fix(email): stop flooding the log when email polling fails

[tryingET]

Problem

EmailReceiver polls /email/poll every 10 seconds and logs the entire axios error on every failure: request, socket and agent included. On one desktop install that was 83,334 log entries in a week, which pushed the daily log to 366 MB:

• 41,599 × HTTP 410 { error: 'retired', reason: 'pro_required', service: 'mail.agnt.gg' }
• 41,581 × HTTP 401 { error: 'Authentication required', reason: 'missing' } (before sign-in)

Change

• A failure logs one short line (status plus the server's reason), and only when the reason changes. A later success logs one "recovered" line.
• 410 means the mail service will not serve this client, so polling stops until the next start instead of retrying every 10 seconds forever.
• Behaviour on success is unchanged.

Please check

• Is stopping on 410 right for your pro_required rollout? If a user upgrades while the app is running, email triggers resume only after a restart. An alternative is a long back-off (for example hourly) instead of stopping.

Verification

• New EmailReceiver.poll.test.js: 3 tests (410 stops with one short line, repeated 401 logs once, reason change and recovery). All 3 failed before the change and pass after.
• Full backend: 6,340 passed. The 2 failing files also fail on main: sessionTokenCache.test.js (date-expired fixture, fixed in fix(storage): prevent test boot of real user database #145) and browserActDriver.live.test.js (needs a real browser).