无 format 的 autonumber 字段两侧渲染不同:driver-sql 兜底成 `{0000}` 发出 `0001`,引擎兜底路径发出裸 `1` —— 同一份元数据换驱动号形不同

[baozhoutao]

发现于 #6468 的实现(PR #6553)途中,未在该 PR 内顺手修 —— #6468 修的是播种时
怎么从已存值里读回计数器,本条是渲染兜底:两侧对「字段没写 format」这件事
给出的默认格式不同。机制不同、代码位置不同,按 Prime Directive #10 单独开单。

缺陷

autonumberFormat / format 在字段上是可选的。两侧对缺省的处理不一致:

• driver-sql 把空格式兜底成 {0000}:
  packages/drivers/driver-sql/src/sql-driver.ts:4322 与 :4404
  同一行 const fmt = rawFmt || '{0000}';(两处分别在 initObjects 与外部对象注册路径)。
  渲染结果是 4 位零填充 —— 0001、0002、…
• 引擎兜底路径(驱动不声明 supports.autonumber 时)不兜底,直接用空格式:
  packages/objectql/src/engine.ts:2115
const tokens = parseAutonumberFormat(typeof fmt === 'string' ? fmt : '');
  空 token 列表在 renderAutonumber 里走 width === null 分支
  (packages/spec/src/data/autonumber-format.ts:210-212),渲染出裸计数器 —— 1、2、…

实测

在 #6468 的工作树上,同一份元数据 { rec_no: { type: 'autonumber' } }(无 format),
库内已有 1 / 2 / 10 三行:

• SqlDriver.create() ⇒ rec_no = '0011'
• 引擎兜底路径(InMemoryDriver,supports = {})⇒ rec_no = '11'

两侧的计数器都是 11(播种一致,#6468 已 pin 住这一点),分歧纯粹在渲染宽度。
PR #6553 的 packages/drivers/driver-sql/src/sql-driver-autonumber-suffix.test.ts
里那条 legacy 用例已把 driver 侧的 '0011' 显式 pin 住,并在注释里点名这条分歧,
以免下一个读者把它误当成 #6468 的一部分。

可达性

不需要任何异常写法 —— 声明 { type: 'autonumber' } 且不写 format 即可。典型暴露面:
开发/测试跑 memory 驱动、生产跑 SQL 驱动,同一份元数据在两处发出的单号字面不同
(1 vs 0001),测试里断言的号形到了生产就不成立;数据迁移或换驱动时,同一对象
的历史号形也会在切换点前后突变。

供分诊参考(不预设结论)

三条路各有代价,需要先定「无 format 的 autonumber 该长什么样」这个契约问题:

1. 引擎跟随 driver-sql,空格式也兜底成 {0000};
2. driver-sql 跟随引擎,去掉 || '{0000}',无格式即裸计数器 —— 但这会改变已落地
   数据的号形(现存 0001 之后接 2),需要兼容判断;
3. 在 spec 侧把默认写进契约(例如 FieldSchema 对 autonumber 给出 format 的
   声明式默认),让「declared = enforced」,两侧都没有自己的兜底。

⚠️ 与 #6468(PR #6553,播种解析)同区域但不同函数:#6468 动的是 seedAutonumber /
scanMaxNumericTail 的读回,本条在 applyAutonumbers / initObjects 的格式
兜底。#6468 落地后再动本条不会冲突。

[os-zhuang]

Triage: needs-user-decision + domain:spec.

• Classification: the divergence is real and measured, but every fix route changes a shipped default number shape for someone — route 1 flips engine-fallback deployments from bare 1 to 0001, route 2 changes driver-sql's already-landed numbering mid-sequence, route 3 writes the default into the contract. "What does a format-less autonumber render as" is a product-semantics / public-contract call with stored-data continuity implications ⇒ maintainer inbox, not the queue. No prior ruling found: 自增号格式带「序号槽之后的后缀」时,播种解析读错数字段:引擎读成年份(2026),driver-sql 读成拼接串(12026),两侧还互不一致 #6468's scope is counter seeding, and the open decision card 决策:自增号「计数器反解」规则是否作为 renderAutonumber 的逆进 packages/spec(PR #6553 的 open question,维护者裁) #6560 is about the counter-inverse rule — neither answers the render default.
• Landing anchors verified on origin/main @ f549a0d: packages/drivers/driver-sql/src/sql-driver.ts:4330/:4412 — const fmt = rawFmt || '{0000}'; (both sites); packages/objectql/src/engine.ts:2115 parses the empty string when no format; the width === null branch in packages/spec/src/data/autonumber-format.ts renders the bare counter and its own comment calls that "legacy behaviour". Body's cited lines (:4322/:4404) drifted; code unchanged.
• Routing rationale: the deciding surface is the contract default — renderAutonumber and any declared default live in packages/spec/src/data/, and a declared default changes normative metadata semantics ⇒ domain:spec under the acceptance-face criterion (not spec-surface). If the ruling picks route 1 or 2, the execution card re-routes to engine-core/drivers at that point.
• Dedup: autonumber family 自增号格式带「序号槽之后的后缀」时,播种解析读错数字段:引擎读成年份(2026),driver-sql 读成拼接串(12026),两侧还互不一致 #6468 (dispatched), 决策:自增号「计数器反解」规则是否作为 renderAutonumber 的逆进 packages/spec(PR #6553 的 open question,维护者裁) #6560, Autonumber counter neither syncs to MAX(existing) per tenant nor re-checks on collision — warm-DB creates 409 in bursts, each failure burning a number (25 retries observed) #5495, [观察] autonumber builder 不注入 readonly: true —— readonly 契约的「表单不可编辑」那一半对单号字段不成立 #5628, insert 路径 stripRuntimeOwnedFields 同样删的是 hook 覆写后的当前值 —— beforeInsert hook 写入的 autonumber 在调用方也提交了同名键时被连坐删除(与其自身文档矛盾) #6339 — none covers the render-fallback divergence.
• Not placed on the v17 board: both sides agree on the counter value (自增号格式带「序号槽之后的后缀」时,播种解析读错数字段:引擎读成年份(2026),driver-sql 读成拼接串(12026),两侧还互不一致 #6468 pins seeding parity); the divergence is number formatting across drivers — loud in tests, no data loss. Flag if you read it as ground ①.

本评论来自分诊座位 Routine（#5474 试点），不构成认领。

---

Generated by Claude Code

[os-zhuang]

Maintainer ruling (2026-08-08): Route 3, with the default fixed at {0000}. The default format for autonumber fields becomes a declared contract default in FieldSchema (spec side), and both hand-written fallbacks go away: driver-sql drops its hardcoded fallback and the engine fallback path stops emitting bare integers — both sides read the declared default.

Rationale (three-axis review): routes 1 and 2 both fix the fork but leave "the default" living in two hand-written fallbacks — and route 2 would break number-shape continuity for already-stored SQL data, the largest install base (business). Route 3 is the only one that puts the default itself into the contract: declared = enforced (long-term), and an AI metadata author sees the default in the schema instead of guessing per-driver behavior (containment).

Compatibility note (must be in the changeset): choosing {0000} keeps stored driver-sql data undisturbed; engine-fallback deployments flip from bare 1 to 0001 for newly issued numbers. Counter continuity itself is unaffected (#6468 pinned it).

For the dispatch: contract-first ordering — the spec default lands first (spec lane), the two de-fallback changes follow (engine-core / drivers as routed by triage).

State: needs-user-decision removed; pm:queue restored — spec lane leads.

Maintainer directive (verbatim, covering all 14 inbox cards): 「你的建议全部接受」. Recorded by PM session session_01JaVVMrSxt7Tgi1uwEuDtH7.

---

Generated by Claude Code