feat(spec): retire the inert additionalTypes key from MetadataPluginConfig (#8586, ADR-0049) - #8702
Merged
Merged
Conversation
… ADR-0049) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01E5tUwGM3LQoqErTfkvRW7W
…ire-additionaltypes
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
Contributor
📓 Docs Drift CheckThis PR changes 3 package(s): 107 hand-written doc(s) reference the affected code and may need an implementation-accuracy re-verification:
⛔ 7 release-owned page(s) also reference the affected code. These are read-only:
|
github-actions
Bot
added
documentation
This was referenced Aug 14, 2026
…ire-additionaltypes
…g main (Batch D analytics entry absorbed; 86 semantic)
os-zhuang
marked this pull request as ready for review
August 14, 2026 15:21
os-zhuang
enabled auto-merge
August 14, 2026 15:22
This was referenced Aug 14, 2026
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #8586
Implements the maintainer ruling of 2026-08-14 (verbatim 「同意」, joint with #8421): ADR-0049 disposition remove.
MetadataPluginConfig.additionalTypeswas declared, authorable, documented on four docs pages as THE way a plugin registers a custom metadata type — and read by nothing. Measured basis re-verified on this tree before implementing: every occurrence is a declaration or a mention, the only production writer of the manager's type registry issetTypeRegistry(DEFAULT_METADATA_TYPE_REGISTRY), declared count == live count (27 == 27). Premise still valid.Note: #8421 is not addressed here — this card lands first; the refuse-by-static-registry work on #8421 remains open and follows it.
Route:
retiredKey()tombstone, not strict deletionThe dispatch prompt presumed a strict object ("authoring it becomes a loud unknown-key refusal"). The tree disagrees:
MetadataPluginConfigSchemais a plainz.object, not.strict(), so a plain deletion would be a silent strip — the exact trap the playbook forbids (#3726/#3733, ADR-0104). Per the playbook's route table the removal is aretiredKey()tombstone (thekernel/Manifest:loadingprecedent, #4914): authoring the key is atscerror (typednever) and a parse refusal —invalid_typeat pathadditionalTypes, message carrying the full prescription (notunrecognized_keys; tombstones raiseinvalid_typefromz.never(), asalias-integrity.test.tsrecords). The pin tests assert the strongest envelope this surface has: refusal + issuecode+path+ prescription text. Astatusfield does not exist on a ZodError issue, so it is deliberately not asserted.What this PR comprises
packages/spec/src/kernel/metadata-plugin.zod.ts, guidance following the five house conventions (noos migrate metasentence — no conversion covers this surface, see below).${…}placeholders — the #8336 shape one surface over, deliberately outside its connection-material class #8495 / PR feat(spec): refuse${…}placeholder syntax in memory persistence.path / persistence.key at publish (#8495) #8666 precedent):18.kernel__MetadataPluginConfig__additionalTypes.ts(kernel/MetadataPluginConfig:additionalTypes); a newretired-key:18region was added toRETIRED_KEYS_BY_MAJOR(hand-written markers, generator-filled content);18.metadata-plugin-additional-types-retired.ts;applyConversionsToStoredItemmaps types onto stack collections; a metadata-plugin config is neither (PLURAL_TO_SINGULARhas nopluginsentry) — a conversion would be a transform with no seam that ever runs. Same rationale askernel/Manifest:loading.registry.tsregenerated viagen:migration-registry(85 semantic, 30 retired-key, 53 retired-def); step18 rationale extended (hand-written half).spec-changes.json/docs/protocol-upgrade-guide.mdare unchanged, correctly: the projections exclude the uncut major 18 (the on-main [finding][spec] memory-driver persistence.path / persistence.key still accept unresolved${…}placeholders — the #8336 shape one surface over, deliberately outside its connection-material class #8495 semantic entry is absent from them too);check:spec-changes/check:upgrade-guidegreen.authorable-surface/kernel.jsonrow becomeskernel/MetadataPluginConfig:additionalTypes [RETIRED]viagen:schema(tombstone route keeps the row — the dispatch card said the baselines "lose the row", which is the strict-removal shape;[RETIRED]is this route's correct reading).authorable-surface.base.jsonanchor lags legitimately (informational line, per playbook).content/docs/references/kernel/metadata-plugin.mdxis pipeline-generated — regenerated viagen:docs, never hand-edited. Its diff includes sibling optionality flips (default-bearing fields now render required): explained and verified by ablation — the oldadditionalTypesembeddedMetadataTypeRegistryEntryBaseSchema(pullingActionSchema), which made output-shape JSON-Schema emission throw and forced the whole def into the input-shape fallback (x-io: input). With the key tombstoned the def emits output shape like the other ~1460 defs. Ablation proof: restoring the old key and regenerating reverts the page byte-identical to origin/main.content/docs/plugins/adding-a-metadata-type.mdx— all 4 occurrence sites (TL;DR, built-in-vs-plugin section, schema-wiring note, release checklist) rewritten to how a kind actually enters the live set: as a side effect of registering an item of that kind (SchemaRegistry.registerItem/MetadataManager.register), withregisterMetadataTypeSchemafor the schema half.packages/metadata/src/metadata-manager.ts(~2470, false "covers plugin-contributed additionalTypes" claim),packages/metadata-protocol/src/protocol.ts(~4376, kept the real artifact-load half, dropped the phantom),packages/spec/src/kernel/metadata-type-schemas.ts("register the type as well" note now states the real channel). Additionally two comments inmetadata-plugin.zod.tsitself (theMetadataTypeenum docblock andDEFAULT_METADATA_TYPE_REGISTRY) claimed plugins extend the registry viacontributes.kinds— verified false for that registry (registerKindstores kind descriptors as items of thekindtype; it does not extend the type registry); corrected in place since they advertise the same phantom growth path on the exact surface being retired.additional-types-retirement.test.ts: refusal with code/path/prescription (direct and through the manifestconfigembed) + control that the same config without the key parses and grows no property. The old spec test that authored the key re-judged: it merely exercised the alias-free full-config parse, so it was re-spelled (key removed) rather than replaced.minor+ BREAKING annotation + FROM/TO +adr-0087: registeredmarker (launch-window convention, PR feat(spec): refuse${…}placeholder syntax in memory persistence.path / persistence.key at publish (#8495) #8666 precedent;check-changeset-no-majorgreen).check:livenesswalks metadata type schemas (listMetadataTypeSchemaTypes);kernel/MetadataPluginConfigis not in that walk and has no ledger row to keep or delete.check:livenessgreen.*.form.tsinput ever spelled this key; no i18n change.Verification (all at HEAD
54ae32c17, post-merge of origin/main, after the final commit)check:generated13/13 green (migration-registry, spec-changes, upgrade-guide, skill-docs, skill-refs, react-blocks, authorable-surface, api-surface, export-origins, docs, strictness-ledger, liveness, test-typecheck).api-surface/unchanged — the tombstone is key-level narrowing, invisible to that snapshot by design; no exported value schema was orphaned (MetadataTypeRegistryEntryBaseSchemakeeps its consumer and is module-private).check:generatednames as not-run, run as one group:check:empty-state,check:skill-examples,check:template-manifests,check:variant-docs,check:exported-any,check:dual-source-exports,check:scripts-typecheck— all PASS.check:nul-bytesPASS.node scripts/pm/dispatch-gates.mjsover the actual changed paths — families beyond the prompt's list, all PASS:check:changeset-gate-self-tests,check:cross-package-test-inputs,check:doc-formula-expressions,check:docs-audit-scope,check:durability-log-level,check:filter-alias-parity,check:merge-driver,check:objectui-changeset,check:quick-reference-counts,check:role-word,check:spec-parsed-alias,check:type-source-resolution,check:query-options-erasure,check:type-check-coverage,check:type-check-debt(after fullturbo run buildof the packages closure, as lint.yml does),check-adr-0087-registration,check-changeset-no-major,check-empty-changeset,check-dev-prereqs.@objectstack/spec10572/10572 (399 files);@objectstack/metadata603/603 (31 files);@objectstack/metadata-protocol1317/1317 (89 files). Neither metadata package has atypecheckscript — theirbuild(tsup + dts) is the type gate, both green. Pin test 3/3.gen:schemared at gate (b2) with the exact message —1 RETIRED_KEYS_BY_MAJOR entr(ies) name a key that is still LIVE: kernel/MetadataPluginConfig:additionalTypes (registered at major 18). Restored from the commit; regeneration reproduces the committed artifacts byte-identically (cleangit status).additionalTypesauthors/writers in/home/user/objectuiand/home/user/cloud(control probe confirmed the sweep saw both trees).packages/qa/dogfood: nocoversentry references this key (no expression surface on it); the dogfood consumption-radius check was a grep plus the cross-package-test-inputs gate, both clean.Generated by Claude Code
Generated by Claude Code