Dashboard › opencode-lore › Session 1WOHBDvviEvs
1WOHBDvviEvs0oK84Read-only audit of /home/byk/Code/opencode-lore-embedding-runtime current change tmzmtlss, focused on embedding admission/headroom sampling outside cgroup-memory.ts: packages/core/src/embedding/local.ts, embedding/pool.ts, embedding-cap.ts and related tests. Identify concrete correctness gaps, especially inconsistent repeated samples, monotonic clamp behavior, unknown headroom fail-closed behavior, pool growth accounting, cap floors/non-finite inputs, and concurrency. Do not edit or mutate VCS. Return severity-ranked findings with exact file:line references and minimal regression/fix suggestions. Avoid auditing cgroup mount resolution itself because another reviewer owns it.
<skill_content name="jj-guide"> # Skill: jj-guide # jj Guide for AI Agents Jujutsu (jj) is a Git-compatible VCS with mutable commits, automatic snapshotting, no staging area, and first-class conflicts. This skill teaches you how to use it safely from a non-interactive agent environment. If `.jj/` exists in the repo root, this is a jj repo. **Use `jj` commands, not `git`.** In a colocated repo (`.jj/` *and* `.git/`), git tools can read the state, but mutations should go through `jj` so the operation log stays consistent. ## Critical Rules - **NEVER** use interactive flags (`-i`, `--interactive`). TUI prompts hang in agent environments. This applies to `jj split -i`, `jj squash -i`, `jj commit -i`, `jj resolve`, `jj diffedit`, etc. - **ALWAYS** pass `-m "msg"` when describing/committing. Without `-m`, an editor opens and hangs. - **VERIFY** mutations with `jj st` and `jj log` after `squash`, `abandon`, `rebase`, `restore`, `commit`. ...
<path>/home/byk/Code/opencode-lore-embedding-runtime</path> <type>directory</type> <entries> .craft.yml .git .github/ .gitignore .jj/ .lore.md .oxfmtrc.json .oxlintrc.json AGENTS.md CHANGELOG.md codecov.yml LICENSE node_modules/ package.json packages/ patches/ pnpm-lock.yaml pnpm-workspace.yaml quality/ README.md scripts/ stryker.config.mjs supabase/ tsconfig.base.json tsconfig.json vitest.config.ts vitest.evals.config.ts vitest.mutation.config.ts (28 entries) </entries>
<path>/home/byk/Code/opencode-lore-embedding-runtime/AGENTS.md</path> <type>file</type> <content> 1: <!-- This section is maintained by the coding agent via lore (https://github.com/BYK/loreai) --> 2: ## Long-term Knowledge 3: 4: For long-term knowledge entries managed by [lore](https://github.com/BYK/loreai) (gotchas, patterns, decisions, architecture), see [`.lore.md`](.lore.md) in the project root. 5: <!-- End lore-managed section --> 6: 7: ## Review 8: 9: See [`quality/REVIEW.md`](quality/REVIEW.md) for the review heuristics and workflow — regression-test discipline, adversarial-order state setup, fan-out registry coverage, recurring bug-class batteries, and the two-reviewer rule. Every non-trivial PR requires an adversarial correctness review before merge. 10: 11: # Project Overview 12: 13: Lore is a **three-tier memory architecture** for AI coding agents. ...
No files found
<path>/home/byk/Code/opencode-lore-embedding-runtime/package.json</path>
<type>file</type>
<content>
1: {
2: "name": "lore-monorepo",
3: "private": true,
4: "type": "module",
5: "license": "FSL-1.1-Apache-2.0",
6: "description": "Monorepo root for Lore — three-tier memory architecture",
7: "main": "./packages/opencode/src/index.ts",
8: "exports": {
9: ".": "./packages/opencode/src/index.ts"
10: },
11: "packageManager": "pnpm@10.28.0",
12: "volta": {
13: "node": "24.16.0",
14: "pnpm": "10.28.0"
15: },
16: "pnpm": {
17: "onlyBuiltDependencies": [
18: "esbuild",
19: "onnxruntime-node",
20: "sharp"
21: ],
22: "overrides": {
23: "basic-ftp": "5.3.1",
24: "ip-address": "10.2.0",
25: "fast-xml-parser": "5.7.3",
26: "@huggingface/hub": "2.11.0",
27: "onnxruntime-node": "1.27.0",
28: "yaml": ">=2.8.3",
29: "esbuild": ">=0.28.1",
30: "@opentelemetry/core": ">=2.8.0",
31: ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/quality/REVIEW.md</path> <type>file</type> <content> 1: # Review Heuristics — Lore 2: 3: This document codifies the review bar for pull requests. It exists because 4: adversarial review has repeatedly caught correctness bugs that a green test 5: suite missed — skip/early-return branches, lifecycle sequences, and fan-out 6: registry gaps. Institutionalizing these patterns so they compound. 7: 8: The review workflow (§6) is referenced from AGENTS.md. 9: 10: --- 11: 12: ## 1. Regression-test discipline 13: 14: Every adversarial-review finding that surfaces a defect **must** land a 15: deterministic regression test in the same PR. The test: 16: - fails on the base branch (proving it reproduces the bug), 17: - passes on the fix (proving the fix works), 18: - drives the *real* precondition, not an artificially-constructed state (e.g. ...
Working copy changes: A packages/core/src/cgroup-memory.ts M packages/core/src/embedding/local.ts M packages/core/src/embedding/pool.ts M packages/core/src/embedding-cap.ts M packages/core/src/embedding-worker-types.ts M packages/core/src/embedding-worker.ts M packages/core/src/embedding.ts M packages/core/src/ort-native.ts A packages/core/test/cgroup-memory.test.ts M packages/core/test/embedding-cap.test.ts M packages/core/test/embedding-pool-memory.test.ts M packages/core/test/embedding-pool.test.ts M packages/core/test/embedding-wasm-fallback.test.ts M packages/core/test/embedding-worker-stdio.test.ts M packages/core/test/embedding-worker-types.test.ts M packages/core/test/ort-native.test.ts M packages/gateway/script/bundle.ts M packages/gateway/script/ort-platform-package.ts M packages/gateway/test/bundle-exports.test.ts M packages/gateway/test/ort-platform-package.test.ts Working copy (@) : tmzmtlss 01e1ee40 fix/embedding-runtime-admission | fix(core): correct embedding runtime ...
tmzmtlsspstn 01e1ee401324aea61a6bf3a33ae56a344b7034c9 parents=2cbf40060c67df229731cf1b9e34d5bf22ec2145 conflict=false empty=false pxrvwsvvkmpo 2cbf40060c67df229731cf1b9e34d5bf22ec2145 parents=e977b4b1da8f8197c3238466d41de0bccd8eb054 conflict=false empty=false
packages/core/src/cgroup-memory.ts packages/core/src/embedding/local.ts packages/core/src/embedding/pool.ts packages/core/src/embedding-cap.ts packages/core/src/embedding-worker-types.ts packages/core/src/embedding-worker.ts packages/core/src/embedding.ts packages/core/src/ort-native.ts packages/core/test/cgroup-memory.test.ts packages/core/test/embedding-cap.test.ts packages/core/test/embedding-pool-memory.test.ts packages/core/test/embedding-pool.test.ts packages/core/test/embedding-wasm-fallback.test.ts packages/core/test/embedding-worker-stdio.test.ts packages/core/test/embedding-worker-types.test.ts packages/core/test/ort-native.test.ts packages/gateway/script/bundle.ts packages/gateway/script/ort-platform-package.ts packages/gateway/test/bundle-exports.test.ts packages/gateway/test/ort-platform-package.test.ts
eb8a9bc8e2c921d3f055314aa87581b3df0fc5b32955f44b4b51e4a3c5e5a1b6 -
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool-memory.test.ts /home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap.test.ts /home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap-persistence.test.ts /home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool.test.ts
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/local.ts</path>
<type>file</type>
<content>
1: /** Local ONNX embedding worker lifecycle. */
2:
3: import { freemem } from "node:os";
4: import { readLinuxCgroupMemoryHeadroomBytes } from "../cgroup-memory";
5: import { db } from "../db";
6: import * as log from "../log";
7: import { vendorModelInfo } from "../embedding-vendor";
8: import { nativeIntraOpThreads } from "../ort-native";
9: import {
10: MIN_EMBED_TOKENS,
11: MODEL_MAX_TOKENS,
12: backoffEmbedCap,
13: clampFreeToContainerLimit,
14: memoryModelEmbedCap,
15: reconcileEmbedCap,
16: reprobeEmbedCap,
17: resolveMemoryHeadroom,
18: shouldReprobeEmbedCap,
19: type PersistedEmbedCap,
20: } from "../embedding-cap";
21: import {
22: EMBED_OOM_EXIT_CODE,
23: isMissingLocalStackError,
24: isWasmFatalError,
25: type EmbedRequest,
26: type WorkerInbound,
27: type WorkerOutbound,
28: type WorkerInitData,
29: } from ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/local.ts</path>
<type>file</type>
<content>
1001: }
1002: try {
1003: await this.ensureWorker();
1004: } catch {
1005: // Synchronous respawn failure — nothing else will settle these requests
1006: // (the local embed path has no timeout). Reject them here.
1007: for (const [, p] of this.pendingRequests) {
1008: p.reject(
1009: new LocalProviderUnavailableError(
1010: "embedding worker respawn failed after WASM fallback",
1011: ),
1012: );
1013: }
1014: this.pendingRequests.clear();
1015: return;
1016: }
1017: // Re-read through a cast: TS still has `this.worker` narrowed to null from
1018: // the `this.worker = null` above (it can't see that ensureWorker() reassigns
1019: // it across the await), so a direct read would be typed `never`.
1020: const worker = this.worker as ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/pool.ts</path>
<type>file</type>
<content>
1: /** Host-owned queue and pool for local embedding workers. */
2:
3: import { createHash } from "node:crypto";
4: import { freemem } from "node:os";
5: import { config } from "../config";
6: import * as log from "../log";
7: import {
8: DEFAULT_MAX_EMBED_POOL,
9: EMBED_POOL_ABS_MAX,
10: PER_WORKER_MEM_BUDGET_BYTES,
11: clampFreeToContainerLimit,
12: } from "../embedding-cap";
13: import {
14: EmbeddingQueueCapacityError,
15: EmbeddingRequestAbortedError,
16: EmbeddingWorkerWatchdogError,
17: type EmbeddingProvider,
18: LocalProviderUnavailableError,
19: WORKER_SHUTDOWN_TIMEOUT_MS,
20: isRecallEmbed,
21: } from "./contract";
22: import {
23: LocalProvider,
24: availableMemoryHeadroom,
25: clearLocalProviderLatch,
26: constrainedMemoryLimit,
27: localEmbeddingState,
28: } from "./local";
29: import { OwnedRetirements } from ...
Found 44 matches
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool.test.ts:
Line 1547: // desiredEmbedPoolSize) would mask it via the worker count. Guards the
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap.test.ts:
Line 10: PER_WORKER_MEM_BUDGET_BYTES,
Line 16: desiredEmbedPoolSize,
Line 111: expect(PER_WORKER_MEM_BUDGET_BYTES).toBe(
Line 114: expect(PER_WORKER_MEM_BUDGET_BYTES).toBe(1536 * MB);
Line 115: expect(desiredEmbedPoolSize(3 * GB - 1)).toBe(1);
Line 116: expect(desiredEmbedPoolSize(3 * GB)).toBe(2);
Line 123: expect(desiredEmbedPoolSize(clamped)).toBe(1);
Line 125: expect(desiredEmbedPoolSize(7 * GB)).toBe(DEFAULT_MAX_EMBED_POOL);
Line 144: expect(desiredEmbedPoolSize(clamped)).toBe(
Line 145: desiredEmbedPoolSize(hostFree),
Line 392: describe("desiredEmbedPoolSize", () => {
Line 395: expect(desiredEmbedPoolSize(0)).toBe(1);
...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding-cap.ts</path> <type>file</type> <content> 1: /** 2: * Adaptive token-cap math for local ONNX embedding inference. 3: * 4: * Local inference OOMs on long inputs: the O(L²) attention tensor for a long 5: * sequence blows the WASM heap, and because WASM linear memory never shrinks, 6: * an in-process retry cannot recover. We instead cap the input sequence length 7: * up-front and adapt that cap to the host — start from a memory-aware estimate, 8: * and on each OOM lower the cap ×0.7 and respawn the worker on a fresh heap. 9: * 10: * This module holds the pure, dependency-free math so it can be unit-tested in 11: * isolation. The stateful pieces (persistence, worker lifecycle, telemetry) 12: * live in embedding.ts. 13: */ 14: 15: /** Floor token cap. ...
diff --git a/packages/core/src/embedding/local.ts b/packages/core/src/embedding/local.ts
index 5c5e7e600f..32f4307b4d 100644
--- a/packages/core/src/embedding/local.ts
+++ b/packages/core/src/embedding/local.ts
@@ -1,6 +1,7 @@
/** Local ONNX embedding worker lifecycle. */
import { freemem } from "node:os";
+import { readLinuxCgroupMemoryHeadroomBytes } from "../cgroup-memory";
import { db } from "../db";
import * as log from "../log";
import { vendorModelInfo } from "../embedding-vendor";
@@ -13,6 +14,7 @@
memoryModelEmbedCap,
reconcileEmbedCap,
reprobeEmbedCap,
+ resolveMemoryHeadroom,
shouldReprobeEmbedCap,
type PersistedEmbedCap,
} from "../embedding-cap";
@@ -76,6 +78,11 @@
testHostFreememBytes = bytes;
}
+let testAvailableMemoryBytes: number | null = null;
+export function _setAvailableMemoryForTest(bytes: number | null): void {
+ testAvailableMemoryBytes = bytes;
+}
+
/**
* The process's cgroup memory LIMIT in bytes (not free-within-limit), or ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/cgroup-memory.ts</path>
<type>file</type>
<content>
1: import { readFileSync } from "node:fs";
2: import { posix } from "node:path";
3:
4: interface CgroupMembership {
5: version: 1 | 2;
6: path: string;
7: }
8:
9: interface CgroupMemoryFiles {
10: version: 1 | 2;
11: directory: string;
12: mountPoint: string;
13: root: string;
14: }
15:
16: type ReadTextFile = (path: string) => string;
17:
18: function readTextFile(path: string): string {
19: return readFileSync(path, "utf8");
20: }
21:
22: function decodeProcPath(value: string): string | null {
23: const decoded = value.replace(/\\([0-7]{3})/g, (_match, octal: string) =>
24: String.fromCharCode(Number.parseInt(octal, 8)),
25: );
26: if (!decoded.startsWith("/") || decoded.includes("\0")) return null;
27: if (decoded.split("/").includes("..")) return null;
28: return posix.normalize(decoded);
29: }
30:
31: function ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/cgroup-memory.ts</path>
<type>file</type>
<content>
180: const limit = readLimit(readText(posix.join(directory, limitFile)));
181: if (limit === undefined) return undefined;
182: if (Number.isFinite(limit)) {
183: const candidate = Math.max(0, limit - usage);
184: headroom =
185: headroom === undefined
186: ? candidate
187: : Math.min(headroom, candidate);
188: }
189: }
190: } catch {
191: // The cgroup v2 root cannot itself be resource-controlled and may omit
192: // memory.current/high/max. Every descendant boundary still must be
193: // readable so sibling usage is never subtracted from the wrong scope.
194: if (
195: files.version === 2 &&
196: directory === files.mountPoint &&
197: files.root ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap.test.ts</path>
<type>file</type>
<content>
1: import { describe, it, expect } from "vitest";
2: import {
3: BACKFILL_MIN_AUTO_DUTY,
4: BACKFILL_THROTTLE_MAX_SLEEP_MS,
5: DEFAULT_MAX_EMBED_POOL,
6: EMBED_POOL_ABS_MAX,
7: EMBED_WORKER_RESIDENT_RESERVE_BYTES,
8: MIN_EMBED_TOKENS,
9: MODEL_MAX_TOKENS,
10: PER_WORKER_MEM_BUDGET_BYTES,
11: EMBED_TOKEN_CEILING,
12: backfillThrottleSleepMs,
13: backoffEmbedCap,
14: clampEmbedCap,
15: clampFreeToContainerLimit,
16: desiredEmbedPoolSize,
17: memoryModelEmbedCap,
18: reconcileEmbedCap,
19: reprobeEmbedCap,
20: resolveBackfillCpuDuty,
21: resolveMemoryHeadroom,
22: shouldReprobeEmbedCap,
23: } from "../src/embedding-cap";
24:
25: const GB = 1024 * 1024 * 1024;
26: const MB = 1024 * 1024;
27:
28: describe("clampFreeToContainerLimit", () => {
29: it("is a no-op when unconstrained (constrained <= 0)", () => {
30: // ...
Found 36 matches
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-self-heal.test.ts:
Line 11: // It re-probes recoverable causes on a slow cadence but never the terminal
Line 28: it("primes on the first call, then re-probes a recoverable latch after the interval", () => {
Line 33: // First call primes the clock — no re-probe yet.
Line 45: it("never re-probes the terminal missing-stack cause", () => {
Line 49: // Even long after the interval, a missing stack is never re-probed.
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding.test.ts:
Line 361: // the OPTIMISTIC pre-probe state (isAvailable() === true). Without this
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap.test.ts:
Line 19: reprobeEmbedCap,
Line 277: it("re-probes upward via the model when memory has materially grown", () => {
Line 305: it("never re-probes up to or past a persisted known-bad cap ...
Found 1 matches /home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/pool.ts: Line 292: const provider = new LocalProvider(
RUN v4.1.8 /home/byk/Code/opencode-lore-embedding-runtime
❯ packages/core/test/embedding-cap.test.ts (62 tests | 1 failed) 44ms
× always accepts finite Node headroom without a hard limit 11ms
❯ packages/core/test/embedding-pool-memory.test.ts (5 tests | 1 failed) 1881ms
× uses the token floor when cgroup accounting fails without a hard limit 9ms
⎯⎯⎯⎯⎯⎯⎯ Failed Tests 2 ⎯⎯⎯⎯⎯⎯⎯
FAIL packages/core/test/embedding-cap.test.ts > resolveMemoryHeadroom > always accepts finite Node headroom without a hard limit
AssertionError: expected undefined to be 7516192768 // Object.is equality
- Expected:
7516192768
+ Received:
undefined
❯ packages/core/test/embedding-cap.test.ts:102:57
100|
101| it("always accepts finite Node headroom without a hard limit", () =>…
102| expect(resolveMemoryHeadroom(0, undefined, 7 * GB)).toBe(7 * GB);
| ^
103| expect(resolveMemoryHeadroom(Number.NaN, ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/owned-retirements.ts</path>
<type>file</type>
<content>
1: /** Own asynchronous retirement operations until they settle. */
2: export class OwnedRetirements<Owner extends object> {
3: private readonly active = new Set<Promise<void>>();
4: private readonly owners = new WeakSet<Owner>();
5: private readonly errors: unknown[] = [];
6:
7: get size(): number {
8: return this.active.size;
9: }
10:
11: track(operation: Promise<void>): Promise<void> {
12: const active = this.active;
13: const errors = this.errors;
14: let tracked: Promise<void>;
15: tracked = operation
16: .then(
17: () => {},
18: (error: unknown) => {
19: errors.push(error);
20: },
21: )
22: .finally(() => active.delete(tracked));
23: active.add(tracked);
24: return operation;
25: }
26:
27: /** Start at most one retirement operation for an owned resource. ...
Found 10 matches
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding-worker.ts:
Line 54: * Main-thread-owned input token cap (see `WorkerInitData.maxTokens`). Every
Line 61: const maxTokens = init.maxTokens ?? 2048;
Line 457: // raw .ts and can't value-import ort-native — see maxTokens above), a strict
Line 628: * Only texts exceeding `maxTokens` chars are checked — at worst-case
Line 631: function truncateTexts(texts: string[], maxTokens: number): string[] {
Line 635: if (text.length <= maxTokens) return text;
Line 639: if (ids.length <= maxTokens) return text;
Line 640: return tk.decode(ids.slice(0, maxTokens), {
Line 771: const effectiveMax = req.maxTokens ?? maxTokens;
Line 787: const effectiveMax = req.maxTokens ?? maxTokens;
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding-worker.ts</path>
<type>file</type>
<content>
40: }
41: const port = parentPort;
42:
43: const init = workerData as WorkerInitData;
44: const { modelId, dimensions, vendorModel } = init;
45: // Force the bundled WASM ONNX Runtime, skipping native-addon resolution (see
46: // WorkerInitData.forceWasm). Set by the main thread only when respawning after a
47: // native worker posted `init-needs-wasm` (#1379). Default false = prefer native.
48: const forceWasm = init.forceWasm ?? false;
49: // Snapshot of the host's stderr-silence state (see WorkerInitData). When true,
50: // every diagnostic below stays off stderr so it can't corrupt the host's TUI.
51: const stderrSilenced = init.stderrSilenced ?? false;
52:
53: /**
54: * Main-thread-owned input token cap (see `WorkerInitData.maxTokens`). ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding-worker.ts</path>
<type>file</type>
<content>
620: }
621:
622: // ---------------------------------------------------------------------------
623: // Embed processing
624: // ---------------------------------------------------------------------------
625:
626: /**
627: * Truncate texts to a maximum number of tokens using the real tokenizer.
628: * Only texts exceeding `maxTokens` chars are checked — at worst-case
629: * ~1 char/token, shorter texts can never exceed the limit.
630: */
631: function truncateTexts(texts: string[], maxTokens: number): string[] {
632: const tk = tokenizer;
633: if (!tk) return texts;
634: return texts.map((text) => {
635: if (text.length <= maxTokens) return text;
636: // Exclude [CLS]/[SEP] special tokens so ids.length reflects pure content
637: // token count — otherwise the 2 extra tokens skew the limit check.
638: const ids = tk.encode(text, { ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool-memory.test.ts</path>
<type>file</type>
<content>
1: import { afterEach, beforeEach, describe, expect, it } from "vitest";
2: import { EventEmitter } from "node:events";
3: import type { Worker } from "node:worker_threads";
4: import {
5: embed,
6: _persistEmbedCap,
7: _resetLocalProviderProbe,
8: _restoreProvider,
9: _saveAndClearProvider,
10: _setConstrainedMemoryForTest,
11: _setContainerFreeForTest,
12: _setAvailableMemoryForTest,
13: _setEmbedPoolSizeForTest,
14: _setPoolFreememForTest,
15: _setTestWorkerFactory,
16: } from "../src/embedding";
17: import {
18: MIN_EMBED_TOKENS,
19: backoffEmbedCap,
20: memoryModelEmbedCap,
21: } from "../src/embedding-cap";
22: import { EMBED_OOM_EXIT_CODE } from "../src/embedding-worker-types";
23:
24: // Regression suite for the embedding-pool OOM that SIGKILLed the whole gateway
25: // (Onur's report): the pool grew to N workers, ...
Found 15 matches
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool-memory.test.ts:
Line 19: backoffEmbedCap,
Line 241: backoffEmbedCap(first.maxTokens),
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap.test.ts:
Line 13: backoffEmbedCap,
Line 177: describe("backoffEmbedCap", () => {
Line 179: expect(backoffEmbedCap(MODEL_MAX_TOKENS)).toBe(Math.round(8192 * 0.7)); // 5734
Line 180: expect(backoffEmbedCap(1000)).toBe(700);
Line 184: expect(backoffEmbedCap(MIN_EMBED_TOKENS)).toBe(MIN_EMBED_TOKENS);
Line 186: expect(backoffEmbedCap(300)).toBe(MIN_EMBED_TOKENS);
Line 193: const next = backoffEmbedCap(cap);
Line 201: expect(backoffEmbedCap(cap)).toBe(MIN_EMBED_TOKENS);
/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-oom-recovery.test.ts:
Line 18: backoffEmbedCap,
Line 143: ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-cap-persistence.test.ts</path>
<type>file</type>
<content>
1: import { describe, it, expect, beforeEach } from "vitest";
2: import { db } from "../src/db";
3: import { _persistEmbedCap, _readPersistedEmbedCap } from "../src/embedding";
4:
5: // Mirrors EMBED_CAP_KV_KEY in embedding.ts (private). Only used to seed corrupt
6: // rows for the guard tests below.
7: const KV_KEY = "lore:embedding_cap";
8:
9: function writeRaw(value: string): void {
10: db()
11: .query(
12: "INSERT INTO kv_meta (key, value) VALUES (?, ?) ON CONFLICT(key) DO UPDATE SET value = ?",
13: )
14: .run(KV_KEY, value, value);
15: }
16:
17: describe("embedding cap persistence", () => {
18: beforeEach(() => {
19: db().query("DELETE FROM kv_meta WHERE key = ?").run(KV_KEY);
20: });
21:
22: it("returns null when no cap is persisted", () => {
23: expect(_readPersistedEmbedCap()).toBeNull();
24: ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool.test.ts</path>
<type>file</type>
<content>
720: await flush();
721: expect(fakes).toHaveLength(3);
722: expect(fakes[2].embedIds).toHaveLength(1);
723: fakes[0].completeNext();
724: fakes[2].completeNext();
725: await Promise.all([firstQueued, secondQueued]);
726: });
727:
728: it("normalizes an initialization watchdog at the readiness boundary", async () => {
729: _setEmbedPoolSizeForTest(1);
730: _setEmbeddingWorkerWatchdogsForTest(10, 60_000);
731: installFakeWorkers();
732:
733: const outcome = await settle(ensureEmbeddingReady({ deadlineMs: 1_000 }));
734: expect(outcome.ok).toBe(false);
735: if (!outcome.ok) {
736: expect(outcome.err).toBeInstanceOf(LocalProviderUnavailableError);
737: expect(outcome.err).not.toBeInstanceOf(EmbeddingWorkerWatchdogError);
738: expect((outcome.err as Error & { cause?: unknown ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-pool.test.ts</path>
<type>file</type>
<content>
1: import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
2: import { EventEmitter } from "node:events";
3: import type { Worker } from "node:worker_threads";
4: import {
5: embed,
6: embedInTokenBatches,
7: EmbeddingQueueCapacityError,
8: EmbeddingRequestAbortedError,
9: EmbeddingWorkerWatchdogError,
10: ensureEmbeddingReady,
11: isAvailable,
12: LocalProviderUnavailableError,
13: recallEmbedsInFlight,
14: resetProvider,
15: shutdownProvider,
16: _configuredEmbedPoolSize,
17: _getLocalInitRetryAtForTest,
18: _resetLocalProviderProbe,
19: _resetProviderShutdownTrackingForTest,
20: _restoreProvider,
21: _saveAndClearProvider,
22: _setConstrainedMemoryForTest,
23: _setAvailableMemoryForTest,
24: _setEmbedPoolSizeForTest,
25: _setEmbeddingWorkerWatchdogsForTest,
26: ...
<path>/home/byk/Code/opencode-lore-embedding-runtime/packages/core/test/embedding-oom-recovery.test.ts</path>
<type>file</type>
<content>
1: import { afterEach, beforeEach, describe, expect, it } from "vitest";
2: import { EventEmitter } from "node:events";
3: import type { Worker } from "node:worker_threads";
4: import {
5: embed,
6: isAvailable,
7: LocalProviderUnavailableError,
8: _persistEmbedCap,
9: _readPersistedEmbedCap,
10: _resetLocalProviderProbe,
11: _restoreProvider,
12: _saveAndClearProvider,
13: _setConstrainedMemoryForTest,
14: _setContainerFreeForTest,
15: _setTestWorkerFactory,
16: } from "../src/embedding";
17: import {
18: backoffEmbedCap,
19: MIN_EMBED_TOKENS,
20: EMBED_TOKEN_CEILING,
21: } from "../src/embedding-cap";
22: import { EMBED_OOM_EXIT_CODE } from "../src/embedding-worker-types";
23: import { isStderrSilenced, silenceStderr } from "../src/log";
24:
25: // Exercises the stateful OOM-recovery lifecycle in LocalProvider that the ...
Found 3 matches /home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/pool.ts: Line 84: healthy: boolean; Line 308: healthy: false, Line 551: slot.healthy = true;
Found 5 matches /home/byk/Code/opencode-lore-embedding-runtime/packages/core/src/embedding/local.ts: Line 139: function persistEmbedCap( Line 177: export function _persistEmbedCap( Line 182: persistEmbedCap(cap, freeMemBytes, knownBadCap); Line 791: persistEmbedCap(next, free, this.lastOomCap); Line 857: persistEmbedCap(capAfter, free, this.lastOomCap);
Found 26 matches /home/byk/Code/opencode-lore-embedding-runtime/.lore.md: Line 29: * **Embedding token batching and temporal chunk cap**: Use embedInTokenBatches for all embedding paths, including temporal messages, to prevent ONNX padding all inputs to the longest sequence and causing OOM. Batch by token area (MAX\_BATCH\_TOKEN\_AREA 4096), but always admit the first item so the loop progresses. Preserve input order and call storeTemporalChunks once. Cap temporal chunks at 64: retain the first 63 and join the remainder into one tail; an oversized tail may occupy its own batch and is truncated downstream. Line 38: * **embedding worker stdio drain**: Chose silently calling \`resume()\` on owned embedding-worker stdout/stderr over logging sanitized or rate-limited lines because native dependencies can emit model paths, provider diagnostics, IDs, and process state, while every core log reaches persistent storage and \`LogSink\`. ...
RUN v4.1.8 /home/byk/Code/opencode-lore-embedding-runtime
Test Files 2 passed (2)
Tests 13 passed (13)
Start at 09:37:10
Duration 2.89s (transform 1.34s, setup 1.00s, import 854ms, tests 3.33s, environment 0ms)
Working copy changes: A packages/core/src/cgroup-memory.ts M packages/core/src/embedding/local.ts M packages/core/src/embedding/pool.ts M packages/core/src/embedding-cap.ts M packages/core/src/embedding-worker-types.ts M packages/core/src/embedding-worker.ts M packages/core/src/embedding.ts M packages/core/src/ort-native.ts A packages/core/test/cgroup-memory.test.ts M packages/core/test/embedding-cap.test.ts M packages/core/test/embedding-pool-memory.test.ts M packages/core/test/embedding-pool.test.ts M packages/core/test/embedding-wasm-fallback.test.ts M packages/core/test/embedding-worker-stdio.test.ts M packages/core/test/embedding-worker-types.test.ts M packages/core/test/ort-native.test.ts M packages/gateway/script/bundle.ts M packages/gateway/script/ort-platform-package.ts M packages/gateway/test/bundle-exports.test.ts M packages/gateway/test/ort-platform-package.test.ts Working copy (@) : tmzmtlss 6a5b70c9 fix/embedding-runtime-admission | fix(core): correct embedding runtime ...
tmzmtlsspstn 6a5b70c9ee77ac9387a28be6865a1bb3d21cd577 parents=2cbf40060c67df229731cf1b9e34d5bf22ec2145 conflict=false empty=false pxrvwsvvkmpo 2cbf40060c67df229731cf1b9e34d5bf22ec2145 parents=e977b4b1da8f8197c3238466d41de0bccd8eb054 conflict=false empty=false
ab8a298011cb88a89aa2ac41f70ad718afc296605e867007d1c298f679f43b5f -
(no output)
packages/core/src/cgroup-memory.ts
a781cff720fd7624addc687a1830db1cbebde9153c10bd961dc08ecbe374b9f7 -
RUN v4.1.8 /home/byk/Code/opencode-lore-embedding-runtime
❯ packages/core/test/embedding-cap.test.ts (62 tests | 1 failed) 66ms
× always accepts finite Node headroom without a hard limit 16ms
❯ packages/core/test/embedding-pool-memory.test.ts (5 tests | 1 failed) 1001ms
× uses the token floor when cgroup accounting fails without a hard limit 11ms
Test Files 2 failed | 1 passed (3)
Tests 2 failed | 122 passed (124)
Start at 09:39:48
Duration 2.91s (transform 1.52s, setup 1.32s, import 1.45s, tests 2.54s, environment 0ms)
⎯⎯⎯⎯⎯⎯⎯ Failed Tests 2 ⎯⎯⎯⎯⎯⎯⎯
FAIL packages/core/test/embedding-cap.test.ts > resolveMemoryHeadroom > always accepts finite Node headroom without a hard limit
AssertionError: expected undefined to be 7516192768 // Object.is equality
- Expected:
7516192768
+ Received:
undefined
❯ packages/core/test/embedding-cap.test.ts:102:57
100|
101| it("always accepts finite Node headroom without a hard limit", () =>…
102| ...
false false false
{"reprobe":3000,"infiniteFree":null}
freeFinite=false backoffFinite=false backoffNaN=true
256
6a5b70c9ee77ac9387a28be6865a1bb3d21cd577 parent=2cbf40060c67df229731cf1b9e34d5bf22ec2145 conflict=false a781cff720fd7624addc687a1830db1cbebde9153c10bd961dc08ecbe374b9f7 - packages/core/src/cgroup-memory.ts packages/core/src/embedding/local.ts packages/core/src/embedding/pool.ts packages/core/src/embedding-cap.ts packages/core/src/embedding-worker-types.ts packages/core/src/embedding-worker.ts packages/core/src/embedding.ts packages/core/src/ort-native.ts packages/core/test/cgroup-memory.test.ts packages/core/test/embedding-cap.test.ts packages/core/test/embedding-pool-memory.test.ts packages/core/test/embedding-pool.test.ts packages/core/test/embedding-wasm-fallback.test.ts packages/core/test/embedding-worker-stdio.test.ts packages/core/test/embedding-worker-types.test.ts packages/core/test/ort-native.test.ts packages/gateway/script/bundle.ts packages/gateway/script/ort-platform-package.ts packages/gateway/test/bundle-exports.test.ts packages/gateway/test/ort-platform-package.test.ts
6a5b70c9ee77ac9387a28be6865a1bb3d21cd577 parent=2cbf40060c67df229731cf1b9e34d5bf22ec2145 conflict=false a781cff720fd7624addc687a1830db1cbebde9153c10bd961dc08ecbe374b9f7 -
Audited tmzmtlss at 6a5b70c9ee77, parent 2cbf40060c67. I did not edit files or mutate VCS.
Unknown headroom still fails open to host memory
constrainedMemoryLimit() uses 0 for both unconstrained and unknown states (packages/core/src/embedding/local.ts:86-97). When no finite hard limit or headroom exists, clampFreeToContainerLimit() returns hostFree (packages/core/src/embedding-cap.ts:165-173). The pool only blocks unknown headroom when constrained > 0 (packages/core/src/embedding/pool.ts:282-287).
A failed soft-limit sample with constrained === 0 therefore gives the worker the host-sized cap and permits pool growth. The candidate’s own regression proves this: packages/core/test/embedding-pool-memory.test.ts:164-176 expected 256, but received 4948.
Fix: return a tagged sample: known, known-unconstrained, or unknown. Only the explicit unconstrained state may use host memory; unknown must become zero headroom. Add a pool-growth regression for unknown soft-limit accounting, not only a payload-cap assertion.
Pool admission can double-spend memory during OOM recovery
A slot starts unhealthy and becomes healthy after one success (packages/core/src/embedding/pool.ts:290-315, :550-575), but it never becomes unhealthy again. Growth reserves memory only for !slot.healthy slots (:262-269).
An OOM replaces the worker inside the same provider without notifying the pool (packages/core/src/embedding/local.ts:720-727, :876-919). After the old worker exits, sampled headroom rises temporarily. A concurrent request can spend that headroom on a new pool slot while the replacement worker reloads its own model.
Fix: track current worker-generation residency separately from historical health. Reserve a full budget from exit until the replacement completes a real embed. ...
Date: Sep 17, 2026 * 🔴 (09:34) User stated `packages/core/test/embedding-pool.test.ts` exercises cross-worker dispatch via `_setTestWorkerFactory`, including least-busy routing, lazy/memory-gated growth, the broken-provider latch, and pool shutdown; the per-worker OOM/self-heal lifecycle belongs exclusively in `packages/core/test/embedding-oom-recovery.test.ts`, and `embedding-pool.test.ts` neve…
Date: Sep 17, 2026 * 🟡 (09:32) Search found 15 references to `backoffEmbedCap`: `packages/core/test/embedding-pool-memory.test.ts` lines 19 and 241; `packages/core/test/embedding-cap.test.ts` lines 13, 177, 179, 180, 184, 186, 193, and 201; `packages/core/test/embedding-oom-recovery.test.ts` lines 18 and 143; `packages/core/src/embedding/local.ts` lines 12 and 846; and its definition in `package…
Date: Sep 17, 2026 * 🔴 (09:29) User asserted `resolveMemoryHeadroom` always accepts finite Node headroom without a hard limit, including `resolveMemoryHeadroom(0, undefined, 7 * GB) === 7 * GB` and finite `7 * GB` Node headroom when host/cgroup inputs are `Number.NaN`. * 🔴 (09:29) User asserted `clampFreeToContainerLimit` never raises the figure; it is monotonic and can only reduce memory use. …
Date: Sep 17, 2026 * 🔴 (09:27) User asserted `resolveMemoryHeadroom` always accepts finite Node headroom without a hard limit. * 🔴 (09:27) User asserted a host logger/sink must never interrupt stream consumption. * 🔴 (09:27) User asserted embedding-pool behavior must never multiply memory pressure on a constrained host; each added local worker carries a full model/runtime baseline that embeddi…
Date: Sep 17, 2026 * 🔴 (09:26) User asserted the adaptive embedding cap never exceeds the Nomic v1.5 maximum sequence length. * 🔴 (09:26) User asserted that when free memory never improves, embedding-cap recovery never re-probes upward. * 🔴 (09:26) User asserted that after rebooting with more memory, the system never re-probes above a cap imposed by the WASM heap. * 🔴 (09:26) User asserted em…
Date: Sep 17, 2026 * 🔴 (09:26) User asserted the embedding pool’s primary worker is always present; it has its own OOM backoff, separate from pool sizing and replacement-worker admission logic. * 🔴 (09:26) User asserted invalid `LORE_EMBED_POOL_SIZE` values must resolve to `undefined`, never `NaN`; `_configuredEmbedPoolSize()` in `packages/core/src/embedding/pool.ts` is a test seam for validati…
Date: Sep 17, 2026 * 🔴 (09:25) User asserted the project’s TDD rule: “Never trust the fix without seeing it fail.” Every bug fix must start with a deterministic regression test that fails on the base branch and passes on the fix. * 🔴 (09:25) Project regression tests must drive the real precondition; tests of a guard must fail if that guard is deleted. * 🔴 (09:25) Project adversarial-order test…
Date: Sep 17, 2026 * 🔴 [requested-review] (09:22) User requested a read-only audit of change `tmzmtlss` in `/home/byk/Code/opencode-lore-embedding-runtime`, focused on embedding admission/headroom sampling outside `cgroup-memory.ts`, specifically `packages/core/src/embedding/local.ts`, `packages/core/src/embedding/pool.ts`, `packages/core/src/embedding-cap.ts`, and related tests. * 🔴 (09:22) Us…