Repository navigation
metadata-protocol: a pair row declared for package X is served at an addition package Z's exact address as Z's item, where the registry answers Z's own (ADR-0150 D4, red line 2) #32606
Description
epic:#31527#1 · os-bill · session_01Utp6CYGYEUhpi29FpjLmxT
Parent: #31527 (ADR-0150, epic red line 2: ambiguity reported, never guessed; another package's customization is never served at this address). Found by server seat 5 while working #32561, measured at PR #32557's head 075fb3188 and again after #32561's local fix, which does not change it. Detail: /mnt/project-files/server-seatC/card-32561/finding-declared-row-at-addition.md.
Premise
Setup:
- App A (
com.acme.suite) has module packages X and Z. Z depends on A and on X. - X ships
welcomeinen-US. Z ships onlyfr-FR, a bundle addition to X's template. A.customstores afr-FRrow declaringoverrides: { package: X }, both stored and hydrated.
What happens:
?package=X&member=fr-FRserves the row. This is correct: the row is declared for X.?package=Z&member=fr-FRserves the same row, restamped_packageId: Zwithprovenance: packageand_diagnostics.valid: false.- The registry answers Z's own item there:
getPackageScopedItem(Z, 'welcome', 'fr-FR')is Z'sfr-FR. - CONTROL: a full holder Y (which ships
en-USandfr-FR) serves its own item at?package=Y&member=fr-FR.
So a row declared for X is served at an addition's address as that addition's item. The exact qualifier's pair arm never sees it: the item arrives already stamped Z, so exactQualifierItem returns it.
Likely surface (estimated by the finder, not measured): findServedOverlayRow and servedOverlayRowCandidates in packages/metadata-protocol/src/protocol-environment-rows.ts. These produce the pair's served-override candidate for a package that depends on the declared base.
Re-check: the harness above in packages/objectql/src/package-scoped-exact-qualifier.test.ts, reading ?package=Z&member=fr-FR.
Fix shape
- At Z's exact address, a row whose
overrides.packagenames X is not Z's layer. Serve what the registry answers at Z, which is Z's own item. - Sweep every served-override candidate reader for the same "depends on the declared base" shortcut.
- Pin both registration orders, with CONTROLs for X's own address and for a full holder.
- Declare the move FROM → TO with its arm,
narrowing. The first-party count is expected to be 0.
Order
Same lane as #32561 (server seat 5). Lands after #32561, because both touch the exact read's pair rows.
Blocked-by: #32561
Activity
objectstack-fleet commentedon Nov 12, 2026
triage#1 · hotlong · session_01Q4EjjMnyLqJdzuktEmRPXH
Path: none
Triage:
findingpromoted →domain:server· p2 ·pm:blocked· 2026-11-12T07:47Z⛔ Not a claim.
domain:server(packages/metadata-protocolexact read at an addition package's address). Linked under design(metadata): address behavioural metadata by package — key flows, actions, jobs and other non-object types by (package, name); re-open ADR-0028 (replaces #30806 phase 3 renames) #31527.Blocked-by: #32561is indexed; metadata-protocol: the exact read binds an undeclared App-pair row to X even when an addition package ships the same bundle member, where the registry binds it to no package (ADR-0150 D4, red line 2) #32561 is open. Grade: p2, as metadata-protocol: the exact read binds an undeclared App-pair row to X even when an addition package ships the same bundle member, where the registry binds it to no package (ADR-0150 D4, red line 2) #32561 (the same red line 2 family).Generated by Claude Code
objectstack-fleet commentedon Nov 12, 2026
domain:server#1 · os-sales · session_01L6bhgVaKJ9TPfXPRhdqx3e
Parked by the epic #31527 release-line ruling 6824949405 (the maintainer's decision card on the epic seat's thread, project "Objectstack security ui"): this is a p2 card of the epic with no open PR, so it is not on the release line. The maintainer's follow-up (13:30Z) keeps cards that are already built on a pushed branch going to merge; this card has no branch, so it parks.
Restart-when: the release that carries the epic #31527 release line is published (ruling on #31527, comment 6824949405)
State: Unclaimed and unstarted; its
Blocked-by:line stays for the restart.pm:blocked→pm:on-hold; no assignee. Nothing is deleted.objectstack-fleet commentedon Nov 12, 2026
epic:#31527#1 · os-sales · session_01L6bhgXQnNkXTT1Wu7i7Y3i
Un-parked · #32606 · 2026-11-12T21:35Z. The maintainer (Sales) lifted the park of unstarted p2 epic cards (ruling record 6836538545 on #31527, amending 6824949405).
pm:on-hold→pm:queue; the epic seat dispatches it (see that record for the seat).Generated by Claude Code
objectstack-fleet commentedon Nov 13, 2026
triage#1 · hotlong · session_01Q4EjjMnyLqJdzuktEmRPXH
Path: none
Triage: H14 ·
pm:blockingderived · 2026-11-13T12:18Z⛔ Not a claim; grade and state unchanged. Open card(s) #32849 name this card on a line-leading
Blocked-by:, so the derived cache is added. It comes off when they close or drop the line.objectstack-fleet commentedon Nov 14, 2026
domain:server#5 · os-elon-musk · session_016WV4cy1y7UNVbT2QAdmgdn
Claim: area-lease pilot #33059, area E1 (seat domain:server#5, rules /mnt/project-files/server-area-pilot/pilot-rules.md) · 2026-11-14T14:53Z
Session:
session_016WV4cy1y7UNVbT2QAdmgdn(area E1 thread)Account:
os-elon-muskBranch:
claude/project-thread-725beaWorktree:
objectstack-issue-32606Domain:
domain:serverSeat:
domain:server#5(lease E1)File surface:
packages/metadata-protocol/src/**(the moved-edit predicates and their three callers),packages/objectql/src/package-scoped-exact-qualifier.test.ts, the spec semantic entry, changesetsClause-②: no
Metadata-protocol: semantic
Responsibility: an App pair row declared for X is no longer served at a bundle addition Z's address, list, execution list or customizations read while X still ships the name (PR #33251)
Premise: holds — measured on
maine540baf in both registration orders before any edit (Z's address, list and execution list served the row stamped Z)