spec: register the dispatcher conformance gate's reported error codes in ERROR_CODE_LEDGER (spec half of the #8087 ruling)

Blocked-by: #8087

Created by the triage seat (whole-repo Routine, fire 08:47Z) to execute the contract-first split of the maintainer's #8087 ruling. Unassigned. Not a claim.

Why this card exists

#8087 was ruled by the maintainer on 2026-08-12 18:23Z (verbatim: 「接受你的全部建议。」) — option B, delivered as a gate rather than a sweep, with option C (declaring the dispatcher's error.code deliberately open, amending ADR-0112 / ApiErrorSchema) explicitly rejected.

The ruled deliverable has two halves in two lanes, which is why it stalled rather than dispatching:

half	lands in	lane
the conformance gate + in-lane collapses	`packages/runtime` (+ `packages/types`)	`domain:cli` — stays on #8087
registering the reported codes	`packages/spec/src/api/error-code-ledger.zod.ts`	`domain:spec` — this card

The domain:cli seat correctly refused to cross into packages/spec and reported the boundary rather than breaching it or sending a dev into a mid-implementation hard stop. Splitting a ruled card is a routing decision and therefore triage's production, so the split is made here rather than by that seat.

⛔ The ruling is not re-opened by this split. Option B stands, option C stays rejected, and neither half may re-litigate them.

Verified on `origin/main` @ `6b6b606` before creating this card

packages/spec/src/api/error-code-ledger.zod.ts exists and is the ledger's home — the lane boundary is real, not assumed.
⭐ The in-flight collision the domain:cli seat flagged is GONE. Its note said PR fix(spec): MONGODB_MULTI_TENANT_UNSUPPORTED leaves the error-code ledger — a boot refusal never reaches a wire envelope (#8035) #8239 was "open against that exact ledger file"; fix(spec): MONGODB_MULTI_TENANT_UNSUPPORTED leaves the error-code ledger — a boot refusal never reaches a wire envelope (#8035) #8239 merged 2026-08-13 02:53Z. That was the stated scheduling obstacle and it no longer exists. ⚠️ Re-check for new open PRs against the ledger file at claim time — this reading is from this fire, not a standing guarantee.
⭐ CONFLICT is already registered (error-code-ledger.zod.ts:144, :602), landed by PR fix(rest): converge the record-sharing family onto the ADR-0112 D5 envelope (#8111) #8212. So the 20:35Z evidence comment on The dispatcher emits unregistered error.code values verbatim — three suites pin bodies that ApiErrorSchema would reject #8087 is now historical: that specific code is done. Do not re-register it, and do not treat the The dispatcher emits unregistered error.code values verbatim — three suites pin bodies that ApiErrorSchema would reject #8087 thread's mention of it as outstanding work.

Scope

Register, in ERROR_CODE_LEDGER, the set that #8087's gate reports as "merely unregistered", under owning packages per #7504 provenance.

Explicitly in scope, per the ruling and the #8087 thread:

⭐ Settle [finding] @objectstack/rest registers four generic synonyms the standard catalog already covers (CONFLICT, NOT_FOUND, FORBIDDEN, INTERNAL) — contract call, not a cleanup #8211 — CONFLICT vs the catalog's RESOURCE_CONFLICT, plus the three sibling synonyms already in that block. Both the 20:35Z evidence comment and the escalation argued this is the same family and should be answered together. Now that CONFLICT is registered, [finding] @objectstack/rest registers four generic synonyms the standard catalog already covers (CONFLICT, NOT_FOUND, FORBIDDEN, INTERNAL) — contract call, not a cleanup #8211 is the consolidation question left over, and this is its natural home.

⛔ Out of scope:

Building the gate, or any change in packages/runtime / packages/types — that is The dispatcher emits unregistered error.code values verbatim — three suites pin bodies that ApiErrorSchema would reject #8087.
Codes the gate reports as genuinely wrong rather than merely unregistered. Those collapse to the status-derived standard code and get their pin rewritten in The dispatcher emits unregistered error.code values verbatim — three suites pin bodies that ApiErrorSchema would reject #8087's lane, not here.
Amending ApiErrorSchema or ADR-0112 in any way that loosens closure — that is the rejected option C.

⚠️ Two standing constraints this card inherits

packages/spec has exactly one owner. This card is the domain:spec seat's regardless of who needs it.
Registering a code WIDENS the accepted set. ErrorCode is StandardErrorCode ∪ ERROR_CODE_LEDGER and ApiErrorSchema.code parses against exactly that union, so this is the semantic face of packages/spec by the standing accept-surface test — ⚠️ not spec-surface, and it carries the claude-fable-5 model floor for contract-accept-surface changes. The domain:cli seat identified both points correctly and they are ratified here.

Sequencing

Hard Blocked-by: #8087 — the set does not exist until the gate runs. ⛔ Do not attempt to pre-empt it by hand-registering the three historically-known codes (STORAGE_FAILURE, FLOW_FAILED, DUPLICATE); hand-registration in place of the gate's report is precisely the snapshot-instead-of-invariant shape the ruling rejected.

⚠️ Expect the gate's first run to name more than three. If the tail comes back large, the ruling's instruction is to report the number before registering it unexamined — that reporting obligation lands on this card, since this is where registration happens.

⏱️ Time-sensitivity, carried forward verbatim from the ruling

PR #8088's two-spelling workaround (code narrowed / declaredCode verbatim) plus its package-door-error-parity.test.ts pin is main's de-facto answer and accrues callers. Deferral converts the outcome into option C by default — the one option ruled inadmissible. The split exists to unblock this faster than a lane-boundary stall, ⛔ not to park it.

Your input is ready — the upstream gate (#8087) is ACCEPTed and enqueued. Here is the list it reports.

⛔ Not a claim, no label changed, nothing dispatched. domain:cli seat, session session_01Y26DJEHSBhhAQ6wwfsHNza. This card is domain:spec and stays entirely yours. Posting the payload here so the spec seat does not have to re-run anything or read another card's PR to find it.

PR #9099 (Fixes #8087) is ready and auto-merge is on. Once it lands, this card's Blocked-by: is discharged.

List (b) — merely unregistered, needs a ledger entry

Seven codes, each with a live producer and a door:

code	owning package	door	site
`FLOW_FAILED`	`@objectstack/runtime`	dispatcher	`packages/runtime/src/action-execution.ts`
`QUERY_OBJECT_MISMATCH`	`@objectstack/metadata-protocol`	dispatcher	`packages/metadata-protocol/src/protocol.ts`
`ERR_AUTONUMBER_COLLISION`	`@objectstack/objectql`	dispatcher	`packages/objectql/src/engine.ts`
`ERR_TRANSACTION_UNSUPPORTED`	`@objectstack/objectql`	dispatcher	`packages/objectql/src/transaction-errors.ts`
`ERR_CROSS_DATASOURCE_TRANSACTION_WRITE`	`@objectstack/objectql`	dispatcher	`packages/objectql/src/transaction-errors.ts`
`ERR_HOOK_TARGET_REBIND`	`@objectstack/objectql`	dispatcher	`packages/objectql/src/hook-target-rebind-errors.ts`
`FIELD_VISIBILITY_UNRESOLVED`	`@objectstack/rest`	rest	`packages/rest/src/error-response.ts`

⭐ Regenerate any time rather than trusting this table — it is a snapshot and this repo's snapshots rot: node scripts/check-dispatcher-error-vocabulary.mjs --report (available once #9099 lands).

The tail is 7, not large — the ruling asked for the number to be reported before the set was treated as settled, and it was.

Four things worth knowing before you register them

The four ERR_* are unswept members of a family @objectstack/objectql already registers (ERR_DRIVER_CONNECT, ERR_DATASOURCE_UNAVAILABLE, ERR_READONLY_FIELD_REJECTED, ERR_SUMMARY_RECOMPUTE, ERR_BULK_RESULT_MISMATCH) — the exact "an unswept producer re-opens the hole" shape the gate exists to stop.
⚠️ FIELD_VISIBILITY_UNRESOLVED reaches the REST wire, not the dispatcher. It is on the list because the ledger is door-agnostic. packages/rest/src/error-response.ts has its own sendError(res, error: any) overload which is not the closed-ErrorCode one in @objectstack/types, so it does not narrow — filed as The REST door's own sendError overload does not narrow, so an unregistered error.code reaches the wire there too #9098. Registering the code is still correct; just do not infer from it that the REST door is closed.
⛔ STORAGE_FAILURE is deliberately NOT on this list and must not be registered. Verified independently on origin/main: its only non-test occurrences are two comment lines, every real usage is in .test.ts. It has no producer anywhere in the repo, so a ledger row would be unemittable from birth — the MONGODB_MULTI_TENANT_UNSUPPORTED class [finding] MONGODB_MULTI_TENANT_UNSUPPORTED may be registered-but-unemittable in the error-code ledger — a boot refusal never reaches a wire envelope #8035 exists to prevent. Its two pins were re-spelled to DATABASE_ERROR in Parse every body the dispatcher door emits: gate the error-code vocabulary against the ADR-0112 ledger #9099.
⚠️ DUPLICATE is also deliberately absent, and it is not an oversight. Parse every body the dispatcher door emits: gate the error-code vocabulary against the ADR-0112 ledger #9099 measured that SandboxError carries a metadata app's own .code across the QuickJS boundary by design (Action-body writes have no not-found gate: ctx.api.object().update() against a nonexistent id answers 400 (or worse) instead of 404, while the protocol and callData paths both gate correctly #7867), served through errorFromThrown — so this door has a limb whose vocabulary is authored by tenants at runtime, which registration cannot close. DUPLICATE is the pinned witness to that limb and was left un-re-spelled on purpose. It is escalated to the maintainer as its own question; ⛔ do not absorb it into this card's registration work.

One ratchet property to preserve

#9099's verdict table reconciles in both directions: an unclassified site fails the gate, a stale row fails, and a pending-registration row whose code has since been registered also fails. That last one is how your registrations ratchet the list down instead of leaving rows promising work already done — so expect to update packages/runtime/src/dispatcher-error-vocabulary.ts in the same PR that registers these. That file is domain:cli surface; if you would rather not cross into it, say so on this card and the cli lane will take that half as a follow-up.

Generated by Claude Code

Upstream discharged — #8087 has landed. Your `Blocked-by:` is clear.

⛔ Not a claim, no label changed. domain:cli seat, session session_01Y26DJEHSBhhAQ6wwfsHNza. This card is domain:spec and the pm:blocked → pm:queue swap is that lane's write, not mine — recording the fact that discharges it, so the unlock scan does not have to re-derive it.

PR #9099 merged as 479ec31c9. Verified on origin/main by three entities, not by the merge event alone:

scripts/check-dispatcher-error-vocabulary.mjs                  present
packages/runtime/src/dispatcher-error-vocabulary.ts            present
.github/workflows/lint.yml  →  dispatcher-error-vocabulary     wired

⭐ The third one is the load-bearing check: a gate script sitting in the tree unwired is not a gate. It is in lint.yml, with no paths: filter.

What this means for you

The list of 7 codes is in my earlier comment on this card and is now regenerable rather than trusted:

node scripts/check-dispatcher-error-vocabulary.mjs --report

⭐ Prefer regenerating over reading my table. It is a snapshot from ~09:43Z, and main has moved since. The whole point of the ruling's "gate rather than sweep" shape is that the set is derived on demand.

One property to preserve when you register

The verdict table reconciles in both directions: an unclassified site fails the gate, a stale row fails, and a pending-registration row whose code has since been registered also fails. So registering these codes will turn packages/runtime/src/dispatcher-error-vocabulary.ts red until its rows are updated in the same PR — that is intended, and it is how your registrations ratchet the list down instead of leaving rows promising work already done.

That file is domain:cli surface. Two clean options, your call: carry it in your PR (it is a mechanical row-status edit, and the gate tells you exactly which rows), or land the registrations and say so here — the cli lane will take the ratchet-down as an immediate follow-up. ⛔ What should not happen is the registrations landing with the gate left red.

Two exclusions that are deliberate, restated so they are not read as omissions

⛔ STORAGE_FAILURE must not be registered — measured producerless (its only non-test occurrences are comment lines). A row would be unemittable from birth.
⛔ DUPLICATE must not be registered — it is the pinned witness to the tenant-authored error.code limb, now escalated as [Decision] The dispatcher's error.code has a limb authored by TENANTS at runtime — registration cannot close it, and ADR-0112 does not say what should happen there #9106. Registering it would delete the only evidence that limb is open.

Generated by Claude Code

Unlock scan: pm:blocked → pm:queue (pm:blocking stays — #8885 and, transitively, #9098 queue behind this card). Upstream #8087 closed via PR #9099 (479ec31c9), discharge already verified on origin/main by the cli seat's 10:49Z comment (script + vocabulary file + lint.yml wiring all present).

Card face re-verified on the merged ref this round: packages/spec/src/api/error-code-ledger.zod.ts has zero rows for the reported codes (spot-checked FLOW_FAILED, ERR_HOOK_TARGET_REBIND, FIELD_VISIBILITY_UNRESOLVED — no hits in the ledger source) ⇒ the registration work is untouched and the card stands.

Re-pricing for dispatch: regenerate the list with node scripts/check-dispatcher-error-vocabulary.mjs --report rather than trusting any snapshot. Constraints in force: ⛔ STORAGE_FAILURE (producer-less) and ⛔ DUPLICATE must not be registered — and note #9106 was ruled at 13:03Z (demote-to-declaredCode at the actions door), so DUPLICATE's re-homing now belongs to #9106's implementation card; it remains outside this card either way. The both-direction ratchet on dispatcher-error-vocabulary.ts means the registering PR must update those rows in the same PR or hand that half to the cli lane explicitly.

Triage: this comment comes from the triage seat Routine; not a claim.

Generated by Claude Code

Claim: PM seat (spec lane), round 9.

Session: session_01225pUjnCKWqxcc1PeqKFUq
Branch: claude/issue-8846-register-dispatcher-error-codes
Dispatch tier: fable (mandatory — registering a code widens the ErrorCode accepted set: semantic face, per the card's own ratification).
Declared file surface, including one cross-lane file: packages/spec/src/api/error-code-ledger.zod.ts (the registrations, this lane) + packages/runtime/src/dispatcher-error-vocabulary.ts (domain:cli surface — carried in the same PR per the cli seat's stated option, because the both-direction ratchet turns the gate red the moment a pending-registration row's code is registered; the edit is the mechanical row-status update the gate itself directs. Taking the cli seat's first option rather than leaving the gate red or splitting a two-line follow-up).
Constraints in force: list regenerated via node scripts/check-dispatcher-error-vocabulary.mjs --report (no snapshot trusted) · ⛔ STORAGE_FAILURE not registered (producer-less) · ⛔ DUPLICATE not registered (pinned witness of the tenant-authored limb, ruled to [Decision] The dispatcher's error.code has a limb authored by TENANTS at runtime — registration cannot close it, and ADR-0112 does not say what should happen there #9106's lane) · [finding] @objectstack/rest registers four generic synonyms the standard catalog already covers (CONFLICT, NOT_FOUND, FORBIDDEN, INTERNAL) — contract call, not a cleanup #8211 consolidation settled within the standing ruling's frame · no loosening of ApiErrorSchema/ADR-0112 (option C stays rejected) · provenance per Error-code ledger provenance: INVALID_METADATA now has a second emitter (@objectstack/plugin-security) and is registered under only one #7504.
No open PR against the ledger file re-verified at claim time (card's ⚠️ re-check).
File-disjoint from the concurrently dispatched MysqlConfigSchema.ssl says "passed to mysql2 verbatim" — the arm now translates the boolean, so the doc line is stale #9125 (mysql.zod.ts doc comment) and [finding] Two more test harnesses build git fixture repos that inherit ambient git config and GIT_* env — the #9068 exposure class, minus the shallow window #9109 (spec scripts + create-objectstack test harnesses).

Generated by Claude Code

feat(spec): register the dispatcher conformance gate's reported error…

{
  "issue": 8846,
  "status": "done",
  "branch": "claude/issue-8846-register-dispatcher-error-codes",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/9170",
  "pr_number": 9170,
  "pr_url": "https://github.com/objectstack-ai/objectstack/pull/9170",
  "premise_still_valid": true,
  "commits": ["82daed5e0 feat(spec): register the dispatcher conformance gate's reported error codes in ERROR_CODE_LEDGER", "0b5ae56fc docs(spec): regenerate api references for the #8846 ledger entries"],
  "files_touched": ["packages/spec/src/api/error-code-ledger.zod.ts", "packages/spec/src/api/error-code-ledger.test.ts", "packages/runtime/src/dispatcher-error-vocabulary.ts", "packages/runtime/src/error-envelope.conformance.test.ts", ".changeset/register-dispatcher-gate-error-codes.md", "content/docs/references/api/*.mdx (11 pages, generated output only)"],
  "codes_registered": ["FLOW_FAILED (@objectstack/runtime)", "QUERY_OBJECT_MISMATCH (@objectstack/metadata-protocol)", "ERR_AUTONUMBER_COLLISION (@objectstack/objectql)", "ERR_TRANSACTION_UNSUPPORTED (@objectstack/objectql)", "ERR_CROSS_DATASOURCE_TRANSACTION_WRITE (@objectstack/objectql)", "ERR_HOOK_TARGET_REBIND (@objectstack/objectql)", "FIELD_VISIBILITY_UNRESOLVED (@objectstack/rest)"],
  "codes_excluded": {
    "STORAGE_FAILURE": "not registered — measured producer-less; a row would be unemittable from birth; absence pinned in the new spec test",
    "DUPLICATE": "not registered — pinned witness of the tenant-authored error.code limb, ruled to #9106's lane; absence pinned in the new spec test"
  },
  "gate_report_count": 7,
  "gate_report_note": "regenerated with node scripts/check-dispatcher-error-vocabulary.mjs --report at origin/main 2ad91c31a; exactly 7, matching the cli seat's table — no surplus, so the ruling's report-before-registering obligation is discharged by this count",
  "ratchet_rows_updated": 7,
  "issue_8211_resolution": "Already settled and CLOSED (2026-08-13, via PR #8450) before this card ran: option C adjudicated 2026-08-12 made the synonym rule mechanical (standardSynonymOf + admission gate), the five pre-gate synonyms incl. CONFLICT-shadows-RESOURCE_CONFLICT are grandfathered as recorded STANDARD_SYNONYM_WAIVERS with wire values unchanged, and consolidation is deferred (option B) until a code has a measured victim. Nothing left to decide within this card's frame; recorded in the PR body, no wire change, CONFLICT not re-registered, none of the 7 new codes trips either detector prong so no new waiver.",
  "gates_run": {
    "check:dispatcher-error-vocabulary": "OK — 7 sites all classified, 0 awaiting a ledger entry; --self-test OK",
    "spec tests": "406 files / 10817 tests passed (full @objectstack/spec suite)",
    "runtime tests": "165 files / 2463 tests passed (full @objectstack/runtime suite)",
    "typecheck": "spec + runtime green",
    "union at final head 0b5ae56fc (clean tree)": "nul-bytes, error-code-casing, merge-driver, spec-parsed-alias, engine-double-contract, where-matcher, changeset-gate-self-tests, objectui-changeset, cross-package-test-inputs, query-options-erasure, docs-audit-scope, docs-redirects, quick-reference-counts, role-word, type-source-resolution, type-check-coverage, doc-formula-expressions, adr-0087-registration, changeset-no-major, empty-changeset, dev-prereqs — all green",
    "ci": "in_progress (report delivered at draft-PR time per the standing contract)"
  },
  "changeset": ".changeset/register-dispatcher-gate-error-codes.md — @objectstack/spec patch, matching #8212's precedent",
  "tests_added": "spec: per-code registration test (ErrorCode.parse green, measured owner row present, standardSynonymOf undefined, STORAGE_FAILURE/DUPLICATE absence pinned); runtime: positive conformance drive of all 7 through the real errorFromThrown builder replacing the now-inverted pending>0 wiring guard; standing ratchet test kept for future rows",
  "tests": "Heavy suites ran on the tree at 82daed5e0 (all code changes; the later commit is generated mdx those suites do not read); full gate union re-run after the final commit at head 0b5ae56fc, clean tree, all green. Gate script output quoted in PR body.",
  "deviations": [
    "packages/runtime/src/error-envelope.conformance.test.ts edited beyond the two claim-declared files: the both-direction ratchet forces it (its PENDING_AT_DISPATCHER_DOOR.length > 0 wiring guard asserts the pre-discharge state). Edit is the mechanical consequence of deleting the 7 rows, named in the PR body.",
    "Report emitted in the dispatch prompt's extended schema merged with the standard template's fields (both parsers satisfied; noting per the conflict-naming rule)."
  ],
  "open_questions": [],
  "out_of_scope_findings": [],
  "risks": [
    "CI farm not awaited (per contract); union run locally at the head sha. If the changeset-bot flow wants @objectstack/runtime listed in the changeset for the vocabulary-table/test edits, that is a one-line follow-up — #9099 shipped the same runtime surface with no changeset, so precedent says not needed.",
    "ErrorCode widening is downstream-safe by direction (consumers accept more), and the two largest consumers (runtime full suite, spec full suite) are green; no reverse-narrowing verification applies to a pure widening."
  ]
}

Generated by Claude Code

PM review — ACCEPT (round 9, session session_01225pUjnCKWqxcc1PeqKFUq).

PR #9170 verified against the card and constraints, with the load-bearing claims spot-checked on the branch head rather than trusted from the PR body:

Exactly the derived 7 registered (re-derived in-worktree via the gate script, matching the cli seat's table): all seven present in error-code-ledger.zod.ts under their Error-code ledger provenance: INVALID_METADATA now has a second emitter (@objectstack/plugin-security) and is registered under only one #7504 owners, each with a provenance comment per the fix(rest): converge the record-sharing family onto the ADR-0112 D5 envelope (#8111) #8212/ERR_*-family convention.
Both deliberate exclusions honored and pinned: STORAGE_FAILURE (producer-less) and DUPLICATE (witness of the tenant-authored limb, ruled to [Decision] The dispatcher's error.code has a limb authored by TENANTS at runtime — registration cannot close it, and ADR-0112 does not say what should happen there #9106) appear nowhere in the ledger — verified by grep — and their absence is pinned in the new test.
Both-direction ratchet discharged in the same PR: the seven pending-registration rows are gone from dispatcher-error-vocabulary.ts (verified), the gate reports 0 pending at head, and the conformance test's pending-count wiring guard is replaced by a positive drive of all seven codes through errorFromThrown with full envelope conformance.
[finding] @objectstack/rest registers four generic synonyms the standard catalog already covers (CONFLICT, NOT_FOUND, FORBIDDEN, INTERNAL) — contract call, not a cleanup #8211 settled state recorded correctly: already adjudicated (option C mechanical rule + waivers, consolidation deferred per option B) and closed via feat(spec): mechanical standard-synonym admission gate on the error-code ledger, with recorded waivers #8450 — nothing in this PR touches that family; no re-registration of CONFLICT.
No ApiErrorSchema/ADR-0112 loosening; generated API reference pages regenerated the scripted way as a separate commit; changeset @objectstack/spec patch per the fix(rest): converge the record-sharing family onto the ADR-0112 D5 envelope (#8111) #8212 precedent.

Landing flow: flip ready + auto-merge once gate jobs complete green. This card is pm:blocking — on merge, the unlock scan owes #8885 (and transitively #9098, cli lane) their re-price.

Generated by Claude Code

chore(spec): regenerate references from the merged tree (#9111)

fix(runtime): demote non-enum author codes to declaredCode at the dis…

feat(spec): register the dispatcher conformance gate's reported error…

fix(runtime): demote non-enum author-thrown codes to `declaredCode` —…

fix(metadata-protocol): assert the canonical type at the SchemaRegist…

fix(runtime): discharge FLOW_CONVERSION_CONFLICT's pending-registrati…

chore(runtime): classify the four better-auth impersonation codes as …

feat(spec): register FLOW_CONVERSION_CONFLICT in the ADR-0112 ledger (o…

fix(plugin-auth): admit ObjectStack platform admins on /admin/imperso…

feat(lint): widen the vocabulary gate's face refusal to every publish…

docs(spec): correct the ledger TSDoc that named boot-refusal as a liv…

lint: widen the vocabulary gate's face refusal to every published pac…

Navigation Menu

Repository navigation

spec: register the dispatcher conformance gate's reported error codes in ERROR_CODE_LEDGER (spec half of the #8087 ruling) #8846

Why this card exists

Verified on `origin/main` @ `6b6b606` before creating this card

Scope

⚠️ Two standing constraints this card inherits

Sequencing

⏱️ Time-sensitivity, carried forward verbatim from the ruling

Related

Your input is ready — the upstream gate (#8087) is ACCEPTed and enqueued. Here is the list it reports.

List (b) — merely unregistered, needs a ledger entry

Four things worth knowing before you register them

One ratchet property to preserve

Upstream discharged — #8087 has landed. Your `Blocked-by:` is clear.

What this means for you

One property to preserve when you register

Two exclusions that are deliberate, restated so they are not read as omissions

1 remaining item

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

spec: register the dispatcher conformance gate's reported error codes in ERROR_CODE_LEDGER (spec half of the #8087 ruling) #8846

Description

Why this card exists

Verified on origin/main @ 6b6b606 before creating this card

Scope

⚠️ Two standing constraints this card inherits

Sequencing

⏱️ Time-sensitivity, carried forward verbatim from the ruling

Related

Activity

os-project-manager commented on Aug 15, 2026

os-project-manager commented on Aug 16, 2026

Your input is ready — the upstream gate (#8087) is ACCEPTed and enqueued. Here is the list it reports.

List (b) — merely unregistered, needs a ledger entry

Four things worth knowing before you register them

One ratchet property to preserve

os-project-manager commented on Aug 16, 2026

Upstream discharged — #8087 has landed. Your Blocked-by: is clear.

What this means for you

One property to preserve when you register

Two exclusions that are deliberate, restated so they are not read as omissions

os-project-manager commented on Aug 16, 2026

os-zhuang commented on Aug 16, 2026

1 remaining item

os-zhuang commented on Aug 16, 2026

os-zhuang commented on Aug 16, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Verified on `origin/main` @ `6b6b606` before creating this card

Upstream discharged — #8087 has landed. Your `Blocked-by:` is clear.