Release close-out (attended, 2026-10-25): hub 0.170.0, agent 0.165.0 (vouched), controller 0.343.0 (global floor), catalog ebbf1f4; D49 = A and D44–D46 recorded (09 §3 258–261, CONTEXT, 07 §6.3/§8.2e); 9 rows closed, R-1254 hub half, R-1250 dated check 2026-10-27; R-1287–R-1290 filed; STATUS, map, REPORT-release-2026-10-26

documentation/backlog/STATUS-HISTORY.md

@@ -7,6 +7,13 @@ Moved here byte for byte on 2026-10-14 (night session), newest first, so that `S
 
 ## Moved from STATUS on 2026-10-25 (attended release run)
 
+**Updated 2026-10-25 (Sunday).** A reinstall keeps old backups whole; a restore downloads first and tells the truth.
+
+Decided the same day: **D49 = A**, **D46 = yes** (`09` §3 decisions 258, 261):
+
+| D49 | R-1270 | A restore's data is back but the app does not start in 90 s. Keep the restored data and wait, or pour the old data back by itself? | **Keep and wait** (a slow first start is common; the old copy is kept, a late start turns it into a success; you get one alarm) | pour back by itself: a slow but correct restore is thrown away | the data stays restored, the old copy stays for support | — |
+| D46 | R-1272 | Put Homepage behind the family gate, like Glance? | **Yes** (a stranger now sees the box's CPU, memory and disk, and later the family's links; one line in the template, tried on the drill first) | leave it open: anyone who finds the address sees those | it stays open to strangers | — |
+
 The Monday list, run on 2026-10-25 (attended; results in `REPORT-release-2026-10-26.md`):
 
 **Monday 2026-10-26, release list (attended).** 1. Controller v0.343.0: the restore and backup fixes (R-1251–R-1253, R-1269–R-1271), R-1250. If not: the P1 stays open. 2. Agent v0.165.0 + your vouch: BYO keeps its SSH (R-1257). 3. Hub v0.170.0: the recovery package in one step (R-1254). 4. Catalog: n8n usage data off; OpenGist 256 MB.