Gladex Agent Logs

3121:## 2026-09-24T03:45:00Z — STEP 0 clean + shipped the queued SQLITE_BUSY fix: busy_timeout on every DB open (0.4.18)
3199:## 2026-09-24T04:21:29Z — STEP 0 clean + shipped the queued PHP busy_timeout port (0.4.19)
=== prod unread ===
0
=== timers ===
Thu 2026-09-24 06:53:29 CEST     7min Thu 2026-09-24 06:38:29 CEST     7min ago agent-loop-watchdog.timer      agent-loop-watchdog.service
Thu 2026-09-24 08:06:05 CEST 1h 20min -                                       - identity-run@lena.timer        identity-run@lena.service
Thu 2026-09-24 08:11:23 CEST 1h 25min -                                       - identity-run@jonas.timer       identity-run@jonas.service
Thu 2026-09-24 08:15:30 CEST 1h 29min -                                       - identity-run@mia.timer         identity-run@mia.service
Thu 2026-09-24 08:20:16 CEST 1h 34min -                                       - identity-run@leon.timer        identity-run@leon.service
Thu 2026-09-24 08:25:31 CEST 1h 39min -                                       - identity-run@aylin.timer       identity-run@aylin.service
Thu 2026-09-24 08:50:29 CEST  2h 4min -                                       - identity-run@dispatcher.timer  identity-run@dispatcher.service
=== budget ===
month: 2026-09
allowance: 5.00
spent: 1.50 (gladex.de — investor-approved 2026-09-20, purchase pending at registrar)
remaining: 3.50
=== git status ===
 M app/src/php/stats.php
?? tests/test_stats_dbpath.php
> build · mimo-v2.6-flash-free
All checks green. Now the CHANGELOG entry and PROGRESS.md record:
← Edit repo/CHANGELOG.md
Index: repo/CHANGELOG.md
===================================================================
--- repo/CHANGELOG.md
+++ repo/CHANGELOG.md
@@ -3,8 +3,25 @@
 All notable changes to Gladex are documented here.
 
 Format: `[version] - YYYY-MM-DD — description`
 
+## [0.4.20] - 2026-09-24 — prod /stats was showing DEV's message counts: the hardcoded DB path
+
+### Fixed
+- **`stats.php` hardcoded `/opt/startup/dev/data/messages.db`** — the item queued at the end of 0.4.19. The Message counts on `/stats` are per-env data, but the path was a constant pointing at **dev**, so **prod `/stats` displayed dev's message counts**. The two DBs had already diverged (dev **57** rows = 24 investor→agent + 33 agent→investor; prod **22** = 3 + 19), making the bug visible: prod claimed 57 messages while its own DB held 22, and would never move as the prod thread grew.
+  - Fix: `$dbPath = __DIR__ . '/../../data/messages.db';` — resolved from the file's own location, exactly the idiom `app.php` and `init_db.php` already use. With the deployed layout `/opt/startup/<env>/src/php/`, `../../data` is **that env's** data dir, so the same bytes serve both envs correctly.
+
+### Added
+- `tests/test_stats_dbpath.php` — **16 assertions, ~0.7 s**, two independent layers:
+  - **static**: none of the 3 copies (repo/dev/prod) may contain the hardcoded path, all must carry the `__DIR__` expression — a stale deployed copy fails the suite;
+  - **behavioural**: renders `stats.php` as a script inside a synthetic env tree (`fixture/src/php/{stats,db}.php` + `fixture/data/messages.db` seeded from the repo's own `schema.sql` with **known 7/3/4 rows**) and asserts the page prints **7/3/4** — i.e. its own DB — plus a **non-vacuity guard** (fixture 7 ≠ live dev 57, so "not reading the hardcoded path" is actually provable).
+
+### Verified
+- **Mutation check**: restoring the hardcoded path → behavioural layer prints dev's **57/24/33** instead of 7/3/4 (**10 FAIL, exit 1**); restored byte-identically (md5 `decf1849…` re-matched) → **16/16**.
+- **Suites**: PHP **223** (78+49+27+19+16+16+11+7) + shell **26** → **249 passed, 0 failed**.
+- **Deploy (reviewer-gated)**: repo → **dev** (md5-identical, unit restarted, `/stats` 200 with dev's own 57/24/33) → reviewer subagent **APPROVE** (it re-ran the suites itself and confirmed prod was serving dev's 57/24/33 pre-fix) → **prod** (restart; `/stats` now **22/3/19** — its own counts, the direct before/after proof). repo == dev == prod md5 `decf184964196c8b904b81471842a21e`; 14 routes (`/healthz` `/` `/investor` `/stats` `/team` `/info` `/api/changelog` × dev+prod) all **200**; `tools/healthcheck both` → **HEALTHY / HEALTHY**.
+- **STEP 0**: **0 unread** `investor_to_agent` rows (dev **and** prod), **0 unhandled** `INBOX.md` entries → no reply rows owed. No DNS write, no config change, no API key, no spend (2026-09: 5.00 / 1.50 / 3.50, **model spend 0.00**, `*-free` only).
+
 ## [0.4.19] - 2026-09-24 — PHP gets the same `busy_timeout` contract: one helper behind every SQLite open
 
 ### Fixed
 - **PHP SQLite opens carried no explicit busy-timeout** — the item queued at the end of 0.4.18 ("carry the same `busy_timeout` into the PHP-side SQLite opens"). Two traps made this worth doing carefully rather than copying the Go line:

$ date -u +%Y-%m-%dT%H:%M:%SZ; echo "---"; ls /data/repo/mailboxes/*/ 2>/dev/null | grep -v gitkeep | grep -v '^$' | head; echo "(mailboxes checked)"
2026-09-24T04:45:58Z
---
/data/repo/mailboxes/main-to-reviewer/:
/data/repo/mailboxes/main-to-scout/:
/data/repo/mailboxes/reviewer-to-main/:
/data/repo/mailboxes/scout-to-main/:
(mailboxes checked)
> build · mimo-v2.6-flash-free
Now append the run record to PROGRESS.md, re-verify the changelog parses, then commit:
$ cat >> /data/repo/agent-logs/PROGRESS.md <<'EOF'

## 2026-09-24T04:46:00Z — STEP 0 clean + shipped the queued stats.php env-path fix (0.4.20)

**STEP 0 (done before anything else)** — queried `investor_to_agent AND read=0` in
`/opt/startup/dev/data/messages.db` **and** the prod mirror: **0 unread rows** (dev
24/24 read, prod 3/3 read); `INBOX.md`: every `## ` entry carries `~~HANDLED~~`,
0 unhandled → **no reply rows owed this run** (nothing unanswered, nothing to insert).
The standing open asks remain mine→theirs and were NOT re-spammed: **msg #57 A/B/C
public-gating decision** + **SOA MNAME placeholder** (both still unread *by the
investor*) — re-surfaced under NEEDS-INVESTOR below.

**This run's step: the queued item (1) from 0.4.19 — `stats.php` hardcoded
`/opt/startup/dev/data/messages.db`, so prod `/stats` reported DEV's message counts.**
One small, fully verified change.

- **Why it was live-wrong, not theoretical**: the DBs had already diverged — dev
  **57** rows (24 investor→agent, 33 agent→investor) vs prod **22** (3, 19). Prod's
  `/stats` claimed 57 messages while its own DB held 22, and would never move as the
  prod thread grew. `app.php`/`init_db.php` already used `__DIR__ . '/../../data/…'`;
  `stats.php` was the odd one out.
- **Fix (1 line + 3-line comment)**: `$dbPath = __DIR__ . '/../../data/messages.db';` —
  with the deployed layout `/opt/startup/<env>/src/php/`, `../../data` resolves to
  **that env's** data dir, so the same bytes are correct in repo, dev and prod.
- **New test `tests/test_stats_dbpath.php` (16 assertions, ~0.7 s)**, two layers:
  - **static**: no copy (repo/dev/prod) may contain the hardcoded path; all must carry
    the `__DIR__` expression — a stale deployed copy fails the suite;
  - **behavioural**: renders `stats.php` as a script in a synthetic env tree
    (`fixture/src/php/{stats,db}.php` + `fixture/data/messages.db` seeded from the
    repo's own `schema.sql` with **known 7/3/4 rows**) and asserts the page prints
    **7/3/4** (its own DB), plus a **non-vacuity guard** (fixture 7 ≠ live dev 57).
- **Mutation-verified**: restoring the hardcoded path → behavioural layer printed dev's
  **57/24/33** instead of 7/3/4 → **10 FAIL, exit 1**; restored byte-identically
  (md5 `decf184964196c8b904b81471842a21e` re-matched) → 16/16. Tests are load-bearing.
- **Pre-promote suite run**: exactly **3 failures, all prod-pending** (identity guard in
  `test_changelog_api` + the 2 static prod assertions) — nothing else red.
- **Deploy, reviewer-gated (standing rule)**: repo → **dev** (md5-identical, unit
  restarted, `/stats` 200 = dev's own 57/24/33) → reviewer subagent **VERDICT: APPROVE**
  (it re-ran the suites itself and confirmed pre-fix prod served 57/24/33) → **prod**
  (restart; `/stats` now **22/3/19** — its own counts: the direct before/after proof).
  Note: the reviewer executed the copy+restart itself right at its step limit; I
  **independently re-verified everything after it** — nothing taken on trust.
- **Post-promote verification (mine)**: repo == dev == prod md5 `decf1849…`; both units
  `active`; 14 routes (`/healthz` `/` `/investor` `/stats` `/team` `/info`
  `/api/changelog` × dev+prod) all **200**; prod Messages = **22/3/19**, dev = **57/24/33**;
  full suites **249 passed, 0 failed** (PHP 223 = 78+49+27+19+16+16+11+7, shell 26);
  `tools/healthcheck both` → **HEALTHY / HEALTHY**. CHANGELOG **0.4.19 → 0.4.20**.

**Health / rote duties this run**: `mailboxes/*` → **0** pending Dispatcher assignments
(only `.gitkeep`s); all 6 identity timers armed (next 08:06 / 08:11 / 08:15 / 08:20 /
08:25 / 08:50 CEST), watchdog firing (7 min cadence); units dev/prod **active**;
**no spend** — BUDGET 2026-09 unchanged 5.00 / 1.50 / 3.50, **model spend 0.00**
(`*-free` only), no DNS write, no config change, no API key read or touched.

**Discovered while verifying (queued for next run)** — two same-family observations,
neither user-visible today:
1. **`stats.php`'s "Go CLI" row reads `/opt/startup/dev/bin/gladex` only**, even when
   the page is rendered by prod — harmless while both binaries are the same build,
   misleading the moment they diverge. Candidate: read the env's own binary like the
   DB fix just did.
2. **`tools/system-status` counts unread investor messages from the DEV db only**,
   while the investor reads the prod chat — a read on prod would leave dev's counter
   stale (alert noise, not a data bug). Candidate: count both, or the env serving it.

**NEEDS-INVESTOR (unchanged, still open)**: (a) the **A/B/C public-gating decision**
(msg #57 — investor pages/chat/webmail are world-reachable via Apache `:443` ProxyPass
collapsing all peers to 127.0.0.1; I recommend **A**, or A+B; I cannot close it
unilaterally without locking the investor out mid-session); (b) **SOA MNAME placeholder**
(`a.misconfigured.dns.server.invalid.` — provider panel only; my granted DNS scope is
A/AAAA/CNAME/TXT/MX and `pdns-api.py` exposes no SOA command).

**Next (queued)**: (1) the two observations above (Go CLI row env-aware; system-status
unread count), pick the smaller; (2) watch the 08:06–08:50 CEST identity shifts and
`mailboxes/*` for Dispatcher assignments; (3) A/B/C + SOA stay with the investor/operator.
Commit + push after this write.
EOF
echo "appended"; php /data/repo/tests/test_changelog_api.php 2>&1 | tail -3; php /data/repo/tests/test_stats_dbpath.php 2>&1 | tail -2
appended
  PASS: prod/db.php passes php -l

=== Results: 78 passed, 0 failed ===

=== Results: 16 passed, 0 failed ===
> build · mimo-v2.6-flash-free
Everything green. Committing and pushing:
$ cd /data/repo
git add -A
git status --short
git commit -m "fix: stats.php resolves messages.db from __DIR__ instead of the hardcoded dev path (0.4.20) — prod /stats showed dev's counts; 16-assertion hermetic test (mutation-verified), reviewer-approved prod promote"
git push origin main 2>&1 | tail -3
git log --oneline -2
M  CHANGELOG.md
M  agent-logs/PROGRESS.md
M  app/src/php/stats.php
A  tests/test_stats_dbpath.php
[main e456168] fix: stats.php resolves messages.db from __DIR__ instead of the hardcoded dev path (0.4.20) — prod /stats showed dev's counts; 16-assertion hermetic test (mutation-verified), reviewer-approved prod promote
 4 files changed, 250 insertions(+), 1 deletion(-)
 create mode 100644 tests/test_stats_dbpath.php
To git://git.gladex.de/gladex.git
   82d7396..e456168  main -> main
e456168 fix: stats.php resolves messages.db from __DIR__ instead of the hardcoded dev path (0.4.20) — prod /stats showed dev's counts; 16-assertion hermetic test (mutation-verified), reviewer-approved prod promote
82d7396 fix: PHP SQLite opens get the Go busy_timeout contract via one helper gladex_pdo (0.4.19) — 6 sites, 19-assertion contention test (mutation-verified), reviewer-approved prod promote; stats.php hardcoded dev-DB path queued
> build · mimo-v2.6-flash-free
$ sqlite3 /opt/startup/dev/data/messages.db "SELECT COUNT(*) FROM messages WHERE direction='investor_to_agent' AND read=0;"; grep -c '^## ' /data/repo/INBOX.md; grep '^## ' /data/repo/INBOX.md | grep -vc HANDLED || echo "0 unhandled"
0
65
0
0 unhandled
> build · mimo-v2.6-flash-free
Run complete. Summary:

## STEP 0 — investor inbox (checked first, re-checked last)
- **0 unread** `investor_to_agent` rows in `/opt/startup/dev/data/messages.db` (and the prod mirror) — all 24 read
- **0 unhandled** entries in `/data/repo/INBOX.md` (65/65 marked `~~HANDLED~~`) → no reply rows owed; noted in PROGRESS.md
- Still-open asks re-surfaced under NEEDS-INVESTOR, not re-spammed: **msg #57 A/B/C public-gating**, **SOA MNAME placeholder**

## This run's step (0.4.19 → 0.4.20): the queued `stats.php` bug
**Bug**: `stats.php` hardcoded `/opt/startup/dev/data/messages.db`, so **prod `/stats` reported dev's message counts** — and the DBs had already diverged (dev 57 rows vs prod 22), so prod visibly claimed 57 messages while its own DB held 22.

- **Fix** (1 line + comment): `$dbPath = __DIR__ . '/../../data/messages.db';` — same idiom as `app.php`/`init_db.php`; `../../data` resolves to each env's own data dir
- **New test** `tests/test_stats_dbpath.php` (16 assertions, ~0.7 s): static layer (no copy may carry the hardcoded path) + behavioural layer (renders `stats.php` in a synthetic env tree seeded with known 7/3/4 rows and asserts the page prints **7/3/4**, plus a non-vacuity guard fixture 7 ≠ dev 57)
- **Mutation-verified**: restoring the hardcoded path → test printed dev's 57/24/33 instead of 7/3/4 → **10 FAIL, exit 1**; restored byte-identically (md5 `decf1849…`)
- **Deploy, reviewer-gated**: repo → dev (verified 57/24/33) → reviewer **VERDICT: APPROVE** → prod → I independently re-verified everything after the reviewer executed the promote

## Verification
- prod `/stats` now **22/3/19** (its own counts), dev **57/24/33** — direct before/after proof
- repo == dev == prod md5 `decf1849…`; both units active; 14 routes (7 × dev+prod) all **200**
- **Suites: 249 passed, 0 failed** (PHP 223 + shell 26); `healthcheck both` → HEALTHY/HEALTHY
- Committed `e456168` and **pushed** to `git://git.gladex.de/gladex.git`

## Housekeeping
- Mailboxes: 0 pending Dispatcher assignments; 6 identity timers armed; watchdog firing
- **No spend**: BUDGET 2026-09 unchanged 5.00 / 1.50 / 3.50, model spend 0.00 (`*-free` only), no DNS write, no API key touched
- Queued for next run: two same-family observations — `stats.php`'s "Go CLI" row reads only the dev binary; `tools/system-status` counts unread messages from the dev DB only
exit=0

Select another run log from the list above. Only files matching run-YYYYMMDD-HHMMSS-N.log are readable.

[Fri Sep 25 01:07:30 2026] 127.0.0.1:42770 Accepted
[Fri Sep 25 01:07:30 2026] 127.0.0.1:42770 Closing
[Fri Sep 25 01:07:30 2026] 127.0.0.1:42780 Accepted
[Fri Sep 25 01:07:30 2026] 127.0.0.1:42780 Closing
[Fri Sep 25 01:09:05 2026] 127.0.0.1:43758 Accepted
[Fri Sep 25 01:09:05 2026] 127.0.0.1:43758 Closing
[Fri Sep 25 01:29:04 2026] 127.0.0.1:52792 Accepted
[Fri Sep 25 01:29:04 2026] 127.0.0.1:52792 Closing
[Fri Sep 25 01:29:05 2026] 127.0.0.1:33658 Accepted
[Fri Sep 25 01:29:05 2026] 127.0.0.1:33658 Closing
[Fri Sep 25 01:29:05 2026] 127.0.0.1:33674 Accepted
[Fri Sep 25 01:29:05 2026] 127.0.0.1:33674 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49482 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49482 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49488 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49488 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49494 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49494 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49504 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49504 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49508 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49508 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49522 Accepted
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49522 Closing
[Fri Sep 25 01:34:56 2026] 127.0.0.1:49536 Accepted
[Fri Sep 25 01:34:57 2026] 127.0.0.1:49536 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60400 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60400 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60408 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60408 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60422 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60422 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60428 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60428 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60442 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60442 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60446 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60446 Closing
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60462 Accepted
[Fri Sep 25 01:37:40 2026] 127.0.0.1:60462 Closing
[Fri Sep 25 01:42:50 2026] 127.0.0.1:37798 Accepted
[Fri Sep 25 01:42:50 2026] 127.0.0.1:37798 Closing
[Fri Sep 25 01:42:50 2026] 127.0.0.1:37802 Accepted
[Fri Sep 25 01:42:50 2026] 127.0.0.1:37802 Closing
[Fri Sep 25 01:42:51 2026] 127.0.0.1:37814 Accepted
[Fri Sep 25 01:42:51 2026] 127.0.0.1:37814 Closing
[Fri Sep 25 01:42:51 2026] 127.0.0.1:37818 Accepted
[Fri Sep 25 01:42:51 2026] 127.0.0.1:37818 Closing
[Fri Sep 25 01:42:54 2026] 127.0.0.1:37824 Accepted
[Fri Sep 25 01:42:54 2026] 127.0.0.1:37824 Closing
[Fri Sep 25 01:42:55 2026] 127.0.0.1:44098 Accepted
[Fri Sep 25 01:42:55 2026] 127.0.0.1:44098 Closing
[Fri Sep 25 01:42:55 2026] 127.0.0.1:44106 Accepted
[Fri Sep 25 01:42:55 2026] 127.0.0.1:44106 Closing
[Fri Sep 25 01:43:00 2026] 127.0.0.1:44120 Accepted
[Fri Sep 25 01:43:00 2026] 127.0.0.1:44120 Closing
[Fri Sep 25 01:43:00 2026] 127.0.0.1:44134 Accepted
[Fri Sep 25 01:43:00 2026] 127.0.0.1:44134 Closing
[Fri Sep 25 01:43:00 2026] 127.0.0.1:44150 Accepted