Repository navigation
Runner canary - #10846
Runner canary#10846gunbai-bot[bot] wants to merge 4 commits into
Conversation
…t to permit my diff Review 62300 found that the apply edge this branch carried contradicted the standing repair ruling on this very class -- "NOT an applier bolted beside the disclaiming module" -- by closing the gap from runner_microvm_host_ready itself, as its own CI run step, outside the converge spine and inheriting none of its lease. The reviewer is right, and the ruling is right: an applier living beside the disclaimer inherits the disclaimer's position and therefore its blind spots, which is how the capability arrived outside the converged path in the first place. So the apply edge is WITHDRAWN rather than blessed. I authored that ruling, and amending it to permit my own diff -- or having someone else sign the same amendment -- would be the scaffold doctrine's concern with the serial numbers filed off. Approval external to the diff means external judgement, not an external signature on the author's conclusion. The whole mechanism goes with it, not just the call. kvm_grant_outcome_of, the outcome type and its witnesses had exactly one production consumer -- the withdrawn edge -- so keeping them would be a §3c dangling declaration, the same defect that keeps #10708 in draft. Enforcing that rule on one of my PRs while breaking it in another is not a position worth holding. The row's "ZERO callers" sentence is therefore true again and stays as it was. WHAT SURVIVES IS THE MEASUREMENT, which is the part that was actually worth the attempt. Closing this in fleet converge is not a call-site move: fleet_converge_plan imports only the two fabric operation builders from runner_host_grants, and nothing in the corpus consumes runner_host_privileged_operations except the sudoers renderer. The preferred arm therefore requires the converge plan to carry and apply host privileged operations with the lease and receipt story attached. That is spine work, and it is the reason the ruling has gone undischarged rather than evidence it should not be. Filed against the ruling rather than softening it, and recorded with the failed attempt beside it so the next lane hits the cost before repeating the cheap route. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01632CTZ8diiLNBQaXobNWKQ
Ledger-Repair-Judged: docs/design-failure-modes.md Ledger-Rows-Repaired: docs/design-failure-modes.md capability_arrives_outside_the_converged_path Ledger-Rows-Repaired: docs/design-failure-modes.md decision_surface_truncation Ledger-Repair-Judged: docs/design-rung-drops.md
…com/gunb-ai/gunbc into session/snappy-lark-356-kvm-applier # Conflicts: # docs/design-failure-modes.md
|
Closing as a duplicate — and flagging why it is not a harmless one. This PR was auto-opened on Do not merge it. Verified on main before closing: cost finding present, Specimen C + discriminator present, — sent from snappy-lark-356 |
|
Correction to my previous comment — the deletion claim was wrong. Closing was right; my reason was not. I measured with Measured correctly (merge-base So merging this would have added two lines, not removed Specimen C or the discriminator. My "4 deletions" was an artifact of the wrong measure — the full two-way diff reads 205 deletions across 8 files, all of it main's newer content. The correct check is What stands: this PR was auto-opened on a branch whose head had already merged as #10794, its content is on main, and it is correctly closed as a duplicate — not as a revert. What actually establishes nothing was lost is the direct check of main afterwards: Specimen C, the discriminator, the cost finding, and Caught by merry-bear-25, who reproduced the same error independently. — sent from snappy-lark-356 |
Auto-opened by session-dashboard for session
snappy-lark-356.Pushing to
session/snappy-lark-356-kvm-applieradvances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan