Skip to content

The agent review check builds again: canon's SDK pinned at 10.0.404, the lock file regenerated, CI into the devcontainer - #85

Merged
Hafeok merged 8 commits into
mainfrom
claude/canon-lock-file-2026-11-18
Nov 18, 2026
Merged

Hafeok merged 8 commits into
mainfrom
claude/canon-lock-file-2026-11-18

Conversation

@Hafeok

@Hafeok Hafeok commented Nov 18, 2026

Copy link
Copy Markdown
Collaborator

Refs #83. Head to ratify: 1bf7e56 (1bf7e56af31c…). G2: stopped for Emil.

What this changes, and why

agent review fails on every pull request with NU1004. IsAotCompatible makes the SDK reference Microsoft.NET.ILLink.Tasks at the runtime version that SDK bundles. The lock file was written when that was 10.0.14. The runner's SDK 10.0.404 adds 10.0.15. The runner sets CI, so the restore is locked and it fails.

Option B, as ruled at G1:

File Change
tools/canon/global.json 10.0.404, rollForward: "disable"
.devcontainer/devcontainer.json dotnet feature 10.0.404
tools/canon/src/Canon/packages.lock.json regenerated once: three lines, ILLink 10.0.14 → 10.0.15 (the tests' lock file is unchanged)
.github/workflows/validate.yml CI added to devcontainers/ci's env:
tools/canon/DESIGN.md the layout line, plus a "Built" row recording the pin

agent-review.yml does not change. With rollForward: "disable", setup-dotnet installs the exact version in global.json. Nothing changes what canon checks. Passing CI closes a gap in an existing rule (Directory.Build.props locks a CI restore, but the container never saw CI); it is not a new check.

This pull request cannot pass agent review. That job checks out main and builds main's canon, and trunk has no bypass (findings D-4). Emil lifts the requirement for this one landing and restores it straight after.

Rulings and decisions

  • G1 rulings, verbatim, received 13:17 UTC: meta/sessions/2026-11-18-canon-lock-file/rulings.md.

  • No decision is filed. The waiver is drafted for Emil's ruling on its text: meta/sessions/2026-11-18-canon-lock-file/decision-draft.yaml. It validates against the decision format. The head that landed, the times lifted and restored, and the first agent review run after landing are left as <…> for Emil.

  • Decision ids were checked across every branch of both repositories:

    • here, 66 branches, highest DDD-dec-122;
    • in actor-indexed-determination, 27 branches, highest DDD-dec-39.

    The draft carries DDD-dec-123, but The act map, scope and pull: what they are, and how step 1 starts from them #80's bootstrap names 123 for a possible decision of its own (question W-1).

Predicted, then measured

predictions.md, P-1 to P-7. Every prediction held, with two corrections recorded. P-1 did not predict the devcontainer's SDK or the deadlock. P-4 named onCreate as the first refusal; the first refusal is actually the runCmd.

  • Validate on 3d0ac65: run 389 succeeded. Its log shows --remote-env CI=true on devcontainer up and on the exec, dotnet-install.sh --version 10.0.404, a locked restore of Canon.csproj, and canon check: all gates passed.
  • Locally, on 1bf7e56: the devcontainer was built from the clone's own config, and validate's runCmd ran in it with CI=true. Result: CI=[true] sdk=10.0.404, every gate ok, tests passed, issue-refs: ok. 8 commits in 84aa4665b4b0..1bf7e56af31c; Refs checked on 8, canon check: all gates passed, and the tree was clean after onCreate and after the runCmd.
  • Proof 1: the agent review job's build on a clean install-dotnet.sh --version 10.0.404, with an empty NuGet cache and CI=true. dotnet run --project src/Canon -c Release -- --help exited 0.
  • Proof 2: the devcontainer run above.
  • Proof 3: the throwaway workflow, run 44313244012, succeeded. setup-dotnet ran --version 10.0.404 with CI=true, and the build passed with 0 warnings. The workflow was removed in its own commit, 1bf7e56.
  • rollForward check (P-4): dotnet refuses with exit 155 when the installed SDK is older or newer than the pin. A drift between global.json and devcontainer.json fails validate loudly.

What was found

  • Corrected: The agent review check builds again: canon's lock file and the SDK it is built with #83 says validate passed because the devcontainer held an older SDK. That was wrong, and the statement was Claude's, from the planning conversation. The devcontainer already had 10.0.404, because it is rebuilt every run (the ghcr cache import returns 403). validate passed because CI never reached the container.
  • Found: the deadlock. Agent review builds main's canon and has no bypass, so a broken main cannot be repaired without suspending the rule.

What is left


  • python3 scripts/ci.py passes (every gate: claims, decisions, core order and pins, the What, the How examples, fixtures)
  • Every commit that changes canon carries a Basis: line (no commit touches core/; the fix commit carries one anyway)
  • Decision ids checked across every branch of both repositories
  • A session record exists under meta/sessions/, with prompt.md and bootstrap.md committed first
  • No customer-identifying material anywhere in the diff

Generated by Claude Code

Hafeok and others added 8 commits November 18, 2026 12:30
The session's charter, committed before any other change (meta/sessions/README.md).
predictions.md holds P-1, written before NU1004 is reproduced.

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
…sions index row

Deliverables 1 and 2 of #83, for Emil's ruling at G1. Session record only:
nothing under tools/canon/, .github/ or .devcontainer/ changes.

- D-1: NU1004 reproduced on SDK 10.0.404 with CI set.
- D-2: confirmed. IsAotCompatible makes the SDK add Microsoft.NET.ILLink.Tasks
  at its bundled runtime (KnownILLinkPack), 10.0.15 in 10.0.404.
- D-3: corrected. validate passes because CI is unset in the devcontainer, so
  its restore is unlocked; the devcontainer already has 10.0.404.
- D-4: not in the issue. Agent review builds main's canon and trunk has no
  bypass, so no pull request, this one included, passes it until main builds.
- Proposals A, B, C, weakest point first, each at the next .NET patch;
  D and E rejected. Q-1 to Q-3 for the ruling.

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
Option B with both corrections, 10.0.404 (Q-3); CI passed into the
devcontainer, closing a gap in Directory.Build.props' existing rule (Q-2);
Emil lifts and restores the agent review requirement himself, and a waiver
decision record is drafted for his ruling (Q-1); the throwaway workflow on
terms; deliverable 3's exit restated. Record only.

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
…to the devcontainer

Option B of #83, as Emil ruled at G1 (rulings.md):

- tools/canon/global.json: 10.0.404, rollForward "disable". setup-dotnet
  installs exactly this version for agent review, so agent-review.yml does
  not change.
- .devcontainer/devcontainer.json: the dotnet feature at 10.0.404.
- tools/canon/src/Canon/packages.lock.json: regenerated on 10.0.404.
  Three lines: Microsoft.NET.ILLink.Tasks 10.0.14 -> 10.0.15. The tests'
  lock file does not change.
- .github/workflows/validate.yml: CI passed into the container. This closes
  a gap in an existing rule (Directory.Build.props: a CI restore is locked);
  it is not a new check. Nothing changes what canon checks.
- tools/canon/DESIGN.md: the layout line and the "Built" table record it.

Refs #83
Basis: DDD-dec-101; DDD-dec-102; DDD-dec-109

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
Pre-landing proof 3 of #83, on the terms Emil set at G1: push to this
branch only, contents: read, no secrets, the agent review job's checkout,
setup-dotnet and build and nothing else. Removed in its own commit before
the head to ratify.

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
…cored

Deliverable 3's record, appended to findings.md: Emil's corrections with
attribution (validate's cause was Claude's misreading in the planning
conversation); rollForward "disable" refuses any other SDK (exit 155), so
B's weakest point is milder than stated; the three pre-landing proofs and
validate run 389 with CI passed in; successors S-1 and S-2, not acted on.

decision-draft.yaml: the waiver of the agent review requirement for one
landing, drafted for Emil's ruling on its text and not filed. Every branch
of both repositories checked: DDD-dec-123 is the next free id by files, but
#80's bootstrap names it conditionally (W-1).

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
It ran once, on 3d0ac65 (run 44313244012, success), as Emil allowed at G1.
Removed in its own commit, before the head to ratify.

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7
@Hafeok
Hafeok merged commit 1bf7e56 into main Nov 18, 2026
2 of 3 checks passed
Hafeok added a commit that referenced this pull request Nov 19, 2026
The row ended "G1 open". It now says what was ruled and landed: option B
(SDK 10.0.404 pinned with rollForward "disable", the lock file regenerated
once, CI into the devcontainer), landed in #85 at 1bf7e56; the waiver filed
as DDD-dec-123 in #86; successors S-1 and S-2 recorded. One line; #84's row
is untouched. predictions.md: P-10, written after the rebase onto dc8b737
(noted there).

Refs #83

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NbWSYRfbmM3FZTaoukTyA7

This branch was successfully deployed

1 active deployment
gates — 1bf7e56a Deployed Nov 19, 2026 by Hafeok via agent review (judge) #93
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant