hub (unreleased): the delete journal keeps no id after one year — emptied in the transaction of the year-end prune (D86, decision 314)

Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS
This commit is contained in:
admin committed 2026-11-15 15:57:10 +01:00
1 parent b07e4af0c1
commit 0fa1fe56d6
6 files changed
+393 -31

No files matched your search

@@ -454,7 +454,7 @@ or a household. The ruling: a delete removes them, with one exception — the de
| `offsite_large_requests` (the hub's own first sight of a large removed-app request; `07` §8.2f — built, not released) | the household | the household is deleted; the daily orphan sweep | `store.PurgeCustomerResidue`, `store.PurgeOrphanRows`. At most 50 rows per household |
| `offsite_app_presence` (the hub's own record of which apps a box runs; D72, `07` §8.2f — built, not released) | the household | the household is deleted; the daily orphan sweep takes what an earlier delete left | `store.PurgeCustomerResidue`, `store.PurgeOrphanRows`. At most 400 names per household; no age limit on a row (it would move a real app's 30 days) |
| `operator_actions` (the operator's own action log) | the household | **one year after the deletion**, with `events` and `notification_log` | `store.PruneDeletedCustomerAudit`. *Decided by CC unattended on 2026-11-14; accepted by the operator on 2026-11-15 (`09` §3 decision 311):* it is an audit record of the same kind as the other two, and the one-year rule (decision 181) is the design's rule for those; the other option, at once, loses „who did what” on the day of the deletion |
| **`host_deletions`, `customer_resets`** — the delete journal | both | **never** | the ruling's one exception: an id and timestamps. **An id can be a name, so this is not „no personal data”** — the design to change that is §14.3, and the privacy text waits for it (`09` §3 decision 312) |
| **`host_deletions`, `customer_resets`** — the delete journal | both | **never** | the ROWS stay for ever; **their ids go one year after the household's deletion** (§14.3, D86 — built, not released). Until that release the ids stay, and an id can be a name |
**Rows left by boxes and households deleted before this release** (Peti's box, the test boxes of 2026-11): the daily
prune runs `store.PurgeOrphanRows` — the six tables of the first and third rows, for a box with no `hosts` row and for
@@ -469,7 +469,7 @@ its own design, because the kernel lane reads a box's old reports to know which
(`TestD73_*`), 16 mutations seen red (`audits/night-build-2026-11-14/B/redproof-hub-d73.txt`).
### 14.3 The delete journal keeps an id that is not a name `[DESIGN — proposed, not built]` (`09` §3 decision 312; written 2026-11-15)
### 14.3 The delete journal keeps no id after one year `[FACT — built, not released: hub `main` after v0.182.0, 2026-11-15]` (D86 = option A, `09` §3 decisions 312 and 314; designed 2026-11-15)
**Why.** §14.2 says the delete journal stays for ever and holds „an id and timestamps, no personal data”. The second
half is not true. An id of a household or of a box is typed by a person and can be a name: one deleted box carried
@@ -531,8 +531,39 @@ thing the sentence must not promise away.
holds its id or its boxes' ids (searched as text, with a look-alike id that must stay); the rows are still there with
their times; a household deleted 364 days ago keeps its id; a household that still exists keeps everything.
**If nothing is decided:** the journal keeps ids that can be names with no end, and the privacy sentence stays
unpublished.
**Ruled 2026-11-15 evening (D86 = remove, option A).**
**As built (`store.PruneDeletedCustomerAudit`, `pruneOneDeletedCustomer`).** At a deleted household's year-end, in ONE
transaction: its events, mail log and action log are deleted, and the journal rows of that deleted life lose their ids
— `customer_resets.customer_id`, `host_deletions.customer_id` and `host_deletions.host_id` become the EMPTY string.
The rows, their times and their step states stay.
Four things the design above did not say, found by reading it against a broken and a hostile case before the build:
| Case | What is built | Why |
|---|---|---|
| The placeholder | the empty string, never a word („removed”) | a word can be typed as a new household's id; the next day's run would then take that household for a deleted one and delete its log. An id is never empty |
| An id that is used AGAIN for a new household (the test household was deleted and made again three times in 2026-11) | only the rows of the DELETED life are emptied: journal rows started, and box deletes made, at or before that deletion's completion | the new household's box deletes and RESETs are read by live code (the key-package carry, a later delete's purge) |
| The emptied row at the next daily run | the run skips rows with no id | it would otherwise run the year-end for the household „” every day |
| `legs_json` | at the year-end every key and value that is not a plain lower-case word becomes „other” | rule 5 asked for a test that the field holds state words only; a test cannot bind a later writer, the year-end can |
| Rule above | Test (`hub/internal/store/deleted_customer_audit_test.go`) |
|---|---|
| after the year no journal row holds the household's or its box's id; the rows and times stay; a look-alike id, a household 364 days after its delete, a household whose delete never ended and a household that exists keep everything | `TestD86_AfterTheYearTheJournalHoldsNoIDAndKeepsItsRows` |
| an id used again keeps the journal of its new life | `TestD86_AnIDUsedAgainKeepsTheJournalOfItsNewLife` |
| the emptied row is never read as a household; no word in place of the id | `TestD86_TheEmptiedRowIsNeverReadAsAHousehold` |
| the step field holds plain words only after the year | `TestD86_TheStepFieldKeepsOnlyPlainWords` |
| the log and the id go in one transaction (rule 2) | `TestD86_TheLogAndTheIDGoInOneTransaction` |
15 mutations seen red, one of them only after the test was made to ask for the staged error
(`audits/build-release-4b-2026-11-15/C/redproof-hub-journal-id.txt`). **Not proven:** on a deployed hub — and no
household's year ends before 2027-09-25, so the first real run is then.
**Rule 4 (a household whose delete never ended) is NOT built as a page line:** such a household keeps its ids (its
year has not started), and nothing on the System page names it — R-1403.
**The privacy sentence is still not published** (decision 314): it goes to the operator with the lawyer rows
(R-802, R-813). After this build the sentence above is true for a hub that runs it.
## 15. Customer e-mails — what the hub writes to a household [hub v0.118.0, R-558]