Merge pull request #2489 from deepseek-harness/fix/publish-retry-idempotence

fix(release): make publication retry, space out, and skip what already landed

.agents/notes/implemented/process/2026-08-10-npm-release-sequences.i18n.yaml

@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write .agents/notes/implemented/process/2026-08-10-npm-release-sequences.md
-2026-08-10-npm-release-sequences.md: e8138aef923e201cc0883232f48ce6921452ca00
-2026-08-10-npm-release-sequences.zh.md: 7ce9fa01dbe10cbdae5585f64392d0ca6a9fb868
+2026-08-10-npm-release-sequences.md: e74a4ac8f2aadd8665ec0db198c6a317a0c201bc
+2026-08-10-npm-release-sequences.zh.md: e152163976f945224f2524fccd7f831ba98e8161

.agents/notes/implemented/process/2026-08-10-npm-release-sequences.md

@@ -70,6 +70,10 @@ Publication runs only from GitHub Actions; there is no local publication path. P
 
 The third state catches code that changed without a version bump. The first two provide idempotence — re-running publish over one artifact republishes nothing and needs no manual selection of packages. The same rule resolves the tension between one vendor release carrying several tags and a workflow that can only run from one ref: the workflow never infers which packages to publish from the tag it ran from.
 
+All three sequences decide this way, including the native one: it publishes through its own script rather than a shell loop, because a loop of bare `npm publish` calls cannot be retried — the registry answers a repeat of an existing version permanently, so one failure partway through left no way forward.
+
+Two registry behaviours shape how a publish is attempted. Writes are spaced by at least two seconds and retried with a backoff, because publishing several packages back to back outruns the registry's own processing and earns `E409 Failed to save packument`. And every retry re-reads the registry first: a reported failure can answer a write that landed anyway, so a version that now exists with this tarball's integrity counts as published rather than as a version to place again.
+
 ### Workspace-internal references use the `workspace:` protocol
 
 Every reference to a workspace member uses `workspace:^`, so `pnpm pack` substitutes a range matching the target version: sibling `peerDependencies` follow the family version, and a reference to a vendored package follows that package's own line. The Landlock platform packages keep `workspace:*`, which publishes the exact version, because a platform package and its entry must agree exactly.

.agents/notes/implemented/process/2026-08-10-npm-release-sequences.zh.md

@@ -70,6 +70,10 @@ tag 只是 commit 指针，不是发布成功的证明。bump 会向 registry
 
 第三态拦住「改了代码却没 bump 版本」。前两态给出幂等——同一个 artifact 重跑 publish 不会重复发布，也不需要人工挑拣包。同一条规则还解决了「一次 vendor 发布携带多个 tag，而 workflow 只能从一个 ref 触发」的矛盾：workflow 从不从触发它的 tag 去推断该发哪些包。
 
+三条序列都按这套判定，native 也在内：它通过自己的脚本发布，而不是 shell 循环——一串裸 `npm publish` 无法重试，registry 对「重发已存在的版本」的回答是永久失败，因此中途失败一次就没有前路了。
+
+registry 的两个行为决定了「怎么尝试一次发布」。写入之间至少间隔两秒并带退避重试，因为连续背靠背发多个包会超出 registry 自身的处理速度，换来 `E409 Failed to save packument`。而每次重试都先重查 registry：报出来的失败可能对应一次其实已经落地的写入，所以「该版本现在存在且 integrity 与本 tarball 相同」算作已发布，而不是又一个待放置的版本。
+
 ### workspace 内部引用走 `workspace:` 协议
 
 所有指向 workspace 成员的引用都用 `workspace:^`，由 `pnpm pack` 替换成匹配目标版本的范围：兄弟包的 `peerDependencies` 跟随族版本，指向 vendored 包的引用跟随那个包自己的版本线。Landlock 平台包保留 `workspace:*`（发布成精确版本），因为平台包与它的入口必须版本完全一致。

.github/workflows/landlock-run-release.yml

@@ -147,6 +147,12 @@ jobs:
       contents: read
       id-token: write
     steps:
+      # The publish script is the only repository file this job needs, and it
+      # imports nothing outside Node's builtins, so there is no install step.
+      - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+
       - uses: actions/setup-node@v4
         with:
           node-version: 24
@@ -167,12 +173,8 @@ jobs:
           fi
 
       - name: Publish tarballs
-        run: |
-          version="${GITHUB_REF#refs/tags/landlock-run-v}"
-          tag_args=()
-          case "$version" in *-*) tag_args=(--tag next);; esac
-          while IFS= read -r tarball; do
-            # No --access: publishConfig.access in each manifest decides, and a
-            # command-line flag would override it.
-            npm publish "dist/npm/${tarball}" "${tag_args[@]}"
-          done < dist/npm/publish-order.txt
+        # Publication is decided per package against the registry, so re-running
+        # this job over the same artifact skips what already landed instead of
+        # failing on it. A bare `npm publish` loop could not be retried: the
+        # registry answers a repeat of an existing version permanently.
+        run: node ./scripts/publish-release.mjs dist/npm

knip.json

@@ -9,6 +9,7 @@
     "musl-gcc",
     "python3",
     "sandbox-exec",
+    "tar",
     "taskkill",
     "where.exe"
   ],

native/landlock-run/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@deepseek-ai/node-addon-landlock-run-workspace",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "private": true,
   "type": "module",
   "license": "BSD-3-Clause",
@@ -19,6 +19,7 @@
     "release:assemble-prebuilds": "node ./scripts/assemble-prebuilds.mjs",
     "release:verify": "node ./scripts/verify-release.mjs",
     "release:pack": "node ./scripts/pack-release.mjs",
+    "release:publish": "node ./scripts/publish-release.mjs",
     "release:verify-packed-install": "node ./scripts/verify-packed-install.mjs"
   },
   "devDependencies": {

native/landlock-run/packages/entry/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@deepseek-ai/node-addon-landlock-run",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "type": "module",
   "description": "Landlock self-restrict-then-exec launcher for sandboxing subprocesses on Linux: per-platform prebuilt static binaries plus the JS seam that resolves, probes, and speaks their CLI contract",
   "repository": {

native/landlock-run/packages/linux-arm64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@deepseek-ai/node-addon-landlock-run-linux-arm64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Prebuilt landlock-run Landlock launcher binary for linux-arm64 (static musl) — resolved as a file path by @deepseek-ai/node-addon-landlock-run, never imported",
   "repository": {
     "type": "git",

native/landlock-run/packages/linux-x64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@deepseek-ai/node-addon-landlock-run-linux-x64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Prebuilt landlock-run Landlock launcher binary for linux-x64 (static musl) — resolved as a file path by @deepseek-ai/node-addon-landlock-run, never imported",
   "repository": {
     "type": "git",

native/landlock-run/scripts/publish-release.mjs

@@ -0,0 +1,165 @@
+#!/usr/bin/env node
+/**
+ * Publish the packed launcher family from the tarballs `pack-release.mjs`
+ * produced, in `publish-order.txt` order.
+ *
+ * What goes out is decided per package against the registry, never from the
+ * order file alone: a version the registry lacks is published, a version whose
+ * published tarball has the same integrity is skipped, and a version whose
+ * published tarball differs fails the run — that last case means the content
+ * changed without a version bump. Skipping on identical integrity is what makes
+ * re-running the publish step over the same artifact safe, which matters here
+ * because a partial publication used to leave no way forward: republishing an
+ * existing version fails permanently.
+ *
+ * Usage: `node scripts/publish-release.mjs [packed dir]`.
+ */
+
+import fs from 'node:fs';
+import path from 'node:path';
+import crypto from 'node:crypto';
+import { spawnSync } from 'node:child_process';
+import { setTimeout as sleep } from 'node:timers/promises';
+import { root } from './repo.mjs';
+
+/**
+ * Registry codes that answer a write which did not settle, rather than a
+ * rejection of what was sent. `E409 Failed to save packument` is the one this
+ * sequence actually hits: publishing the platform packages and the entry back
+ * to back can outrun the registry's own processing. A rejected payload (`E403`
+ * over an existing version, a malformed manifest) never clears on a retry.
+ */
+const TRANSIENT_PUBLISH_CODES = ['E409', 'E429', 'E500', 'E502', 'E503', 'E504', 'ETIMEDOUT', 'ECONNRESET', 'EAI_AGAIN'];
+
+/** How many times one tarball's publish is attempted before the run fails. */
+const PUBLISH_ATTEMPTS = 4;
+
+/**
+ * Shortest gap between two publishes, and the first retry backoff. The registry
+ * needs a moment to commit a packument before the next write; back to back
+ * publishes are what produce `E409`.
+ */
+const PUBLISH_SPACING_MS = 2_000;
+
+const destination = path.resolve(process.argv.slice(2).find((arg) => !arg.startsWith('--')) || path.join(root, 'dist', 'npm'));
+
+/**
+ * @param {string} output Combined npm output.
+ * @returns {boolean} True when the registry reported a write it did not commit.
+ */
+function isTransientFailure(output) {
+  return TRANSIENT_PUBLISH_CODES.some((code) => output.includes(`code ${code}`));
+}
+
+/**
+ * @param {string} tarball Absolute tarball path.
+ * @returns {string} The `sha512-<base64>` integrity npm records for it.
+ */
+function integrityOf(tarball) {
+  return `sha512-${crypto.createHash('sha512').update(fs.readFileSync(tarball)).digest('base64')}`;
+}
+
+/**
+ * @param {string} tarball Absolute tarball path.
+ * @returns {{name: string, version: string}} What the packed manifest declares.
+ */
+function packedIdentity(tarball) {
+  const result = spawnSync('tar', ['-xOzf', tarball, 'package/package.json'], { encoding: 'utf8' });
+  if (result.status !== 0) throw new Error(`cannot read the manifest inside ${tarball}:\n${result.stderr}`);
+  const manifest = JSON.parse(result.stdout);
+  if (typeof manifest.name !== 'string' || typeof manifest.version !== 'string') {
+    throw new Error(`${tarball} manifest lacks name/version`);
+  }
+  return { name: manifest.name, version: manifest.version };
+}
+
+/**
+ * Ask the registry whether a version exists, and with what integrity.
+ * @param {string} name Package name.
+ * @param {string} version Package version.
+ * @returns {{kind: 'absent'} | {kind: 'present', integrity: string}} Registry state.
+ */
+function registryState(name, version) {
+  const result = spawnSync('npm', ['view', `${name}@${version}`, 'dist.integrity', '--json'], { encoding: 'utf8' });
+  if (result.status !== 0) {
+    const output = `${result.stdout}${result.stderr}`;
+    if (output.includes('E404') || output.includes('404 Not Found')) return { kind: 'absent' };
+    throw new Error(`npm view ${name}@${version} failed:\n${output}`);
+  }
+  const parsed = JSON.parse(result.stdout);
+  if (typeof parsed !== 'string' || parsed === '') {
+    throw new Error(`registry reported no dist.integrity for ${name}@${version}`);
+  }
+  return { kind: 'present', integrity: parsed };
+}
+
+/**
+ * Publish one tarball, retrying a registry write that did not settle.
+ *
+ * Every retry re-reads the registry first, because `E409` can answer a write
+ * that landed anyway: republishing a version that now exists fails permanently,
+ * so the same integrity appearing under the failed attempt counts as success.
+ * @param {string} tarball Absolute tarball path.
+ * @param {string} name Package name the tarball declares.
+ * @param {string} version Package version the tarball declares.
+ */
+async function publishTarball(tarball, name, version) {
+  // A prerelease version never takes the latest dist-tag.
+  const tagArgs = version.includes('-') ? ['--tag', 'next'] : [];
+  for (let tries = 1; tries <= PUBLISH_ATTEMPTS; tries += 1) {
+    // No --access: publishConfig.access in each manifest decides, and a
+    // command-line flag would override it.
+    const result = spawnSync('npm', ['publish', tarball, ...tagArgs], { encoding: 'utf8' });
+    const output = `${result.stdout}${result.stderr}`;
+    if (result.status === 0) return;
+
+    const settled = registryState(name, version);
+    if (settled.kind === 'present' && settled.integrity === integrityOf(tarball)) {
+      console.log(`landlock publish: ${name}@${version} landed despite a reported failure, continuing`);
+      return;
+    }
+    if (tries === PUBLISH_ATTEMPTS || !isTransientFailure(output)) {
+      throw new Error(`npm publish ${name}@${version} failed:\n${output}`);
+    }
+    const backoff = PUBLISH_SPACING_MS * 2 ** (tries - 1);
+    console.log(
+      `landlock publish: ${name}@${version} hit a transient registry failure`
+      + ` (attempt ${tries} of ${PUBLISH_ATTEMPTS}), retrying in ${backoff}ms`,
+    );
+    await sleep(backoff);
+  }
+}
+
+const order = fs
+  .readFileSync(path.join(destination, 'publish-order.txt'), 'utf8')
+  .split('\n')
+  .filter((line) => line !== '');
+
+let published = 0;
+let skipped = 0;
+for (const filename of order) {
+  const tarball = path.join(destination, filename);
+  const { name, version } = packedIdentity(tarball);
+  const state = registryState(name, version);
+  if (state.kind === 'present') {
+    const local = integrityOf(tarball);
+    if (state.integrity !== local) {
+      throw new Error(
+        `${name}@${version} is already published with different content`
+        + `\n  registry: ${state.integrity}\n  packed:   ${local}`
+        + '\nBump the version, or investigate why the build is not reproducible.',
+      );
+    }
+    console.log(`landlock publish: ${name}@${version} already published, skipping`);
+    skipped += 1;
+    continue;
+  }
+  // Space out the writes: the gap belongs between publishes, so a run that only
+  // skips does not wait at all.
+  if (published > 0) await sleep(PUBLISH_SPACING_MS);
+  await publishTarball(tarball, name, version);
+  console.log(`landlock publish: ${name}@${version} published`);
+  published += 1;
+}
+
+console.log(`landlock publish: ${published} published, ${skipped} already present`);

scripts/release/publish.ts

@@ -15,16 +15,46 @@
 import { createHash } from 'node:crypto'
 import { readFileSync } from 'node:fs'
 import { join, resolve } from 'node:path'
+import { setTimeout as sleep } from 'node:timers/promises'
 import { parseArgs } from 'node:util'
 import { releaseFamily } from './families.ts'
-import { attempt, isEntry, run } from './process.ts'
+import { attempt, isEntry } from './process.ts'
 import { packedIdentity, readPublishOrder } from './tarball.ts'
 
+/**
+ * Registry codes that answer a write which did not settle, rather than a
+ * rejection of what was sent. `E409 Failed to save packument` is the one this
+ * sequence actually hits: publishing several packages in a row can outrun the
+ * registry's own processing. A rejected payload (`E403` over an existing
+ * version, a malformed manifest) never clears on a retry and must surface.
+ */
+const TRANSIENT_PUBLISH_CODES = ['E409', 'E429', 'E500', 'E502', 'E503', 'E504', 'ETIMEDOUT', 'ECONNRESET', 'EAI_AGAIN'] as const
+
+/** How many times one tarball's publish is attempted before the run fails. */
+const PUBLISH_ATTEMPTS = 4
+
+/**
+ * Shortest gap between two publishes, and the first retry backoff.
+ *
+ * The registry needs a moment to commit a packument before the next write; back
+ * to back publishes are what produce `E409`.
+ */
+const PUBLISH_SPACING_MS = 2_000
+
 /** What the registry knows about one version. */
 type RegistryState =
   | { readonly kind: 'absent' }
   | { readonly kind: 'present'; readonly integrity: string }
 
+/**
+ * Whether a failed publish is worth another attempt.
+ * @param output - combined npm output.
+ * @returns True when the registry reported a write it did not commit.
+ */
+function isTransientFailure(output: string): boolean {
+  return TRANSIENT_PUBLISH_CODES.some(code => output.includes(`code ${code}`))
+}
+
 /**
  * The subresource integrity string npm records for a tarball.
  * @param tarball - absolute tarball path.
@@ -54,8 +84,47 @@ function registryState(name: string, version: string): RegistryState {
   return { kind: 'present', integrity: parsed }
 }
 
+/**
+ * Publish one tarball, retrying a registry write that did not settle.
+ *
+ * Every retry re-reads the registry first, because `E409` can answer a write
+ * that landed anyway: republishing a version that now exists fails permanently,
+ * so the same integrity appearing under the failed attempt counts as success.
+ * @param tarball - absolute tarball path.
+ * @param name - package name the tarball declares.
+ * @param version - package version the tarball declares.
+ */
+async function publishTarball(tarball: string, name: string, version: string): Promise<void> {
+  // A prerelease version never takes the latest dist-tag.
+  const tagArgs = version.includes('-') ? ['--tag', 'next'] : []
+  for (let tries = 1; tries <= PUBLISH_ATTEMPTS; tries += 1) {
+    // No --access: the sequences do not share one access level, so a
+    // command-line flag could not serve both and would override the manifest
+    // that does. Each packed manifest decides, and
+    // check-workspace-constraints holds every manifest to its sequence's level.
+    const result = attempt('npm', ['publish', tarball, ...tagArgs])
+    const output = `${result.stdout}${result.stderr}`
+    if (result.status === 0) return
+
+    const settled = registryState(name, version)
+    if (settled.kind === 'present' && settled.integrity === integrityOf(tarball)) {
+      console.log(`release publish: ${name}@${version} landed despite a reported failure, continuing`)
+      return
+    }
+    if (tries === PUBLISH_ATTEMPTS || !isTransientFailure(output)) {
+      throw new Error(`npm publish ${name}@${version} failed:\n${output}`)
+    }
+    const backoff = PUBLISH_SPACING_MS * 2 ** (tries - 1)
+    console.log(
+      `release publish: ${name}@${version} hit a transient registry failure`
+      + ` (attempt ${String(tries)} of ${String(PUBLISH_ATTEMPTS)}), retrying in ${String(backoff)}ms`,
+    )
+    await sleep(backoff)
+  }
+}
+
 /** Publish the family named by `--family` from the directory named by `--from`. */
-function main(): void {
+async function main(): Promise<void> {
   const { values } = parseArgs({
     options: { family: { type: 'string' }, from: { type: 'string' } },
     allowPositionals: false,
@@ -86,17 +155,15 @@ function main(): void {
       skipped += 1
       continue
     }
-    // A prerelease version never takes the latest dist-tag.
-    const tagArgs = version.includes('-') ? ['--tag', 'next'] : []
-    // No --access: the sequences do not share one access level, so a
-    // command-line flag could not serve both and would override the manifest
-    // that does. Each packed manifest decides, and
-    // check-workspace-constraints holds every manifest to its sequence's level.
-    run('npm', ['publish', tarball, ...tagArgs])
+    // Space out the writes: the gap belongs between publishes, so a run that
+    // only skips does not wait at all.
+    if (published > 0) await sleep(PUBLISH_SPACING_MS)
+    await publishTarball(tarball, name, version)
+    console.log(`release publish: ${name}@${version} published`)
     published += 1
   }
 
   console.log(`release publish: family ${family.id}, ${String(published)} published, ${String(skipped)} already present`)
 }
 
-if (isEntry(import.meta.url)) main()
+if (isEntry(import.meta.url)) await main()