[frahlg]

Problem and result

Moves optimizer/native/bundle to Energyplan 0.7.0 (srcfl/energyplan 847936a, tag v0.7.0-beta.1).

• What changes for households.

  • With a peak-power tariff, the worker now takes the demand hour in progress and the day's completed peak in that day's group. fix(mpc): price the clock hour in progress for planners that announce demand_elapsed_hours #1966 already sends that shape to a worker that lists demand_elapsed_hours, so each local day counts as one sample.
  • A new plan is never worse for the car goals or the bill than the plan the worker already holds.
  • Limits that cannot all be met get a clear infeasible reply.
  • The worker needs less memory.

• Bundle.

  • It now carries linux-arm64 and linux-amd64 only. fix(native): verify a bundle that has no worker for this host #1977 lets verify.py check such a bundle on a Mac and skip execution there. A Mac keeps Core's own planner.
  • The manifest, the two version pins in the tests and the README version change.
  • The README's budget sentence now points to the contract.
  • Licenses and forecast schemas are unchanged. The notices list two more vendored microlp modifications.

• Contract. docs/energyplan-contract.md gets the text that becomes true with this worker:

  • the demand_elapsed_hours feature and its negotiation;
  • the demand-charge fields and bill;
  • each call's time limit since fix(mpc): fit each worker call's time limit to its own deadline #1968;
  • the recovery rule for the batteries' total;
  • optional reply diagnostics as opaque numeric codes, which Core never uses to judge a plan.

  Energyplan's copy is byte-identical (sha256 d2b24ad9…).

• Tests. TestNativeEnergyplanTakesElapsedDemandHours now requires the feature from the bundled worker instead of skipping.

Scope and safety

• Protocol v1 is unchanged.
• Without the feature, Core keeps sending the 0.6.0 demand shape, as fix(mpc): price the clock hour in progress for planners that announce demand_elapsed_hours #1966 does.
• Rollback. A revert of this PR restores the 0.6.0 bundle with all three workers. That revert was prepared and checked: verify.py, the native suite with -count=5 and make verify pass on it.

Verification

• Bundle files are byte-identical to the bytes Energyplan's release gate ran:
  • ftw-solver-linux-arm64 sha256 a6f1609d…;
  • ftw-solver-linux-amd64 sha256 67ff308a….
• python3 optimizer/native/verify.py: Verified Energyplan 0.7.0, 2 platforms, linux-arm64 handshake passed.
• Native suite with the bundled worker, -count=5 (go test ./internal/mpc ./internal/energyforecast ./cmd/ftw -run 'Native|RustForecastHost'): 360 passed, 0 failed. The 5 skips are TestNativeCapturedSiteReplay, which needs a captured request. The elapsed-demand test runs and passes.
• make native-solver-test and make verify: pass.

Checklist

• The change follows VISION.md and one selected scope.
• I checked overlapping PRs and coordinated shared files/contracts.
• Relevant checks cover the changed behaviour and failure paths.
• A human reviewed changed UI in a browser, or no UI changed.
• A Changeset is included, or the change is exempt.
• Every commit has a DCO sign-off.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-06T20:26:31.088180Z	079d691	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.