.message, stack: scriptStack(err) };
     }
     // A script may throw a non-Error (`throw {code: "x"}`), so name/message must not assume one.
     return { kind, name: "Thrown", message: safeText(err) };

package/dist/worker.js

@@ -19,8 +19,13 @@ const POLL_MS = Number(process.env.POLL_MS ?? 250);
 // should return its task quickly rather than holding it for the whole budget.
 const LEASE_MS = Number(process.env.LEASE_MS ?? 30_000);
 const RENEW_MS = Math.max(1_000, Math.floor(LEASE_MS / 3));
-const PROCESS_FILTER = process.env.PROCESS ?? "";
-const TASK_FILTER = process.env.TASK ?? "";
+// The answer has to land before the task's deadline fires, or a script that overran surfaces as
+// external.timeout ("nobody answered") rather than as its own `timeout`.
+const ANSWER_MARGIN_MS = 250;
+// script-node's names. A renamed copy needs PROCESS set; an explicit "" drops that filter, and
+// both "" claims every external task, other runtimes' included.
+const PROCESS_FILTER = process.env.PROCESS ?? "script-node";
+const TASK_FILTER = process.env.TASK ?? "eval_node";
 // Large values arrive as refs; a bundle is one object per definition version, fetched once.
 // Refs are content hashes, so the cache never invalidates.
 const objectCache = new Map();
@@ -94,17 +99,13 @@ async function call(path, body) {
 }
 /** A task input that is not an EvalRequest is the definition's fault, reported as compile_error:
  *  the nearest permanent kind, since no retry fixes the definition. */
-function asEvalRequest(input) {
+function asEvalRequest(input, timeoutMs) {
     if (typeof input !== "object" || input === null)
         return "the task input is not an object";
     const r = input;
     if (typeof r.code !== "string")
         return "the task input has no `code` string";
-    return {
-        code: r.code,
-        input: r.input,
-        timeout_ms: typeof r.timeout_ms === "number" ? r.timeout_ms : undefined,
-    };
+    return { code: r.code, input: r.input, timeout_ms: timeoutMs };
 }
 const inFlight = new Map();
 let running = true;
@@ -161,6 +162,9 @@ async function answer(token, outcome) {
     await release(token);
 }
 async function run(job, signal) {
+    // The task's own timeout is the script's budget. Taken on arrival, because fetching the
+    // objects below spends the same deadline.
+    const due = job.deadline_in_ms === undefined ? undefined : performance.now() + job.deadline_in_ms;
     let resolved;
     try {
         resolved = await resolveObjects(job);
@@ -172,7 +176,16 @@ async function run(job, signal) {
         await release(job.token);
         return;
     }
-    const req = asEvalRequest(resolved);
+    let timeoutMs;
+    if (due !== undefined) {
+        timeoutMs = Math.floor(due - performance.now() - ANSWER_MARGIN_MS);
+        if (timeoutMs <= 0) {
+            // Too late to run anything: the deadline fires as external.timeout, which is the truth.
+            await release(job.token);
+            return;
+        }
+    }
+    const req = asEvalRequest(resolved, timeoutMs);
     if (typeof req === "string") {
         await answer(job.token, {
             error: { code: "compile_error", message: req, data: { name: "BadTaskInput" } },

package/import.ts

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
-// The code-phase resolver: manifest on stdin, `{"code": [...]}` on stdout, diagnostic on stderr
-// with a non-zero exit (specs/source-resolution.md). Two modes share one `tsc` pass: "types"
-// writes declarations only, "build" typechecks and bundles.
+// The typed-phase resolver: manifest on stdin, `{"values": [...]}` on stdout, diagnostic on stderr
+// with a non-zero exit (specs/source-resolution.md). Two modes share one `tsc` pass: "generate"
+// writes declarations only, "resolve" typechecks and bundles.
 
 import { existsSync } from "node:fs";
 import { access, writeFile } from "node:fs/promises";
@@ -25,9 +25,9 @@ type Site = {
   child?: string;
   // The slot as a path of keys and indices, so nothing has to unescape a JSON Pointer.
   pointer: (string | number)[];
-  // Everything after `$<resolver>:`, verbatim. genctl does not read it as a path — that is
-  // this resolver's reading of it.
-  argument: string;
+  // The argument's words, split as sh splits single quotes. genctl does not read them as paths —
+  // that is this resolver's reading of them.
+  args: string[];
   // What .genroc asked genctl to type, keyed by the name it chose. This resolver reads Input
   // and Output; which ADDRESS each came from is the config's business, not ours.
   types?: Record<string, Schema>;
@@ -45,8 +45,7 @@ type ManifestProcess = {
 };
 
 type Manifest = {
-  mode: "types" | "build";
-  root: string;
+  mode: "generate" | "resolve";
   processes: ManifestProcess[];
 };
 
@@ -170,7 +169,7 @@ function declarations(at: Located): string {
   }
 
   return [
-    "// Generated by genroc. Do not edit - regenerate with `genctl types`.",
+    "// Generated by genroc. Do not edit - regenerate with `genctl generate`.",
     `// ${where.name}  (${address(site.pointer)})`,
     "",
     ...emitted,
@@ -401,11 +400,16 @@ function address(pointer: (string | number)[]): string {
     .replace(/^\./, "");
 }
 const located: Located[] = manifest.processes.flatMap((where) =>
-  where.sites.map((site) => ({
-    site,
-    where,
-    file: resolve(where.dir, site.argument),
-  })),
+  where.sites.map((site) => {
+    const [script, ...rest] = site.args;
+    if (script === undefined || rest.length > 0) {
+      die(
+        `${join(where.dir, where.file)}: ${address(site.pointer)}: $import takes one script, ` +
+          `and is given ${site.args.length} arguments`,
+      );
+    }
+    return { site, where, file: resolve(where.dir, script) };
+  }),
 );
 
 // The `code`-field contract is this resolver's, not genctl's, so it checks the directive landed
@@ -428,7 +432,7 @@ for (const at of located) {
   if (!existsSync(at.file)) {
     die(
       `${join(at.where.dir, at.where.file)}: ${address(at.site.pointer)}: ` +
-        `"$import: ${at.site.argument}" names no file (looked at ${at.file})`,
+        `"$import: ${at.site.args[0]}" names no file (looked at ${at.file})`,
     );
   }
 }
@@ -455,15 +459,14 @@ for (const at of byPath.values()) {
   await writeFile(typesPathFor(at.file), declarations(at));
 }
 
-if (manifest.mode === "types") {
-  console.log(JSON.stringify(manifest, null, 2));
+if (manifest.mode === "generate") {
   process.exit(0);
 }
 
 await typecheck([...byPath.values()]);
 
-const code: string[] = [];
+const values: string[] = [];
 for (const at of located) {
-  code.push(await bundle(at));
+  values.push(await bundle(at));
 }
-process.stdout.write(JSON.stringify({ code }));
+process.stdout.write(JSON.stringify({ values }));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@genroc/eval-node",
-  "version": "0.0.0-edge.e4a77190",
+  "version": "0.0.0-edge.e50224ac",
   "type": "module",
   "engines": {
     "node": ">=24"

package/realm.ts

@@ -35,7 +35,10 @@ function scriptStack(err: unknown): string | undefined {
 
 function describe(err: unknown, kind: FailureKind): EvalFailure {
   if (err instanceof Error) {
-    return { kind, name: err.name, message: err.message, stack: scriptStack(err) };
+    // A subclass inherits `name` from Error.prototype, so `class LimitExceeded extends Error {}`
+    // would report "Error"; a name the script set still wins.
+    const name = err.name === "Error" && err.constructor.name ? err.constructor.name : err.name;
+    return { kind, name, message: err.message, stack: scriptStack(err) };
   }
   // A script may throw a non-Error (`throw {code: "x"}`), so name/message must not assume one.
   return { kind, name: "Thrown", message: safeText(err) };

package/worker.ts

@@ -22,8 +22,13 @@ const POLL_MS = Number(process.env.POLL_MS ?? 250);
 // should return its task quickly rather than holding it for the whole budget.
 const LEASE_MS = Number(process.env.LEASE_MS ?? 30_000);
 const RENEW_MS = Math.max(1_000, Math.floor(LEASE_MS / 3));
-const PROCESS_FILTER = process.env.PROCESS ?? "";
-const TASK_FILTER = process.env.TASK ?? "";
+// The answer has to land before the task's deadline fires, or a script that overran surfaces as
+// external.timeout ("nobody answered") rather than as its own `timeout`.
+const ANSWER_MARGIN_MS = 250;
+// script-node's names. A renamed copy needs PROCESS set; an explicit "" drops that filter, and
+// both "" claims every external task, other runtimes' included.
+const PROCESS_FILTER = process.env.PROCESS ?? "script-node";
+const TASK_FILTER = process.env.TASK ?? "eval_node";
 
 type ObjectEntry = { path: (string | number)[]; ref: string; size: number };
 
@@ -34,6 +39,7 @@ type QueueTask = {
   external_input: unknown;
   objects?: ObjectEntry[];
   raises?: Record<string, unknown>;
+  deadline_in_ms?: number;
 };
 
 // Large values arrive as refs; a bundle is one object per definition version, fetched once.
@@ -106,15 +112,11 @@ async function call(path: string, body: unknown): Promise<{ ok: boolean; status:
 
 /** A task input that is not an EvalRequest is the definition's fault, reported as compile_error:
  *  the nearest permanent kind, since no retry fixes the definition. */
-function asEvalRequest(input: unknown): EvalRequest | string {
+function asEvalRequest(input: unknown, timeoutMs: number | undefined): EvalRequest | string {
   if (typeof input !== "object" || input === null) return "the task input is not an object";
   const r = input as Record<string, unknown>;
   if (typeof r.code !== "string") return "the task input has no `code` string";
-  return {
-    code: r.code,
-    input: r.input,
-    timeout_ms: typeof r.timeout_ms === "number" ? r.timeout_ms : undefined,
-  };
+  return { code: r.code, input: r.input, timeout_ms: timeoutMs };
 }
 
 /** genroc cannot call us, so a cancellation arrives on our own heartbeat and reaches the
@@ -183,6 +185,9 @@ async function answer(token: string, outcome: Record<string, unknown>): Promise<
 }
 
 async function run(job: QueueTask, signal: AbortSignal): Promise<void> {
+  // The task's own timeout is the script's budget. Taken on arrival, because fetching the
+  // objects below spends the same deadline.
+  const due = job.deadline_in_ms === undefined ? undefined : performance.now() + job.deadline_in_ms;
   let resolved: unknown;
   try {
     resolved = await resolveObjects(job);
@@ -193,7 +198,16 @@ async function run(job: QueueTask, signal: AbortSignal): Promise<void> {
     await release(job.token);
     return;
   }
-  const req = asEvalRequest(resolved);
+  let timeoutMs: number | undefined;
+  if (due !== undefined) {
+    timeoutMs = Math.floor(due - performance.now() - ANSWER_MARGIN_MS);
+    if (timeoutMs <= 0) {
+      // Too late to run anything: the deadline fires as external.timeout, which is the truth.
+      await release(job.token);
+      return;
+    }
+  }
+  const req = asEvalRequest(resolved, timeoutMs);
   if (typeof req === "string") {
     await answer(job.token, {
       error: { code: "compile_error", message: req, data: { name: "BadTaskInput" } },