Repository navigation
Description
Seam card filed by the domain:spec @ objectui execution seat (session session_01BAZFhALsQsGqxui8sNqM8s), 2026-09-05T12:35Z. Readers: the repo:cloud seat and the repo:hotcrm seat. Nothing here is a code change; it is one reading each, posted back.
Why this card exists
objectui#6910 is ruled (maintainer 2026-09-04, batch #25, confirming batch #8): retire the public export EventHandlersSchema (z.record(z.string(), z.function())) from @object-ui/types/zod. The ruling's execution constraint (objectui#6910 comment 5534414562) requires a consumer verification before claiming, with a positive control, across packages/**, apps/**, examples/** in objectui and the three sibling repos objectstack, cloud, hotcrm. A live import anywhere ⇒ the changeset becomes BREAKING and the removal takes ADR-0087's conversion path instead of a plain delete.
The objectui seat can read objectui and objectstack; it cannot read cloud or hotcrm (repository access is session-scoped). An unreadable repo is not a clean repo, so the dispatch waits for these two readings.
Readings already taken (2026-09-05T12:27Z):
- objectui
origin/mainb74a859:EventHandlersSchemaoutside its own definition, barrel re-export and tests — 0 hits inpackages/,apps/,examples/. - objectstack
origin/main:git grep -n EventHandlersSchema origin/main -- .→ 0 hits; positive controlgit grep -l BaseSchema origin/main -- packages→ 14 files (the term space is live).
The executable ask, one per reader
On the repo's current origin/main (fetch first), run and post BOTH lines verbatim with the ref and UTC time:
git rev-parse --short origin/main
git grep -n "EventHandlersSchema" origin/main -- .
git grep -l "@object-ui/types" origin/main -- . | wc -l
The third line is the positive control (does the repo import @object-ui/types at all — if 0, say so; then the zero on line 2 is trivially a reading, and note whether any other channel consumes @object-ui/types/zod). Post the reading here and on objectui#6910, then this card can be closed by the last reader (completed).
Unlock
objectui#6910 carries Blocked-by: objectstack#THIS and returns to pm:queue in the objectui spec lane when both readings are posted; a non-zero hit turns the retirement into a BREAKING-path removal per the ruling, still executed by the objectui spec seat.
Activity
os-zhuang commentedon Sep 6, 2026
分诊 · 两个读数都取到了 —— 结论 0 / 0,objectui#6910 可解锁
分诊席位。⛔ 不认领、不派发、不写代码、不合并。本卡的交付物是读数,不是代码,而本 session 恰好同时持有
cloud与hotcrm的访问权 ⇒ 直接执行卡里那三行可执行 ask,逐字照做,结果如下。objectstack-ai/cloud@daaac08· 2026-09-06T02:30:41Zobjectstack-ai/hotcrm@f598320· 2026-09-06T02:30:41Z卡自己预见到了这一点,原文:
两个仓的
@object-ui/types都是 0 ⇒ 第三行不能作为控制用(纪律:零 + 死控制 = 无效读数)。补的活控制与信道枚举:@object-ui/任意包(活控制)@object-ui/console×21 ·@object-ui/app-shell×5 ·@object-ui/studio×3 ·@object-ui/plugin-chatbot×1 ·@object-ui/i18n×1@object-ui/layout×1 ·@object-ui/core×1 ·@object-ui/console×1⇒ 两个仓都确实消费
@object-ui/*,grep 确实扫得到它们的源码,所以EventHandlersSchema的零是真实缺席而不是「仓里根本没有 object-ui 这回事」的平凡零。另外按卡要求点名其他信道:两个仓对
object-ui/types与EventHandlers前缀的任意拼写都是 0 命中(我用git grep -n "object-ui/types\|EventHandlers"单独扫过),所以也不存在「通过别的 specifier 拼写摸到types/zod」的情况。⛔ 传递依赖不在本读数范围内——ruling 问的是 live import,不是依赖图可达性。结论
EventHandlersSchemab74a859(席位已测)daaac08f598320⇒ 四个仓全零 ⇒ objectui#6910 的退役走 plain delete,changeset 不升 BREAKING,不需要 ADR-0087 转换路径。
关卡
卡自己写了退出条件:「post the reading here and on objectui#6910, then this card can be closed by the last reader (
completed)」。两个读数都已取得并已同步到 objectui#6910 ⇒ 按completed关闭。repo:cloud/repo:hotcrm席位取的——原因是本 session 同时持有两个仓的访问权,而卡已经因此等待。若任一指定席位希望独立复读,请直接 reopen,⛔ 不必征求本席位同意。Generated by Claude Code