docs: file the six ADR residuals no open item tracks, and correct 21 ADR records that misstate build state (BACKLOG #1497-#1502) - #1003
Merged
Conversation
…acks An ADR-to-backlog sweep over all 172 numbered ADRs cross-referenced each against docs/BACKLOG.md and the closed archive. Seven ADRs name work that is not built and that no OPEN item carries. In every case a citation search returns a hit, and the hit is a CLOSED item that shipped a different slice -- which is why none of them read as missing. #1496 ADR 0056 engine-managed VIP failover, design-only, zero references #1497 ADR 0157 increments 0, 2 and 3 unbuilt; Inc 2 mis-specified #1498 ADR 0173 revocation/OCSP on the hops #1005 did not reach #1499 ADR 0085 Direct S/MIME phases after PR1, orphaned when #157 closed #1500 ADR 0090 the web console Resend UI residual of #123 #1501 ADR 0084 message_events mitigation, gate shipped in #899 #1502 ADR 0102 option (a), deferred to an owner ruling nothing surfaces Severity is conditional throughout per CLAUDE.md section 0: zero deployments, so 0157 and 0173 describe what a first deployment would hit, not a live exposure. Numbers allocated with scripts/coord/alloc.ps1. Open-count control run per docs/LEDGER-GATE.md: items 482 to 489, open 293 to 300, closed 189 unchanged. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…radicts Same sweep as the previous commit. Twenty-one ADR index rows and Status lines assert a build state that shipped code refutes. Each correction below was verified by locating the code or the closing backlog banner, never by trusting another document. The one that matters most is 0053. Its Status still said the cp314t commitment was ratified and the spike was Phase 1 of execution. BACKLOG #90 closed DECLINED 2026-07-09 and #91 closed DECLINED 2026-07-20, so the record told a reader that free-threading is the committed scale path two months after measurement retired it. The ADR body already carried a 2026-07-06 NO-GO note; the Status line above it did not. Index rows corrected: 0003 0005 0035 0047 0053 0081 0082 0105 0119 0121 0124 0169. Status or Built lines corrected in the ADR files: 0015 0016 0022 0026 0027 0042 0053 0169. Evidence, in brief: 0015 ws_security/ws_addressing/client_cert_file knobs, transports/soap.py 0016 the TA1 classifier and blocking reply in transports/x12.py 0022 parsing/fhir/ and transports/fhir.py exist 0026 the zero-egress arm shipped in 0.2.10; live egress is still unbuilt 0027 built in 0.2.9 per BACKLOG #34 0035 untrustedWorkspaces plus the isTrusted gate in ide/ 0042 increment (a) built in 0.2.9 per #47; (b) still deferred 0047 BACKLOG #41 lists all six deliverables shipped 0081 SetMeta, applied exactly-once in transform_handoff 0082 built 2026-07-10 per #134 0105 Phase 3b built: the attachment download route and the console view 0119 ide/src/hexdump.ts 0121 ide/src/testCollections.ts 0124 the no_ack knob in config/wiring.py 0169 #1268 shipped both limbs; the creation-guard residual is now stated No superseded wording was deleted where a reader would carry it forward: 0026 and 0053 keep the sentence that was wrong and say why, per the ADRs' own append-only rule. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…r #1494 The manager session (worktree manager-424d8b) reported that a Builder is running on ADR 0056's stepdown control plane under BACKLOG #1494, titled "Build ADR 0056 engine-managed VIP failover: config seam, stepdown control plane, VIP controller". #1494 strictly supersedes #1496, which asked only whether to build ADR 0056 or retire it. That question has been answered. THE MEASUREMENT IN #1496 WAS NOT WRONG, AND THE DISTINCTION IS THE POINT. Verified after the report, not assumed from it: git show origin/main:docs/BACKLOG.md | grep -c "^## 1494\." -> 0 git show a2bfdb2:docs/BACKLOG.md | grep -c "^## 1494\." -> 0 git show origin/main:docs/BACKLOG.md | grep -c "0056" -> 0 So no filed item tracks ADR 0056 on origin/main, at this branch's base or now. #1494 is an ALLOCATED NUMBER held in an unmerged worktree, which the ledger cannot see and a sweep over docs/BACKLOG.md cannot find. The gap was real against the record; the work was real against the fleet; nothing joins those two views. That is the same shape as BACKLOG #1448. #1496 is burned, per docs/LEDGER-GATE.md: numbers are never reclaimed and holes are free. It is not re-used and not re-filed. Open-count control per docs/LEDGER-GATE.md: items 489 to 488, open 300 to 299, closed 189 unchanged -- the expected delta for withdrawing one open item that was never merged as closed. No ADR file is touched by this branch's 0056 subject: the manager session is editing docs/adr/0056-engine-managed-vip-failover.md concurrently, and this branch has never modified that file. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
wshallwshall
changed the title
github-actions
Bot
added
the
ci-red
|
CI failed while this pull request was in the merge queue, so the queue ejected it. Its own head can still be green: the queue revalidates the merge, and the path gates that skip on a pull request run there. Read the run before retrying. https://github.com/MEFORORG/MessageFoundry/actions/runs/34358161146 |
github-merge-queue
Bot
removed this pull request from the merge queue due to failed status checks
Sep 9, 2026
|
CI failed while this pull request was in the merge queue, so the queue ejected it. Its own head can still be green: the queue revalidates the merge, and the path gates that skip on a pull request run there. Read the run before retrying. https://github.com/MEFORORG/MessageFoundry/actions/runs/34359732510 |
github-merge-queue
Bot
removed this pull request from the merge queue due to failed status checks
Sep 9, 2026
|
CI failed while this pull request was in the merge queue, so the queue ejected it. Its own head can still be green: the queue revalidates the merge, and the path gates that skip on a pull request run there. Read the run before retrying. https://github.com/MEFORORG/MessageFoundry/actions/runs/34361155905 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this is
An ADR-to-backlog sweep over all 172 numbered ADRs (173 files; 0013 carries a declared companion), cross-referenced against
docs/BACKLOG.mdanddocs/archive/backlog/BACKLOG-CLOSED.md. The question asked was the owner's: is every ADR either built out, or on the backlog?Two things came back, and they are separate commits. A third commit withdraws one item after a peer session reported it was already in hand.
1. Six ADRs name unbuilt work that no OPEN item tracks (
#1497-#1502)message_eventsdeclined-handler mitigationThe shape is the same five times over: a citation search for the ADR number returns a hit, and the hit is a closed item that shipped a neighbouring slice. That is why an ADR-number grep does not find these.
Severity is conditional per CLAUDE.md section 0 — zero deployments. #1497 and #1498 describe what a first deployment would hit, not a live exposure.
#1496 was filed and then withdrawn, and the reason is worth more than the item
#1496 said ADR 0056's engine-managed VIP design was untracked. The manager session (
manager-424d8b) then reported a Builder already running on it under #1494. Verified rather than assumed:Both readings are true. No filed item tracks ADR 0056 on
origin/main, at this branch's base or now — and a Builder is nonetheless building it, against an allocated number held in an unmerged worktree. The ledger cannot see that, and a sweep overdocs/BACKLOG.mdcannot find it. Same family as BACKLOG #1448. #1496 is burned perdocs/LEDGER-GATE.md; numbers are never reclaimed and holes are free.This branch has never touched
docs/adr/0056-engine-managed-vip-failover.md, which that session is editing concurrently.2. Twenty-one ADR records state a build state the tree contradicts
Every one of these is built; the record says otherwise, so anyone planning off the index plans wrong.
The one that matters is 0053. Its Status still read "the commitment is ratified; the cp314t feasibility + scaling spike is Phase 1 of execution". BACKLOG #90 closed DECLINED 2026-07-09 and #91 closed DECLINED 2026-07-20. The record told a reader that free-threading is the committed scale path two months after measurement retired it and sharding took over. The ADR body already carried a 2026-07-06 NO-GO note; the Status line above it did not.
The rest: index rows 0003 0005 0035 0047 0081 0082 0105 0119 0121 0124 0169, and Status/Built lines in 0015 0016 0022 0026 0027 0042 0169. Each verified by locating the code or the closing backlog banner. Details and evidence are in the commit message.
A finding that turned out to be a non-finding, recorded because it looks like one
Thirteen ADR numbers have no file and no index row (0097, 0116, 0117, 0125, 0145, 0146, 0162, 0163, 0166, 0168, 0174, 0175, 0181). This is correct behaviour and nothing was changed.
alloc.ps1never reclaims a number: "holes are free, collisions are not." Seven of the thirteen hold live allocation records naming a title, branch and worktree; the other six predate the registry, and the allocator's floor is the max overorigin/main, every ref and every allocation, which now sits at 185. None can be reissued.Checks
pre-commiton all three commitsdocs/LEDGER-GATE.md)scripts/docs/backlog_status_check.py --quietscripts/docs/link_check.py docsscripts/quality/control_char_check.pyscripts/docs/backlog_citation_check.py,dangling_citation_check.pyruff,mypy,pytestscripts/docs/citation_line_check.pyexits 1 with 188 driftedfile:lineanchors. That is pre-existing (BACKLOG #1263, #1473), it is not a pre-commit gate, and none of the drifted anchors is in this diff. The new items deliberately cite symbols and paths, never line numbers.Method, and where it could be wrong
"On the backlog" was tested as: an open item cites the ADR number in its body, with items classified by
parse_itemsfromscripts/docs/backlog_status_check.pyrather than a hand-rolled scan. Two known weaknesses, one of which fired:Every "built" claim was confirmed by finding the code, not by trusting a banner.
Merge note: this branch and
manager-424d8bboth append to the tail ofdocs/BACKLOG.md, so a textual conflict at the append point is likely even though the items are disjoint.🤖 Generated with Claude Code