Fork: 0
OpenStack / openstack-caracal-dc-dc
Transfer to URL with SHA Find file
Newer
Older
openstack-caracal-dc-dc / docs / changelog-20260719-phase5-sweep.md
@JANeumatrix JANeumatrix 2 days ago 17 KB AUDIT EXIT: operator re-signed CURRENT-STATE; G3+G11 CLOSED, G9 OPEN; session close (GA-R4)
Raw Blame History

Changelog 2026-07-19 -- GA Phase-5 sweep session (one changelog for the whole

session per GA-R2 amendment D1; batches append here as they close)

Executing session for the Phase-5 repair sweep (docs/audit/phase5-sweep-plan.md, reviewed PASS + S1-S5). Every batch is operator-gated at its stated gate point; every commit names its GA-F/GA-R trace and carries its own revert note. Status lives in docs/CURRENT-STATE.md (GA-R1) -- this file records what CHANGED.

Batch 0 -- tooling first (OPENED by operator 2026-07-19; repo-only, no cloud contact)

  1. 0.1 repo-lint L10 status-coupling check (GA-R1 rule 8 / C1) -- commit 0b991a5.

    • WHAT: repo_lint.py gains L10: a change set modifying a state-bearing surface (anything under docs/audit/; a **Status: line in design-decisions.md; a gate row | G<n> | in session-ledger.md) without touching CURRENT-STATE.md is flagged. Pending working-tree set when dirty, HEAD commit when clean. WARN this batch (L10_FAIL = False); the flip to FAIL is a one-constant change due at Batch 1 open (sweep plan: "WARN for one batch").
    • WHY: amendment C1's Phase-5 implementation clause, verbatim. DIVERGENCE NOTED: the sweep plan's item 0.1 paraphrase said "CURRENT-STATE gate rows", which is vacuous (editing a CURRENT-STATE row touches CURRENT-STATE by construction); the ratified utterance says "ledger gate rows" and the ruling wins. The ledger carries no gate rows today; the pattern guards their reappearance.
    • Harness: tests/repo-lint T34-T41 (git fixtures), 42/42 PASS.
    • REVERT: git revert 0b991a5.

  2. 0.2 ledger-scan SEC reporting repointed at security-ledger.md (GA-R4/F3) -- commit b504cb6.

    • WHAT: scan emits open COUNT + ID list + register-of-record pointer; row-level disposition text dropped from scan output. Session-ledger machine-derived SEC bullet re-seeded to pointer + count form. G14 evidence cell repointed at docs/security-ledger.md same commit.
    • WHY: F3 verbatim ("pointer and a scan-derived count, never row-level dispositions"). MEASUREMENT CORRECTION (GA-R1/C2): G14's count said 9; its own enumeration and the scan both measure 8 OPEN rows -- corrected to 8, divergence noted in the cell.
    • Harness: tests/ledger-scan +4 checks, 41/41 PASS.
    • REVERT: git revert b504cb6 (restores the 2026-07-18 cell incl. its count error).

  3. 0.3 ledger-scan wrap-aware next-free exclusion (GA-F15) -- commit 527f1fd.

    • WHAT: the "next-free"/"reproduc*" exclusion now suppresses the matching line AND the immediately following line (window exactly one line, reset per file), so a quoted next-free value whose hyphenated token wraps no longer counts as an assignment. Chose this over assignment-site counting: DOCFIX/BUNDLEFIX have no single register file to key assignment sites on (unlike D headers).
    • WHY: GA-F15 -- the wrap class fired twice live inside the audit itself (CURRENT-STATE section 9, now marked FIXED). Live counters re-verified UNCHANGED post-fix: D=130, DOCFIX=197, BUNDLEFIX=052.
    • Harness: tests/ledger-scan +3 checks, 44/44 PASS.
    • REVERT: git revert 527f1fd.

  4. 0.4 record-audit status scan keyed to the legal token set (GA-F10 / vocab H1) -- commit f7580fd.

    • WHAT: extractor knows Option A: legal OPEN/BLOCKED/HELD/CLOSED (OPEN was previously UNSCANNED -- post-migration OPEN-vs-CLOSED contradictions would have been invisible); DONE/TO RUN/RUNNING added; new report section 6 classifies every distinct token and flags illegal tokens + event words (VOID/REOPENED/ SUPERSEDED) minted in status position. Prose event words stay legitimate row-history vocabulary. Report-only: 0/1/2 exit contract unchanged.
    • WHY: H1 requires the scan before any token migration (Batch 2.5 dependency). Live sanity run (scratch --out; the committed Phase-1 capture untouched): 12 distinct tokens, 163 flagged counted occurrences = Batch 2.5's work-list. Contradiction groups 12 -> 13 on the scratch run: the OPEN-token detection gain, not drift.
    • Harness: tests/record-audit +9 checks, 33/33 PASS.
    • REVERT: git revert f7580fd.

Batch 0 verification: all three touched harnesses green (42/42, 44/44, 33/33), full gauntlet + repo-lint results recorded in the batch completion report and the ledger entry for this session.

Batch 1 -- D-130 verification pair, decision, and the plan gate rebuild

(OPENED by operator 2026-07-19)

  1. L10 flip WARN -> FAIL at batch open -- commit 9683a18. One constant (L10_FAIL = True) + harness rc updates, 42/42. Per the Batch 0 revert note. REVERT: git revert 9683a18.

  2. 1.1 v8 plan-only verification (GA-F01, G4) -- captures docs/audit/outer-plan-20260719-baseline.txt (7/2/7, zero drift from 2026-07-18) and outer-plan-20260719-v8-ignorechanges.txt (6/2/6, ZERO forces-replacement). Suppression CONFIRMED; the operator's contingent lean survived its verification.

  3. 1.2 v7 throwaway exercise (GA-F01, G4) -- capture docs/audit/throwaway-v7-20260719.txt. Isolated throwaway-v7-* root (own scratch pool under the AppArmor-covered parent, no networks, no OS): defect reproduced (seed volume "must be replaced" after staging-ISO loss); replacement applied UNDER the running domain, rc=0, NO bounce (domid 3->3); teardown verified zero residue (0 domains, 0 pools, dir gone). New fact recorded: the bounce path is the autostart/domain Update, not the volume replace.

  4. 1.3+1.4 D-130 ADOPTED (a) + implemented -- commit dac10e0. GA-R5 ruling (question + utterance in the D-130 register entry); lifecycle { ignore_changes = [create] } on libvirt_volume.seed in modules/cloudinit-vm; NEW harness tests/cloudinit-vm/run-tests.sh (6/6: guard placement, exact attribute, no leak onto the data disk, tofu validate). G4 + G5 CLOSED. REVERT: git revert dac10e0.

  5. 1.5-1.7 G6 reconcile + triple + G7 re-plan -- commit 5a4f5eb. Operator ruled (ii) state surgery (S3/GA-R5): autostart:true injected into both domain resources (were None), serial 22->23, backup opentofu/terraform.tfstate.pre-G6-20260719; guests untouched (ids 1/2, running). Re-plan capture docs/audit/outer-plan-20260719-postG6.txt = 6/0/6, equals the section-5 recorded triple exactly -- G6 + G7 CLOSED. REVERT: docs via git revert 5a4f5eb; state via push of the pre-G6 backup.

Out-of-batch: permission posture extension (operator-directed 2026-07-19)

  • WHAT: .claude/settings.json ask list extended -- Bash(git merge*), Bash(git revert*), Bash(git reset*), Bash(git rebase*), Bash(git branch -D *), Bash(bash scripts/* apply*), Bash(bash scripts/* install*), Bash(ssh * sudo *), Skill(*). Nothing added to allow; deny + the PreToolUse guard hook unchanged.
  • WHY: operator drives sessions via Remote Control without a second terminal; remaining auto-mode classifier HARD-DENIALS (walls) on mutation surfaces matching no ask rule blocked the approve flow entirely (hit live: a Skill invocation hard-denied 2026-07-19). Extends the 2026-07-18 "ask preempts classifier" posture; targeted ask rules chosen over a catch-all Bash(*) per doc-cited guidance (ask appears to preempt allow -- a catch-all would destroy frictionless reads). Skill(*) syntax is undocumented; verified EMPIRICALLY (post-change Skill invocation prompted + ran; pre-change identical call hard-denied). Remaining walls get their own ask rule when hit -- the settings edit itself now prompts instead of denying.
  • REVERT: git revert this commit, or cp .claude/settings.json.bak per the operator backup if the bak is refreshed first.

Batch 2 -- record fixes (OPENED by operator 2026-07-19; commits per-GA-F)

  1. 2.4 GA-F05 hostname (02af860): vcloud on CLAUDE.md + SKILL.md + a discovered fourth surface (environment.md:43). Memory copy = Batch 3.
  2. 2.2 GA-F03 tofu pins (b5981a9): four sites demoted to the CURRENT-STATE section-7 pointer; the one LIVE false claim (workflow :360) corrected. A fifth site (platform-traps:19) fixed at batch close.
  3. 2.3 GA-F04 OPNsense rows (6232dd3): skill pin row -> pointer; trap sections requalified as entered-at-26.1-persisting; env.md's stale tofu pin (GA-F03 class) caught on the same surface.
  4. 2.1 GA-F02 readiness doc (360e8d6): three contradictory plan-count forms -> section-5 pointers; falsified deploy-ready blockquote reframed as history; fresh-session banner -> one-line pointer at the G9 canonical entry doc (operator scope addition, probe hole H1).
  5. 2.5+2.6 H1 migration + GA-R1 demotions (47423c6): mapping record docs/audit/h1-token-map-20260719.md (no token unmapped); workflow doc 26 flags -> 0 (8 State lines demoted, tracker model dropped from the header); runbooks 18 -> 0 with close-out instructions RETARGETED at CURRENT-STATE; CURRENT-STATE 6 -> 0 (G9 STOPPED -> BLOCKED w/ named blocker; stale "GA-R1 NOT yet ratified" header fixed per C2); gate rows gained GA-R6/E1 [V]/[R] type tags; CLAUDE.md expectation line demoted (S5b). History classes deliberately untouched (Batch 4 / GA-R3).
  6. 2.8 GA-F14 banner on the D-127 changelog (content preserved).
  7. 2.9 GA-F11/F12 adjudication: one real live version row fixed (env.md:29 -> pin authorities); 192 non-history broken-ref rows adjudicated ALL false-positive (provider-repo doc paths, glob/range forms, slash-phrases, wraps) or VR0-era history docs routed to Batch 4. Spot-checks: as-built:59 "netbox/tailscale" = slash-phrase; workflow:473 retired-script mention = history; env.md:86 "phase-00..08" = range glob.
  8. 2.10 S1 R3-F disposition: all 13 items accounted -- F01-F07/F09/F10 closed by the 2026-07-16 Phase A/B + rulings + SEC-010 artifact; F08 CLOSED-SUPERSEDED (R-3's 4th storage node restored rebuild headroom); F11 FIXED (supernet-unassigned clause dropped; vr1-dc1 stays HELD on transit addressing, G12); F12 CLOSED (committed budget model, 186 GiB headroom covers transient N+1 amphora; Stage-6 drill exercises the real path); F13 CLOSED as recorded observation. Stale readiness C5 queue corrected (it queued amendments that landed 2026-07-16).
  9. 2.7 GA-R3 D-tags (operator-ruled): D-121..125 [ARCH], D-126..129 [OPS], D-130 [ARCH]; register template gains the class-tag convention; D-129's ARCH fragment (Roosevelt metal-edge profile) split to a dated D-122 AMENDMENT per A1 (utterance: "Tags as proposed + D-122 amendment").

Batch-2 verification + the criterion-mechanics finding. Extractor re-run: live-surface vocab flags = 0; the remaining flagged occurrences (110) and contradiction groups (12) sit ENTIRELY on history-class surfaces (session ledger, changelogs, the append-only register, VR0-era docs -- Batch 4's rotation/consolidation) or are extractor MECHANICS: the global plan key conflates different roots' plans (Stage-1's 13-object plan vs the DC0 triple), charm channels/step numbers read as juju versions, the provider (2.7.2) vs MAAS-server (3.7.2) attribution, compatibility thresholds (LXD >= 6.7), and prose like "Stage-4 concerns" keying a stage status. FINDING (logged, not silently absorbed): the sweep plan's Batch-2 exit criterion "contradiction groups on counted non-changelog surfaces reach ZERO" is not literally computable by the extractor's changelog-only exclusion -- ledger/register/finding classes are counted non-changelog but their ruled disposition is Batch 4/GA-R3 retention, and the status authority itself must narrate history (7/2/7 -> 6/2/6 -> 6/0/6) to comply with GA-R1 rule 2. The achieved, verifiable state: zero contradiction among LIVE-surface status claims (mechanical classification script in the session record); final ZERO-on-counted comes with Batch 4's rotation + Batch 6.1 as the plan already sequences.

Batch 3 -- memory hygiene (OPENED by operator 2026-07-19; edits per-file gated)

Pre-edit evidence: full memory dir copied to docs/archive/memory-20260718/.

  1. GA-F06: review-after-each-step.md REMOVED. Repo copy of the provenance ruling: grounding-audit-20260718.md GA-F06 ("It was not a standing instruction. It was only supposed to be for a one time / very short sequence of steps."). No durable observation survives it -- the drift-catch anecdote is repo-recorded in the review changelogs.
  2. GA-F07: project-posture-closed-test.md REMOVED. Repo copy of the provenance ruling: GA-F07 (stance never ruled; posture returns NEUTRAL; the closed-test FACT is repo-carried; SEC gates were never deferred).
  3. GA-F05: hostname -> vcloud in multi-workstation-remote-control.md (the last of the four surfaces).
  4. GA-F08: MEMORY.md index rebuilt -- two defective lines gone, the two surviving entries re-described accurately.
  5. GA-R7 rule 6 skill line: "Claude Code sessions launch from the repo root, always" added to SKILL.md's session-bootstrap loop with the GA-F06-origin rationale (subdirectory launch = divergent auto-memory).
  6. G1 full review (rules 1-4): creds-folder-convention.md found duplicating a repo-carried convention + linking both removed entries -- registered as GA-F16 and reduced to a pointer entry (one possibly-unduplicated fact retained with a verify-then-graduate note). multi-workstation-remote-control.md reviewed: workflow facts operator- confirmed (the 2026-07-19 walls-vs-gates entry quotes a same-session directive; policy copy lives in the repo changelog it points at). Post-review state: 3 files, zero rule-1..4 violations, zero stale refs (grep-verified). REVERT: restore any file from docs/archive/memory-20260718/ (pre-edit copies); git revert the batch commit for the repo surfaces. Note on the archive copies: memory-20260718/MEMORY.md's em-dashes were ASCII-normalized (--) to satisfy the repo L1 rule; the transform is whitespace/punctuation-only and content-faithful (this note is the record of it). No gauntlet run for this batch: zero script/tooling changes (memory + docs + one skill prose line); repo-lint 0-fail + fence check OK.

Batch 4 -- consolidation + rotation (OPENED by operator 2026-07-19)

  • C1 ledger rotation (GA-R4 rule 6, GA-F13): 1187 -> 131 lines; bodies verbatim to docs/archive/session-ledger-rotated-20260719.md; lessons/facts ROUTED first (format=line trap -> appendix-A; guard discipline -> operating-discipline; VR0 facts + admin/controller -> maas-as-built; lessons 1-6 verified already covered). F1 cap now enforceable.
  • C2-C5 changelog consolidation (GA-R2, GA-F09): 96 files (95 + the v1-redeploy running record) -> docs/archive/changelogs/, four stage records under docs/archive/stage-records/ (v1-era; stage0-1 = 45; stage2 = 38; stage3-partial = 11), each commit message MANIFESTS its sources (D2). The live session changelog stays top-level until session close. NOTE: v1-redeploy-changelog's own "completion consolidation" TODO is NOT discharged by the move -- it remains in the ledger's project-completion block.
  • C6 docs disposition (D2 scope): 24 non-changelog history docs -> docs/archive/ (v1-era packs/handoffs/findings/drafts; executed plans; the R3-F register; Model-A fallback; DR seed; ipv6/netem research; D-068-openbao DRAFT). RETAINED top-level, 16 files each justified: status authority, registers (D/SEC), ledger, workflow (identity), design doc, readiness checklist, two as-builts, tenant contract (lint L7 auto-covered), log convention, live finding doc (CURRENT-STATE-cited), D-129 review (open sub-decisions), NetBox scope (live apex reference), D-068 analysis (open-decision evidence), live session changelog. Live-surface references to every moved file rewritten to archive paths (18 files touched); history-internal refs left verbatim.

Batch 5 -- skill sweep (OPENED by operator 2026-07-19; single commit)

SKILL.md reconciled against the RATIFIED GA-R texts (not the draft): new "Record authority and session discipline" section (~35 lines per the draft's budget); header phase claim + the false "opentofu UNVALIDATED / no tofu binary" claim DEMOTED to CURRENT-STATE delegation; session bookends rewritten (OPEN = CURRENT-STATE first; CLOSE = 15-line summary

  • archived body + push); stage-close loop gains the GA-R6 close-out set (named-check closes, no conditional close, consolidation commit, memory review); change-delivery loop gains D1 (one session changelog, never citable); routing rows re-pointed (VR1 status, versions, resume). The draft's reconciliation circuit-breaker KEPT as discipline (no ruling quote -- marked as such). Line-by-line checklist: docs/audit/skill-sweep-checklist-20260719.md. 176 -> 224 lines.

Batch 6 -- Phase-6 exit runs (OPENED by operator 2026-07-19)

  • 6.1 ruling + run: operator ruled the docs/archive/ counting exclusion IN ("Yes, extend exclusion"); record-audit.py + harness updated (35/35). Post-run: 9 groups, per-group adjudication = ZERO contradictory pairs (full verdict table: docs/audit/phase6-exit-runs-20260719.md; group-count floor 9 recorded as the re-run baseline). Missed-disposition fix: stage3-review-base.patch archived; CURRENT-STATE known-stale note made token-free.
  • 6.2: docs/*.md = 16 (< 25), history preserved via git mv + manifests.
  • 6.4: fresh capture outer-plan-20260719-phase6-exit.txt = 6/0/6 == section-5 triple EXACTLY.
  • 6.5 (S2): G9 row now names SEC-010's step-B apply+verify dependency (the only apply-gated SEC row).
  • 6.3 + 6.6 PENDING operator execution: probe trio (3 clean sessions, auto-memory OFF, 7-question set, 7/7 on 2-of-3) + the top-to-bottom re-read and re-signature.

SESSION CLOSE (2026-07-19)

Batch 6 completed: probe trio 21/21 (7/7 all three); operator re-signed CURRENT-STATE ("Reviewed, approved, continue."). ALL exit runs PASS; the grounding audit is EXITED, the freeze lifted, G9 OPEN. This changelog is the session's complete WHAT-CHANGED record; it consolidates into the Stage-3 record at stage close per GA-R2.