feat(spec)!: 退役 DriverCapabilities 31 个零读者能力位 —— 全表活性审计,3 活 31 死 (#4634)

#4484 findStream 收尾发现 streaming 位指向已删除的契约方法;按 issue 自述倾向对
DriverCapabilitiesSchema 全部 34 位做两仓活性审计(objectstack + cloud,objectui
零引用确认):仅 queryDateGranularity(engine 聚合分派 + date-bucket parity)、
autonumber(engine 让渡自增生成)、batchSchemaSync(engine 与方法存在性 AND)
三位有决策读者;其余 31 位人人写、无人读,describe 承诺的引擎降级路径从未存在,
且零读者让写错的值(SqlDriver streaming:false 却实现 findStream;InMemoryDriver
streaming:true 却全表物化)三个版本无人发现 —— ADR-0078 假可供性。

- driver.zod.ts:31 位 retiredKey() 墓碑(非 strict schema,裸删会被静默剥离;
  DriverConfigSchema.capabilities 及 SQL/NoSQL 扩展确有 parse 路径),逐位
  prescription 指明真实机制(方法存在性门控:beginTransaction / aggregate /
  syncSchema;#4001 教训);batchSchemaSync .default(false) → .optional()
  (两处读者本就按缺省=false 消费)
- ADR-0087 评估(#4484/#4764 先例):driver 是代码不是 stack 树元数据,
  supports 字面量在 driver 类里、DriverConfig 是插件 TS 配置,均非 sys_metadata
  形态(stack 树近邻 datasource.capabilities 已于 #4583 单独退役)⇒ 不登记 D2
  conversion,登记 D3 semantic migration driver-capabilities-inert-bits-removed
- driver 字面量同步:memory {} / mongodb { batchSchemaSync: true } /
  sql { queryDateGranularity, autonumber, batchSchemaSync: false }
- 回归 pin(driver.test.ts):31 位逐位 parse 拒绝 + prescription 断言;
  compiler-API 断言 34 键形状且退役位类型坍缩为 undefined、活位可写(反空转
  守卫);S1(schema 复活 streaming)5 pin 红、S2(driver 字面量复活)TS2416
  红,均实跑验证后还原
- 生成物:authorable-surface 31 行 [RETIRED]、spec-changes / upgrade-guide /
  references docs 全量再生成;14 项 spec 门禁全绿,全仓 typecheck 122 任务绿
- changeset:spec + 三 driver 包 major,FROM→TO 映射与一行修复齐备

Co-Authored-By: Claude <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0176qgxgCXTJCUv4YFLtusP9

Author: claude

.changeset/driver-capabilities-inert-bits-removed.md

@@ -0,0 +1,78 @@
+---
+"@objectstack/spec": major
+"@objectstack/driver-memory": major
+"@objectstack/driver-mongodb": major
+"@objectstack/driver-sql": major
+---
+
+refactor(spec)!: retire the 31 inert `DriverCapabilities` bits — declared by every driver, read by nothing (#4634, ADR-0049)
+
+The #4484 findStream close-out left one loose end: `DriverCapabilities.streaming`
+described a contract method that no longer exists — and a full liveness audit of
+the record (#4634, across objectstack + cloud, objectui confirmed clean) found
+`streaming` was not the exception but the rule. Of 34 declared bits, **three**
+have a decision-making reader and **thirty-one** were written by every driver
+and consulted by no engine, planner, REST layer or renderer:
+
+- Their `.describe()` strings promised engine adaptation that was never built
+  ("If false, ObjectQL will fetch all records and filter in memory" — no such
+  fallback ever keyed off the bit).
+- Zero readers let values go WRONG unnoticed: `SqlDriver` declared
+  `streaming: false` while implementing `findStream`; `InMemoryDriver` declared
+  `streaming: true` over a full-table read — the exact inverse of the guarantee.
+- The real mechanism everywhere else is **method presence**: transactions gate
+  on `driver.beginTransaction`, aggregate pushdown on
+  `typeof driver.aggregate === 'function'`, schema sync on
+  `typeof driver.syncSchema === 'function'`, and the REQUIRED CRUD/bulk methods
+  are called unconditionally.
+
+Survivors (each with a named reader — the bits method presence cannot carry):
+
+| bit | reader |
+|---|---|
+| `queryDateGranularity` | engine aggregate dispatch (`engine.ts`), `checkDateBucketParity` (`@objectstack/verify`) |
+| `autonumber` | engine defers autonumber generation to the driver (`engine.ts`) |
+| `batchSchemaSync` | engine ANDs it with `syncSchemasBatch` presence (`engine.ts` / `plugin.ts`) |
+
+Migration (FROM → TO):
+
+- Any of the 31 bits (`create`/`read`/`update`/`delete`, `bulkCreate`/
+  `bulkUpdate`/`bulkDelete`, `transactions`/`savepoints`/`isolationLevels`,
+  `queryFilters`/`queryAggregations`/`querySorting`/`queryPagination`/
+  `queryWindowFunctions`/`querySubqueries`/`queryCTE`/`joins`,
+  `fullTextSearch`/`jsonQuery`/`geospatialQuery`/`streaming`/`jsonFields`/
+  `arrayFields`/`vectorSearch`, `schemaSync`/`migrations`/`indexes`,
+  `connectionPooling`/`preparedStatements`/`queryCache`) in a `supports`
+  literal or a `DriverConfig.capabilities` object → **delete the key**. Each is
+  tombstoned (`retiredKey()`), not silently stripped: authoring one is a `tsc`
+  error against `IDataDriver.supports` and a parse error carrying the per-key
+  prescription, which names the mechanism that actually decides the behaviour.
+- `batchSchemaSync` dropped its `.default(false)` for `.optional()` — absence
+  already meant `false` at both readers, so `supports: {}` is now a valid,
+  minimal advertisement. If you read `capabilities.batchSchemaSync` from a
+  *parsed* config and relied on the materialised `false`, treat absence as
+  `false` (both engine readers always did).
+- Driver packages: `InMemoryDriver.supports` is now `{}`,
+  `MongoDBDriver.supports` is `{ batchSchemaSync: true }`, `SqlDriver.supports`
+  is `{ queryDateGranularity, autonumber: true, batchSchemaSync: false }`.
+  Reading a removed bit off these literals no longer type-checks — and no code
+  in any repository did.
+- A future capability (streaming reads, vector search, …) returns **with its
+  caller and its reader in the same change** — the enforce route of ADR-0049 —
+  never as a dangling boolean.
+
+The retirement kit: 31 `retiredKey()` tombstones on the non-strict schema
+(parse + `tsc` both audible; the schema IS parsed via
+`DriverConfigSchema.capabilities` and its SQL/NoSQL extensions); ADR-0087 D3
+semantic migration `driver-capabilities-inert-bits-removed` (a driver is CODE,
+never stack metadata — `supports` lives in driver classes and `DriverConfig`
+is plugin TS configuration, so there is no stored row or stack source for a D2
+conversion to rewrite; the stack-tree neighbour `datasource.capabilities` was
+retired separately in #4583); baselines (`authorable-surface.json` [RETIRED]
+lines, `json-schema.manifest.json`) regenerated deliberately; compiler-API pin
+asserting every retired bit is unwritable (`undefined`) and every live bit is
+not, sabotage-verified both ways (S1 schema resurrection, S2 driver literal
+resurrection).
+
+No runtime behaviour changes — that impossibility is the point: every removed
+bit had zero readers, and the three live bits keep theirs.

content/docs/protocol/objectql/query-syntax.mdx

@@ -749,8 +749,9 @@ form.
 The `search` parameter does **not** reach a full-text index. The engine expands it into
 an `$or` of `$contains` predicates across the object's server-resolved searchable fields
 (ADR-0061) and deletes `search` from the AST before the driver sees it — every driver
-already runs `$or`/`$contains`, so no driver support is needed (`SqlDriver` reports
-`supports.fullTextSearch: false`).
+already runs `$or`/`$contains`, so no driver support is needed (which is also why the
+old `supports.fullTextSearch` capability bit had no reader and was retired in 17.0.0,
+#4634).
 
 `search` takes the **query text itself** — that is the canonical spelling (ADR-0061 D1:
 the client says *what* to search for, the server decides *which fields*), and it is what

content/docs/references/data/driver-nosql.mdx

@@ -123,7 +123,7 @@ const result = AggregationPipelineSchema.parse(data);
 | :--- | :--- | :--- | :--- |
 | **name** | `string` | ✅ | Driver instance name |
 | **type** | `'nosql'` | ✅ | Driver type must be "nosql" |
-| **capabilities** | `{ create: boolean; read: boolean; update: boolean; delete: boolean; … }` | ✅ | Driver capability flags |
+| **capabilities** | `{ queryDateGranularity?: Record<string, boolean>; autonumber?: boolean; batchSchemaSync?: boolean; create?: any; … }` | ✅ | Driver capability flags |
 | **connectionString** | `string` | optional | Database connection string (driver-specific format) |
 | **poolConfig** | `{ min: number; max: number; idleTimeoutMillis: number; connectionTimeoutMillis: number }` | optional | Connection pool configuration |
 | **databaseType** | `Enum<'mongodb' \| 'couchdb' \| 'dynamodb' \| 'cassandra' \| 'redis' \| 'elasticsearch' \| 'neo4j' \| 'orientdb'>` | ✅ | Specific NoSQL database type |

content/docs/references/data/driver-sql.mdx

@@ -65,7 +65,7 @@ const result = DataTypeMappingSchema.parse(data);
 | :--- | :--- | :--- | :--- |
 | **name** | `string` | ✅ | Driver instance name |
 | **type** | `'sql'` | ✅ | Driver type must be "sql" |
-| **capabilities** | `{ create: boolean; read: boolean; update: boolean; delete: boolean; … }` | ✅ | Driver capability flags |
+| **capabilities** | `{ queryDateGranularity?: Record<string, boolean>; autonumber?: boolean; batchSchemaSync?: boolean; create?: any; … }` | ✅ | Driver capability flags |
 | **connectionString** | `string` | optional | Database connection string (driver-specific format) |
 | **poolConfig** | `{ min: number; max: number; idleTimeoutMillis: number; connectionTimeoutMillis: number }` | optional | Connection pool configuration |
 | **dialect** | `Enum<'postgresql' \| 'mysql' \| 'sqlite' \| 'mssql' \| 'oracle' \| 'mariadb'>` | ✅ | SQL database dialect |