Skip to content

Fix Docker store seed target packages#91547

Merged
sallyom merged 1 commit into
openclaw:mainfrom
sallyom:codex/docker-store-seed-targets
Jun 9, 2026
Merged

Fix Docker store seed target packages#91547
sallyom merged 1 commit into
openclaw:mainfrom
sallyom:codex/docker-store-seed-targets

Conversation

@sallyom

@sallyom sallyom commented Jun 9, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Summary

Fixes #91035.

The Docker runtime-assets store seed now follows the current build target instead of adding every lockfile package. This keeps platform optional packages, such as @zed-industries/codex-acp-*, limited to the architecture Docker is actually building for.

Root Cause

pnpm install was already configured with Docker target architecture filters, but the later offline prune store seed used scripts/list-prod-store-packages.mjs to add every package from pnpm-lock.yaml. That explicitly reintroduced non-target optional packages before pnpm store add, including Darwin/Windows and the wrong Linux architecture ACP packages.

Changes

  • Filter pnpm list --prod and lockfile snapshot closure through os/cpu/libc selectors for the current process target.
  • Keep snapshot dependency closure, including target-matching optional dependencies, so the offline prune store still has the packages it needs.
  • Handle pnpm peer-resolved snapshot keys when walking snapshot dependencies.
  • Add focused tests for target platform optional dependencies and peer-resolved snapshot closures.

Verification

  • pnpm lint --threads=8
  • node scripts/run-vitest.mjs test/scripts/list-prod-store-packages.test.ts
  • git diff --check
  • .agents/skills/autoreview/scripts/autoreview --mode local
  • Local Docker target probe, OPENCLAW_EXTENSIONS=acpx, native linux/arm64: verified the store seed includes @zed-industries/codex-acp-linux-arm64@0.15.0, excludes Darwin/Windows/Linux x64 ACP packages, and pnpm store add succeeds.
  • Local Docker target probe, OPENCLAW_EXTENSIONS=acpx, --platform=linux/amd64: verified process.arch is x64, the store seed includes @zed-industries/codex-acp-linux-x64@0.15.0, excludes Darwin/Windows/Linux arm64 ACP packages, and pnpm store add succeeds.
  • Testbox-through-Crabbox full Docker build: provider blacksmith-testbox, id tbx_01ktn3vp481ggkcwhh0wfnt8t6, Actions run https://github.com/openclaw/openclaw/actions/runs/27180308958, command docker build --progress=plain --target runtime-assets --build-arg OPENCLAW_EXTENSIONS=acpx -t openclaw:pr91547-runtime-assets -f Dockerfile ., HEAD ec9f55e742ee428f631c8f6e591c9eaab265ffee, exit 0. The runtime-assets stage completed pnpm list | scripts/list-prod-store-packages.mjs | pnpm store add, offline pnpm prune --prod, plugin dist prune, postinstall, and package dist import checks, then exported image sha256:dcb33fbedf7480f2c7b559468875f53c3f63243ddd3d1b44872d475e4a5d4a00.

Real behavior proof

Behavior addressed: Docker builds with ACPX no longer seed non-target optional platform packages from the pnpm lockfile into the offline production store. The runtime-assets stage now keeps the target Linux architecture package and avoids Darwin, Windows, and wrong-Linux-architecture ACP packages.

Real environment tested: Blacksmith Testbox through Crabbox on Linux amd64 Docker Engine 28.0.4, provider blacksmith-testbox, id tbx_01ktn3vp481ggkcwhh0wfnt8t6, Actions run https://github.com/openclaw/openclaw/actions/runs/27180308958, PR commit ec9f55e742ee428f631c8f6e591c9eaab265ffee.

Exact steps or command run after this patch: docker build --progress=plain --target runtime-assets --build-arg OPENCLAW_EXTENSIONS=acpx -t openclaw:pr91547-runtime-assets -f Dockerfile .

Evidence after fix: The Testbox run exited 0. Docker completed the runtime-assets stage, including pnpm list --prod --depth Infinity --json | node scripts/list-prod-store-packages.mjs | xargs -r pnpm store add, offline pnpm prune --prod, plugin dist prune, postinstall, package dist import checks, and image export.

Observed result after fix: The runtime-assets build succeeded and exported image sha256:dcb33fbedf7480f2c7b559468875f53c3f63243ddd3d1b44872d475e4a5d4a00. The store seed step completed without trying to fetch incompatible ACP platform tarballs.

What was not tested: A full final runtime image smoke was not run here; this proof targeted the failing Docker runtime-assets store/prune path from #91035. Local Docker Desktop full build was attempted before Testbox proof but hit an out-of-memory SIGKILL during pnpm build:docker, before the patched runtime-assets step.

Notes

The temporary openclaw/openclaw:codex/docker-store-seed-targets branch was used only because Blacksmith workflow dispatch requires a base-repository ref. It was deleted after the passing Testbox run; the PR head remains sallyom:codex/docker-store-seed-targets.

@openclaw-barnacle openclaw-barnacle Bot added scripts Repository scripts size: M maintainer Maintainer-authored PR labels Jun 9, 2026
@clawsweeper

clawsweeper Bot commented Jun 9, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Codex review: needs maintainer review before merge. Reviewed June 8, 2026, 10:45 PM ET / 02:45 UTC.

Summary
The PR updates scripts/list-prod-store-packages.mjs and its tests so Docker runtime-assets store seeding keeps only target-matching production packages and target optional snapshot dependencies.

PR surface: Tests +121, Other +52. Total +173 across 2 files.

Reproducibility: yes. Source inspection on current main shows Docker runtime-assets feeds pnpm store add from a script that also adds every lockfile package, while the lockfile records non-target @zed-industries/codex-acp optional packages matching the linked failure logs.

Review metrics: 1 noteworthy metric.

  • Proof/head delta: 1 helper-only loop rewrite after Docker proof. The real Docker proof ran before the final force-push, so maintainers should notice the small current-head delta before relying on that proof.

Merge readiness
Overall: 🦞 diamond lobster
Proof: 🦞 diamond lobster ✨ media proof bonus
Patch quality: 🦞 diamond lobster
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Risk before merge

  • [P1] The store seed feeds Docker offline prune; if the target filter or snapshot closure misses a real production dependency, users would see it as a Docker runtime-assets build failure.
  • [P1] The strongest Docker proof ran on 90349eef56d3, while current head ec9f55e742ee only rewrites dependency iteration; maintainers may still want exact-head Docker proof or a final runtime image smoke before merge.

Maintainer options:

  1. Accept Targeted Runtime-Assets Proof (recommended)
    Maintainers can land after exact-head CI and mergeability if they accept the Testbox runtime-assets build plus the small post-proof helper-only delta as covering the reported ACPX store/prune failure.
  2. Require Exact-Head Final Image Smoke
    Ask for a fresh Docker proof on ec9f55e742ee or a full final runtime image smoke if maintainers want assurance beyond the patched runtime-assets stage.

Next step before merge

  • No automated repair is queued because no concrete patch defect remains; the remaining action is maintainer acceptance of targeted runtime-assets proof for a protected, availability-sensitive Docker build path.

Security
Cleared: The diff changes a local package-listing script and tests only; it does not add dependencies, workflow permissions, secrets handling, action refs, or new executable supply-chain sources.

Review details

Best possible solution:

Merge the target-aware store seed after maintainers accept the targeted Testbox runtime-assets proof and exact-head CI; run a final runtime image smoke only if broader container assurance is required.

Do we have a high-confidence way to reproduce the issue?

Yes. Source inspection on current main shows Docker runtime-assets feeds pnpm store add from a script that also adds every lockfile package, while the lockfile records non-target @zed-industries/codex-acp optional packages matching the linked failure logs.

Is this the best way to solve the issue?

Yes. Filtering the existing store-seed closure by lockfile os/cpu/libc selectors is the narrow maintainable fix and avoids ACPX-specific package hardcoding.

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 329fa44d23f4.

Label changes

Label justifications:

  • P2: This is a focused fix for a Docker build regression with limited blast radius to container runtime-assets package seeding.
  • merge-risk: 🚨 availability: The PR changes which packages are seeded before Docker offline prune, so a missed target or closure case could make container builds fail.
  • rating: 🦞 diamond lobster: Overall readiness is 🦞 diamond lobster; proof is 🦞 diamond lobster and patch quality is 🦞 diamond lobster.
  • status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (linked_artifact): Sufficient linked-artifact proof is present: the PR body reports a Testbox-through-Crabbox Docker runtime-assets build exit 0, and the current-head change after that proof is a small dependency-iteration refactor with the same behavior.
  • proof: sufficient: Contributor real behavior proof is sufficient. Sufficient linked-artifact proof is present: the PR body reports a Testbox-through-Crabbox Docker runtime-assets build exit 0, and the current-head change after that proof is a small dependency-iteration refactor with the same behavior.
Evidence reviewed

PR surface:

Tests +121, Other +52. Total +173 across 2 files.

View PR surface stats
Area Files Added Removed Net
Source 0 0 0 0
Tests 1 123 2 +121
Docs 0 0 0 0
Config 0 0 0 0
Generated 0 0 0 0
Other 1 77 25 +52
Total 2 200 27 +173

What I checked:

  • Repository policy read: Root AGENTS.md plus scoped scripts/test/extension policy were read; the review applied the deep PR review, dependency-contract, protected-label, and script-wrapper guidance. (AGENTS.md:9, 329fa44d23f4)
  • Current Docker entry point: Current main pipes pnpm list --prod --depth Infinity --json through scripts/list-prod-store-packages.mjs into pnpm store add before offline pnpm prune --prod with Linux target selectors. (Dockerfile:132, 329fa44d23f4)
  • Current main failure surface: Current main's list script still adds every package key from pnpm-lock.yaml, so non-target optional platform packages can be fed to pnpm store add. (scripts/list-prod-store-packages.mjs:55, 329fa44d23f4)
  • Shipped behavior check: The latest release tag v2026.6.1 contains the same runtime-assets store-seed path and old add-all-lockfile-package script shape, so the PR is not obsolete on the latest release. (scripts/list-prod-store-packages.mjs:55, 2e08f0f4221f)
  • ACP optional package shape: The workspace lockfile records @zed-industries/codex-acp plus Darwin, Linux, and Windows optional platform packages, and the snapshot closure lists them as optional dependencies. (pnpm-lock.yaml:4274, 329fa44d23f4)
  • PR implementation: The PR reads target os/cpu/libc, filters package specs through lockfile selectors, resolves peer-decorated snapshot keys, and includes both normal and optional snapshot dependency closure. (scripts/list-prod-store-packages.mjs:9, ec9f55e742ee)

Likely related people:

  • @vincentkoc: Blame shows 4b55a0e04d41 introduced the current scripts/list-prod-store-packages.mjs and Docker runtime-assets seed path; recent history also ties Vincent Koc to release and runtime packaging work. (role: introduced current script path and recent release/runtime contributor; confidence: high; commits: 4b55a0e04d41, 2e08f0f4221f, 6d5e142b9312; files: scripts/list-prod-store-packages.mjs, Dockerfile, pnpm-lock.yaml)
  • @steipete: Recent history shows Docker live-lane, bundled plugin dependency, and runtime image work near the same runtime-assets and plugin dependency surfaces. (role: recent Docker and bundled plugin dependency contributor; confidence: medium; commits: ae3b10c15c64, 8e0ab35b0ed0, 17085ec1a4b5; files: Dockerfile, scripts/docker/setup.sh, extensions/acpx/package.json)
  • @sallyom: Sally O'Malley has prior merged Docker work around image size and OPENCLAW_EXTENSIONS, so this PR is not only a drive-by proposal even though she is also the PR author. (role: prior Docker feature contributor and current patch author with adjacent history; confidence: medium; commits: 57f19f0d5c10, 499c1ee6e32d, ec9f55e742ee; files: Dockerfile, scripts/docker/setup.sh, scripts/list-prod-store-packages.mjs)
What the crustacean ranks mean
  • 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
  • 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
  • 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
  • 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
  • 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
  • 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
  • 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works
  • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
  • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
  • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
  • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
  • Maintainers can also comment @clawsweeper review to request a fresh review only.
  • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
  • Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
  • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

@clawsweeper clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. P2 Normal backlog priority with limited blast radius. merge-risk: 🚨 availability 🚨 May cause crashes, hangs, restart loops, stalls, or process outages. labels Jun 9, 2026
@sallyom sallyom marked this pull request as ready for review June 9, 2026 02:02
@clawsweeper clawsweeper Bot added rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. and removed rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. labels Jun 9, 2026
@sallyom sallyom force-pushed the codex/docker-store-seed-targets branch from 90349ee to ec9f55e Compare June 9, 2026 02:36
@sallyom sallyom merged commit c8a8152 into openclaw:main Jun 9, 2026
148 of 149 checks passed
vincentkoc pushed a commit that referenced this pull request Jun 9, 2026
@sallyom @vincentkoc
fix docker store seed target packages (#91547)
0717cd5 
(cherry picked from commit c8a8152)
github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jun 9, 2026
@sallyom
fix docker store seed target packages (openclaw#91547)
638fbad
@github-actions github-actions Bot mentioned this pull request Jun 9, 2026
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

maintainer Maintainer-authored PR merge-risk: 🚨 availability 🚨 May cause crashes, hangs, restart loops, stalls, or process outages. P2 Normal backlog priority with limited blast radius. proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. scripts Repository scripts size: M status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[Bug]: Build fails on v2026.6.1

1 participant

@sallyom