@genroc/eval-node 0.0.0-edge.51cbf37 → 0.0.0-edge.51f2744b

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/realm.js CHANGED
@@ -1,81 +1,45 @@
1
- // The evaluation realm. One Worker per execution: a fresh global object per script, and a
2
- // thread the host can kill mid-loop — the only thing that bounds a synchronous busy loop.
3
- // eval.ts owns the budget and does the killing; nothing here knows about time.
4
- //
5
- // Everything that touches the script's VALUE lives on this side of the boundary — compiling,
6
- // classifying, serialising — because this is the only realm the value exists in.
7
- import { createRequire } from "node:module";
1
+ // The evaluation realm: one Worker per execution, killable mid-loop by eval.ts, which owns the
2
+ // budget. Everything touching the script's VALUE lives here, the only realm it exists in.
3
+ var __rewriteRelativeImportExtension = (this && this.__rewriteRelativeImportExtension) || function (path, preserveJsx) {
4
+ if (typeof path === "string" && /^\.\.?\//.test(path)) {
5
+ return path.replace(/\.(tsx)$|((?:\.d)?)((?:\.[^./]+?)?)\.([cm]?)ts$/i, function (m, tsx, d, ext, cm) {
6
+ return tsx ? preserveJsx ? ".jsx" : ".js" : d && (!ext || !cm) ? m : (d + ext + "." + cm.toLowerCase() + "js");
7
+ });
8
+ }
9
+ return path;
10
+ };
11
+ import { registerHooks } from "node:module";
8
12
  import { parentPort } from "node:worker_threads";
9
- const AsyncFunction = async function () { }.constructor;
10
- const STRICT = '"use strict";\n';
11
- // Bundled `node:*` imports survive as `require` calls — the importer externalises builtins and
12
- // inlines everything else — and a function built by the AsyncFunction constructor has no
13
- // `require` in scope. Passing one in is what makes an import of a builtin work at runtime rather than at
14
- // typecheck only. Resolution is anchored here, which is right: only builtins reach it.
15
- const scriptRequire = createRequire(import.meta.url);
13
+ // IMPORTED under our own URL rather than compiled from a string, so frames carry the author's
14
+ // line numbers and builtin imports resolve normally.
15
+ const SCRIPT_URL = "script:main";
16
16
  const STACK_BYTES = 2_048;
17
- /**
18
- * Line offset the AsyncFunction preamble adds, measured rather than assumed: the generated
19
- * wrapper's shape is engine-specific, and a hardcoded number silently misreports every
20
- * script's error location the day it changes.
21
- */
22
- const lineOffset = (async () => {
23
- // Same parameter list as a real compile: the preamble is what is being measured.
24
- const probe = new AsyncFunction("input", "require", STRICT + "throw new Error('probe');");
25
- try {
26
- await probe();
27
- return 0;
28
- }
29
- catch (err) {
30
- return reportedLine(err) - 1;
31
- }
32
- })();
33
- // V8 marks a frame compiled by the AsyncFunction constructor with the site that CALLED the
34
- // constructor, then the script's OWN position:
35
- // at inner (eval at run (file:///…/worker.ts:107:10), <anonymous>:6:9)
36
- // `eval at` is therefore what separates script frames from runner plumbing — matched without
37
- // the function name, which is whatever encloses the `new AsyncFunction` below. The LAST such
38
- // frame is the body's top level: frames interleave, since a script can throw inside a native
39
- // callback.
40
- const SCRIPT_FRAME = /\(eval at /;
41
- // The trailing `<anonymous>:LINE:COL` — the script's position, after the host file's own.
42
- const POSITION = /<anonymous>:(\d+):(\d+)\)?\s*$/;
43
- // ` at name (` — absent on the top-level frame, which V8 names `eval`.
44
- const FRAME_NAME = /^\s*at\s+(?:async\s+)?([^\s(]+)\s*\(/;
45
- /** Line number of the throw as the engine reported it, or 1 if the stack is unreadable. */
46
- function reportedLine(err) {
47
- const stack = err instanceof Error && typeof err.stack === "string" ? err.stack : "";
48
- const frame = stack.split("\n").find((l) => SCRIPT_FRAME.test(l)) ?? "";
49
- const m = frame.match(POSITION);
50
- return m ? Number(m[1]) : 1;
51
- }
52
- /** Renumbers each script frame to the line the AUTHOR wrote and drops the runner's own.
53
- * Rewriting the whole location is also what keeps the runner's path out of a script's
54
- * stack — V8 puts it inside every compiled frame. */
55
- function scriptStack(err, offset) {
17
+ // The only channel a load hook has to the source. Written per request and read once — a realm
18
+ // evaluates one script and is then discarded, so no second execution can observe it.
19
+ let source = "";
20
+ registerHooks({
21
+ resolve: (specifier, context, next) => specifier === SCRIPT_URL ? { url: SCRIPT_URL, shortCircuit: true } : next(specifier, context),
22
+ load: (url, context, next) => url === SCRIPT_URL ? { format: "module", source, shortCircuit: true } : next(url, context),
23
+ });
24
+ /** Keeps the frames that are the script's own. Everything below the last of them is runner
25
+ * plumbing the author cannot act on, and V8 puts this file's path in it. */
26
+ function scriptStack(err) {
56
27
  if (!(err instanceof Error) || typeof err.stack !== "string")
57
28
  return undefined;
58
29
  const lines = err.stack.split("\n");
59
- let boundary = -1;
30
+ let last = 0;
60
31
  for (let i = 0; i < lines.length; i++)
61
- if (SCRIPT_FRAME.test(lines[i]))
62
- boundary = i;
63
- const frames = (boundary >= 0 ? lines.slice(0, boundary + 1) : lines.slice(0, 1))
64
- .map((line) => {
65
- const pos = line.match(POSITION);
66
- if (!pos)
67
- return line; // the `Error: message` header and native frames, kept as-is
68
- const name = line.match(FRAME_NAME)?.[1];
69
- const at = name && name !== "eval" && name !== "anonymous" ? `at ${name} ` : "at ";
70
- const indent = line.match(/^\s*/)[0];
71
- return `${indent}${at}(script:${Math.max(1, Number(pos[1]) - offset)}:${pos[2]})`;
72
- })
73
- .join("\n");
74
- return frames.length > STACK_BYTES ? frames.slice(0, STACK_BYTES) : frames;
32
+ if (lines[i].includes(SCRIPT_URL))
33
+ last = i;
34
+ const stack = lines.slice(0, last + 1).join("\n");
35
+ return stack.length > STACK_BYTES ? stack.slice(0, STACK_BYTES) : stack;
75
36
  }
76
- function describe(err, kind, offset) {
37
+ function describe(err, kind) {
77
38
  if (err instanceof Error) {
78
- return { kind, name: err.name, message: err.message, stack: scriptStack(err, offset) };
39
+ // A subclass inherits `name` from Error.prototype, so `class LimitExceeded extends Error {}`
40
+ // would report "Error"; a name the script set still wins.
41
+ const name = err.name === "Error" && err.constructor.name ? err.constructor.name : err.name;
42
+ return { kind, name, message: err.message, stack: scriptStack(err) };
79
43
  }
80
44
  // A script may throw a non-Error (`throw {code: "x"}`), so name/message must not assume one.
81
45
  return { kind, name: "Thrown", message: safeText(err) };
@@ -88,23 +52,38 @@ function safeText(v) {
88
52
  return String(v);
89
53
  }
90
54
  }
55
+ /** Unparseable code or an unresolvable import is broken code; anything else the import throws is
56
+ * the module's top level running, i.e. the script throwing. */
57
+ function unloadable(err) {
58
+ const code = err?.code;
59
+ return err instanceof SyntaxError || (typeof code === "string" && code.startsWith("ERR_MODULE"));
60
+ }
91
61
  async function run(req) {
92
- const offset = await lineOffset;
93
- let fn;
62
+ source = req.code;
63
+ let main;
94
64
  try {
95
- // No compile cache: the realm is discarded after this execution, so a cache in it could
96
- // never be hit. Repeated compilation is the price of the fresh global object.
97
- fn = new AsyncFunction("input", "require", STRICT + req.code);
65
+ main = (await import(__rewriteRelativeImportExtension(SCRIPT_URL))).default;
98
66
  }
99
67
  catch (err) {
100
- return { ok: false, failure: describe(err, "compile_error", offset) };
68
+ return { ok: false, failure: describe(err, unloadable(err) ? "compile_error" : "threw") };
69
+ }
70
+ if (typeof main !== "function") {
71
+ const got = main === undefined ? "no default export" : `a ${typeof main}`;
72
+ return {
73
+ ok: false,
74
+ failure: {
75
+ kind: "compile_error",
76
+ name: "NoDefaultExport",
77
+ message: `a script must export default a function; this one has ${got}`,
78
+ },
79
+ };
101
80
  }
102
81
  let value;
103
82
  try {
104
- value = await fn(req.input, scriptRequire);
83
+ value = await main(req.input);
105
84
  }
106
85
  catch (err) {
107
- return { ok: false, failure: describe(err, "threw", offset) };
86
+ return { ok: false, failure: describe(err, "threw") };
108
87
  }
109
88
  try {
110
89
  // undefined stringifies to undefined, not "undefined"; an empty body is how genroc
@@ -112,11 +91,18 @@ async function run(req) {
112
91
  return { ok: true, body: value === undefined ? "" : JSON.stringify(value) ?? "" };
113
92
  }
114
93
  catch (err) {
115
- return { ok: false, failure: describe(err, "nonserializable", offset) };
94
+ return { ok: false, failure: describe(err, "nonserializable") };
116
95
  }
117
96
  }
97
+ // eval.ts terminates this thread the moment the reply lands, while stdio may still be in the
98
+ // pipe. An empty write's callback fires once the queue drains, making the reply a barrier.
99
+ function flush(stream) {
100
+ return new Promise((resolve) => stream.write("", () => resolve()));
101
+ }
118
102
  // Non-null because this module only ever runs as a worker entry point; a null port here
119
103
  // would mean eval.ts loaded it as a plain module, which nothing does.
120
104
  parentPort.on("message", async (req) => {
121
- parentPort.postMessage(await run(req));
105
+ const reply = await run(req);
106
+ await Promise.all([flush(process.stdout), flush(process.stderr)]);
107
+ parentPort.postMessage(reply);
122
108
  });
package/dist/worker.js CHANGED
@@ -1,50 +1,44 @@
1
1
  #!/usr/bin/env node
2
- // The queue worker: claims parked `external` script tasks from genroc, evaluates each in its
3
- // own realm, and answers. This is the whole genroc-facing half — eval.ts and realm.ts know
4
- // nothing about the queue, which is what keeps the containment strategy swappable.
5
- //
6
- // See README.md for the contract, and specs/external-task-queue.md for the queue itself.
2
+ // The queue worker: claims parked `external` script tasks, evaluates each in its own realm, and
3
+ // answers. The only genroc-facing half (README.md; specs/external-task-queue.md).
7
4
  import { readFileSync } from "node:fs";
8
- import { evaluate } from "./eval.js";
5
+ import { Cancelled, evaluate } from "./eval.js";
9
6
  const SERVER = (process.env.GENROC_SERVER ?? "http://localhost:8448").replace(/\/$/, "");
10
7
  const WORKER_ID = process.env.WORKER_ID ?? `evaluator-${process.pid}`;
11
- // The credential, when the server runs with --auth token. A worker needs exactly the `worker`
12
- // permission — the four queue verbs plus GET /api/objects — so mint it scoped rather than
13
- // handing a worker an admin token: this is the credential most likely to sit on a machine you
14
- // trust least. specs/api-auth.md §5.
15
- //
16
- // Sent as a header rather than in the URL because Node's fetch REFUSES a URL carrying
17
- // credentials ("Request cannot be constructed from a URL that includes credentials"), so the
18
- // basic-auth-in-the-URL trick that works for genctl is not available here.
19
- // GENROC_TOKEN_FILE is the mounted-secret shape: a credential in a file rather than an
20
- // environment variable, so it stays out of `docker inspect` and out of the process environment
21
- // any child inherits. The inline variable wins when both are set.
8
+ // Needs only the `worker` permission (specs/api-auth.md §5): this credential sits on the machine
9
+ // you trust least. A header, because Node's fetch refuses credentials in a URL. The file form
10
+ // keeps it out of `docker inspect` and child environments; the inline variable wins.
22
11
  const TOKEN = process.env.GENROC_TOKEN ??
23
12
  (process.env.GENROC_TOKEN_FILE ? readFileSync(process.env.GENROC_TOKEN_FILE, "utf8").trim() : "");
24
13
  const authHeaders = TOKEN ? { authorization: `Bearer ${TOKEN}` } : {};
25
- // Concurrency is the worker's to set, and that is the point of pulling: under the old fetch
26
- // shape genroc decided how many scripts ran at once (--max-concurrent, default 200) and the
27
- // evaluator accepted every one of them. Here it claims what it can run and no more, so a
28
- // backlog is a queue rather than 200 threads fighting over a core.
14
+ // The worker claims only what it can run, so a backlog is a queue rather than threads fighting
15
+ // over a core.
29
16
  const CONCURRENCY = Number(process.env.CONCURRENCY ?? 4);
30
17
  const POLL_MS = Number(process.env.POLL_MS ?? 250);
31
18
  // The visibility timeout. Short, and renewed while work is in flight: a worker that dies
32
19
  // should return its task quickly rather than holding it for the whole budget.
33
20
  const LEASE_MS = Number(process.env.LEASE_MS ?? 30_000);
34
21
  const RENEW_MS = Math.max(1_000, Math.floor(LEASE_MS / 3));
35
- const PROCESS_FILTER = process.env.PROCESS ?? "";
36
- const TASK_FILTER = process.env.TASK ?? "";
37
- // Values too large to ship inline are listed rather than carried, and a bundle is exactly that:
38
- // one object shared by every instance of a definition version, fetched once instead of copied
39
- // into each task. A ref is a content hash, so it is immutable — the cache never invalidates.
22
+ // The answer has to land before the task's deadline fires, or a script that overran surfaces as
23
+ // external.timeout ("nobody answered") rather than as its own `timeout`.
24
+ const ANSWER_MARGIN_MS = 250;
25
+ // script-node's names; a renamed copy sets PROCESS. The claim requires both: this worker answers
26
+ // one task's contract, so it must never be handed another's.
27
+ const PROCESS_FILTER = process.env.PROCESS ?? "script-node";
28
+ const TASK_FILTER = process.env.TASK ?? "eval_node";
29
+ if (!PROCESS_FILTER || !TASK_FILTER) {
30
+ console.error("PROCESS and TASK must both name the queue this worker serves; neither may be empty.");
31
+ process.exit(1);
32
+ }
33
+ // Large values arrive as refs; a bundle is one object per definition version, fetched once.
34
+ // Refs are content hashes, so the cache never invalidates.
40
35
  const objectCache = new Map();
41
36
  async function fetchObject(ref) {
42
37
  const cached = objectCache.get(ref);
43
38
  if (cached !== undefined)
44
39
  return cached;
45
- // Left to throw on purpose: this runs while a task is IN FLIGHT, and a task whose input
46
- // cannot be fetched must fail rather than silently run against a missing value. The caller
47
- // releases the claim, so the task returns to the queue.
40
+ // Left to throw: a task whose input cannot be fetched must not run against a missing value.
41
+ // The caller releases the claim.
48
42
  const res = await fetch(`${SERVER}/api/objects/${encodeURIComponent(ref)}`, { headers: authHeaders });
49
43
  if (!res.ok)
50
44
  throw new Error(`fetch object ${ref}: HTTP ${res.status}`);
@@ -62,15 +56,15 @@ async function fetchObject(ref) {
62
56
  /** Put each listed value back where its path says it belongs. Paths are arrays of keys, so this
63
57
  * needs no parser: the whole reason they are not JSON Pointer strings. */
64
58
  async function resolveObjects(job) {
65
- let input = job.input;
59
+ let input = job.external_input;
66
60
  for (const e of job.objects ?? []) {
67
61
  const value = await fetchObject(e.ref);
68
62
  if (e.path.length === 0) {
69
63
  input = value;
70
64
  continue;
71
65
  }
72
- // The path is rooted at the entry and starts with "input", which is the value being rebuilt.
73
- const rest = e.path[0] === "input" ? e.path.slice(1) : e.path;
66
+ // The path is rooted at the entry and starts with "external_input", the value being rebuilt.
67
+ const rest = e.path[0] === "external_input" ? e.path.slice(1) : e.path;
74
68
  if (rest.length === 0) {
75
69
  input = value;
76
70
  continue;
@@ -84,10 +78,8 @@ async function resolveObjects(job) {
84
78
  return input;
85
79
  }
86
80
  async function call(path, body) {
87
- // A network error is a REPLY, not a throw. A worker outlives the server it polls — a
88
- // restart, a rolling deploy, a container coming up before genroc is listening — and an
89
- // unhandled rejection here kills it for a condition the next poll would clear. Status 0
90
- // says "never reached the server", which is distinct from anything genroc answers.
81
+ // A network error is a REPLY (status 0, "never reached the server"), not a throw: a worker
82
+ // outlives restarts and deploys, and an unhandled rejection would kill it.
91
83
  let res;
92
84
  try {
93
85
  res = await fetch(SERVER + path, {
@@ -109,49 +101,39 @@ async function call(path, body) {
109
101
  }
110
102
  return { ok: res.ok, status: res.status, data };
111
103
  }
112
- /** The task input IS an EvalRequest: `code` required, `input` and `timeout_ms` optional. A task
113
- * whose input is not that shape is the definition's fault, not the script's, and is reported
114
- * as a compile_error — the nearest permanent kind, since no retry can fix the definition. */
115
- function asEvalRequest(input) {
104
+ /** A task input that is not an EvalRequest is the definition's fault, reported as compile_error:
105
+ * the nearest permanent kind, since no retry fixes the definition. */
106
+ function asEvalRequest(input, timeoutMs) {
116
107
  if (typeof input !== "object" || input === null)
117
108
  return "the task input is not an object";
118
109
  const r = input;
119
110
  if (typeof r.code !== "string")
120
111
  return "the task input has no `code` string";
121
- return {
122
- code: r.code,
123
- input: r.input,
124
- timeout_ms: typeof r.timeout_ms === "number" ? r.timeout_ms : undefined,
125
- };
112
+ return { code: r.code, input: r.input, timeout_ms: timeoutMs };
126
113
  }
127
114
  const inFlight = new Map();
128
115
  let running = true;
129
- // Whether the last claim reached genroc. A worker polls several times a second, so an
130
- // unreachable server would otherwise emit a line per poll — thousands during a restart, which
131
- // buries the one line that mattered. Announce the TRANSITIONS instead: going away, and coming
132
- // back. Silence in between is the report that nothing changed.
116
+ // Announce reachability TRANSITIONS only: polling several times a second would otherwise emit
117
+ // thousands of lines during a restart.
133
118
  let serverReachable = true;
134
119
  async function claim(n) {
135
120
  const { ok, status, data } = await call("/api/external-tasks/claim", {
136
121
  worker_id: WORKER_ID,
137
122
  limit: n,
138
123
  lease_ms: LEASE_MS,
139
- ...(PROCESS_FILTER ? { process: PROCESS_FILTER } : {}),
140
- ...(TASK_FILTER ? { task: TASK_FILTER } : {}),
124
+ process: PROCESS_FILTER,
125
+ task: TASK_FILTER,
141
126
  });
142
127
  if (!ok) {
143
- // A credential problem is not transient, and polling through it looks like a healthy
144
- // worker that never picks anything up — the worst shape for an operator to debug. Exit
145
- // instead, so a supervisor restarts it and the failure is visible where it happened.
128
+ // A credential problem is not transient, and polling through it looks like a healthy idle
129
+ // worker. Exit, so a supervisor restarts it visibly.
146
130
  if (status === 401 || status === 403) {
147
131
  console.error(`claim rejected (${status}): ${JSON.stringify(data)}\n` +
148
132
  `The server requires authentication. Set GENROC_TOKEN to a token with the 'worker' ` +
149
133
  `permission — mint one with: genctl token create --perms worker --label evaluator -q`);
150
134
  process.exit(1);
151
135
  }
152
- // status 0 is "never reached the server" (see call): a restart, a rolling deploy, a
153
- // network blip. Not an error to act on — the next poll clears it — so it is reported once
154
- // and then waited out.
136
+ // status 0 (see call): a restart or blip the next poll clears, so it is reported once.
155
137
  if (status === 0) {
156
138
  if (serverReachable) {
157
139
  serverReachable = false;
@@ -174,10 +156,8 @@ async function release(token) {
174
156
  if (!ok)
175
157
  console.error(`release failed: ${JSON.stringify(data)}`);
176
158
  }
177
- /** answer submits the outcome. A refusal is NOT retried with a different one: the definition
178
- * declared a contract this worker does not satisfy (an undeclared code, a payload that does
179
- * not fit `raises`), and guessing again would only pick a second wrong answer. Release it, so
180
- * the task returns to the queue and an operator sees it waiting rather than silently gone. */
159
+ /** A refused outcome is not retried with another: the worker does not satisfy the declared
160
+ * contract. Released, so an operator sees the task waiting rather than gone. */
181
161
  async function answer(token, outcome) {
182
162
  const { ok, data } = await call("/api/external-tasks/resolve", { token, ...outcome });
183
163
  if (ok)
@@ -185,7 +165,10 @@ async function answer(token, outcome) {
185
165
  console.error(`genroc refused the outcome for ${token}: ${JSON.stringify(data)}`);
186
166
  await release(token);
187
167
  }
188
- async function run(job) {
168
+ async function run(job, signal) {
169
+ // The task's own timeout is the script's budget, moved onto the monotonic clock so a wall-clock
170
+ // step mid-run cannot stretch it. Fetching the objects below spends the same deadline.
171
+ const due = job.deadline === undefined ? undefined : performance.now() + (Date.parse(job.deadline) - Date.now());
189
172
  let resolved;
190
173
  try {
191
174
  resolved = await resolveObjects(job);
@@ -197,7 +180,16 @@ async function run(job) {
197
180
  await release(job.token);
198
181
  return;
199
182
  }
200
- const req = asEvalRequest(resolved);
183
+ let timeoutMs;
184
+ if (due !== undefined) {
185
+ timeoutMs = Math.floor(due - performance.now() - ANSWER_MARGIN_MS);
186
+ if (timeoutMs <= 0) {
187
+ // Too late to run anything: the deadline fires as external.timeout, which is the truth.
188
+ await release(job.token);
189
+ return;
190
+ }
191
+ }
192
+ const req = asEvalRequest(resolved, timeoutMs);
201
193
  if (typeof req === "string") {
202
194
  await answer(job.token, {
203
195
  error: { code: "compile_error", message: req, data: { name: "BadTaskInput" } },
@@ -206,13 +198,18 @@ async function run(job) {
206
198
  }
207
199
  let result;
208
200
  try {
209
- result = await evaluate(req);
201
+ result = await evaluate(req, signal);
210
202
  }
211
203
  catch (err) {
212
- // The RUNNER faulted, not the script — the one class where a retry can help. There is no
213
- // error code for it on purpose: releasing the claim is how a queue spells "retryable", and
214
- // it puts the task in front of a different worker instead of burning the definition's
215
- // on_error budget on this one's bad day.
204
+ // Cancelled is not a fault. Still released: the row is terminal so nothing re-claims it,
205
+ // but the release stops it waiting out a lease nobody serves.
206
+ if (err instanceof Cancelled) {
207
+ console.log(`cancelled: ${job.token}`);
208
+ await release(job.token);
209
+ return;
210
+ }
211
+ // The RUNNER faulted, the one retryable class. Releasing is how a queue spells "retryable",
212
+ // and it reaches another worker without spending the definition's on_error budget.
216
213
  console.error(`evaluator fault on ${job.token}: ${err instanceof Error ? err.message : String(err)}`);
217
214
  await release(job.token);
218
215
  return;
@@ -245,11 +242,18 @@ async function renewLoop() {
245
242
  tokens,
246
243
  lease_ms: LEASE_MS,
247
244
  });
248
- // A short count means a claim lapsed and was taken over. Nothing to do about it — the work
249
- // continues and its answer will be refused — but say so, because it is the signal that
250
- // LEASE_MS is too short for what these scripts actually take.
251
- if (ok && data?.renewed < tokens.length) {
252
- console.error(`renewed ${data.renewed}/${tokens.length} claims; a lease lapsed under load`);
245
+ if (!ok)
246
+ continue;
247
+ // Abort first rather than let a stopped script burn a core; the release rides run()'s
248
+ // catch, so nothing is released twice.
249
+ for (const token of (data?.cancelled ?? [])) {
250
+ inFlight.get(token)?.abort.abort();
251
+ }
252
+ // Lost: another worker holds it now and our answer will be refused; logged as the sign that
253
+ // LEASE_MS is too short. NOT released: that would bump the epoch out from under the holder.
254
+ const lost = (data?.lost ?? []);
255
+ if (lost.length) {
256
+ console.error(`lost ${lost.length}/${tokens.length} claims; a lease lapsed under load`);
253
257
  }
254
258
  }
255
259
  }
@@ -258,8 +262,9 @@ async function pollLoop() {
258
262
  const free = CONCURRENCY - inFlight.size;
259
263
  const jobs = free > 0 ? await claim(free) : [];
260
264
  for (const job of jobs) {
261
- inFlight.set(job.token, job);
262
- void run(job).finally(() => inFlight.delete(job.token));
265
+ const abort = new AbortController();
266
+ inFlight.set(job.token, { job, abort });
267
+ void run(job, abort.signal).finally(() => inFlight.delete(job.token));
263
268
  }
264
269
  // Only idle when there was nothing to take: a full queue should be drained at the speed the
265
270
  // realms allow, not at the poll interval.
package/eval.ts CHANGED
@@ -1,22 +1,18 @@
1
- // Evaluation core: run a code string in its OWN realm and classify every outcome into one of
2
- // the failure kinds in README.md. Nothing here knows about genroc — worker.ts is the only
3
- // thing that talks to the queue — so this stays testable and the containment stays swappable.
4
- //
5
- // The containment is a Worker per execution (realm.ts). It is what makes the budget real:
6
- // a synchronous busy loop never yields, so no in-process timer can interrupt it, and only a
7
- // thread the host can kill bounds it.
1
+ // Imports a script module in its own realm (a Worker per execution, realm.ts), calls its default
2
+ // export, and classifies every outcome (README.md). Only a killable thread bounds a synchronous
3
+ // busy loop, which is what makes the budget real. Knows nothing of genroc; worker.ts does.
8
4
 
9
5
  import { Worker } from "node:worker_threads";
10
6
 
11
7
  export type EvalRequest = {
8
+ /** An ES module whose default export is the function to run; `input` is its argument. */
12
9
  code: string;
13
10
  input?: unknown;
14
11
  timeout_ms?: number;
15
12
  };
16
13
 
17
- /** Every kind here is PERMANENT: a retry re-runs the same code on the same input and fails
18
- * identically. The retryable class has no kind because it is not an outcome — a runner that
19
- * faults releases its claim and lets another worker take the task. */
14
+ /** Every kind is PERMANENT: a retry fails identically. A faulting runner is not an outcome; it
15
+ * releases its claim. */
20
16
  export type FailureKind = "compile_error" | "threw" | "timeout" | "nonserializable" | "exited";
21
17
 
22
18
  export type EvalFailure = {
@@ -26,9 +22,8 @@ export type EvalFailure = {
26
22
  stack?: string;
27
23
  };
28
24
 
29
- /** `body` is JSON TEXT, not a value: serialising in the realm is what makes a nonserializable
30
- * return a script fault rather than a 500 thrown out of the response path. It is also what
31
- * crosses the worker boundary — structured clone would refuse a different set of values. */
25
+ /** JSON TEXT serialised in the realm, so a nonserializable return is a script fault, not a 500,
26
+ * and structured clone (which refuses a different set of values) never sees the value. */
32
27
  export type EvalResult =
33
28
  | { ok: true; body: string }
34
29
  | { ok: false; failure: EvalFailure };
@@ -45,9 +40,8 @@ const REALM_URL = new URL(
45
40
  import.meta.url,
46
41
  );
47
42
 
48
- /** Thrown, not returned: a realm that fails to start is the RUNNER faulting, which worker.ts
49
- * answers by releasing the claim rather than by reporting an outcome. A script fault is a
50
- * return value. */
43
+ /** Thrown, not returned: a realm that fails to start is the runner faulting, which worker.ts
44
+ * answers by releasing the claim. */
51
45
  class RealmFault extends Error {
52
46
  constructor(message: string) {
53
47
  super(message);
@@ -55,24 +49,41 @@ class RealmFault extends Error {
55
49
  }
56
50
  }
57
51
 
58
- export async function evaluate(req: EvalRequest): Promise<EvalResult> {
52
+ /** Thrown when `signal` aborts: the answer is no longer wanted, so worker.ts releases rather
53
+ * than answers. Not a RealmFault: nothing is wrong. */
54
+ export class Cancelled extends Error {
55
+ constructor() {
56
+ super("cancelled");
57
+ this.name = "Cancelled";
58
+ }
59
+ }
60
+
61
+ export async function evaluate(req: EvalRequest, signal?: AbortSignal): Promise<EvalResult> {
59
62
  const budget = typeof req.timeout_ms === "number" ? req.timeout_ms : DEFAULT_TIMEOUT_MS;
60
63
 
61
64
  const worker = new Worker(REALM_URL);
62
65
  let timer: ReturnType<typeof setTimeout> | undefined;
66
+ let onAbort: (() => void) | undefined;
63
67
  try {
64
68
  return await new Promise<EvalResult>((resolve, reject) => {
65
69
  timer = setTimeout(() => resolve(timedOut(budget)), budget);
70
+ // The abort is wired to the same promise as the budget, so both settle through the one
71
+ // finally below -- which is what guarantees the thread is gone before either returns.
72
+ if (signal) {
73
+ if (signal.aborted) reject(new Cancelled());
74
+ onAbort = () => reject(new Cancelled());
75
+ signal.addEventListener("abort", onAbort, { once: true });
76
+ }
66
77
  worker.once("message", (reply: WorkerReply) => resolve(reply));
67
- // A script may end its own realm (`process.exit()`), which is not a throw and would
68
- // otherwise present as a hang until the budget expired. Our own terminate() raises
69
- // this too, by which time the promise has settled and the first result stands.
78
+ // A script may `process.exit()`, which would otherwise look like a hang until the budget.
79
+ // Our own terminate() fires this too, after the promise has settled.
70
80
  worker.once("exit", (code: number) => resolve(exited(code)));
71
81
  worker.once("error", (err: Error) => reject(new RealmFault(errorText(err))));
72
82
  worker.postMessage({ code: req.code, input: req.input } satisfies WorkerRequest);
73
83
  });
74
84
  } finally {
75
85
  clearTimeout(timer);
86
+ if (signal && onAbort) signal.removeEventListener("abort", onAbort);
76
87
  // Awaited, and the whole point: on the timeout path a thread is still burning a core, and
77
88
  // resolving before it is gone would report an evaluation the machine is still running.
78
89
  await worker.terminate();