2 Commits

Include renames

Author	SHA1	Message	Date
Bendik LynghaugandClaude Opus 5	44c7ee6870	Render an Apple configuration profile from the server's PACC document ... iOS is the one client none of the mail autoconfiguration reaches: Apple Mail reads neither Mozilla autoconfig nor Autodiscover for IMAP, so adding an account means the device probing for ports - and a probe that meets a filtered port waits out a TCP timeout and reports the account as unverifiable, which the person reads as a wrong password. A profile states the hosts and the ports instead, and carries mail, calendars and contacts in one install. It invents none of them: site.yaml names a PACC document - the standard defined to be fetched from the mail domain itself - and every host, port and collection path comes from there, so the server is described in one place and the profile follows it. The document holds no address, username or password. iOS asks for those at install, which means one profile serves everyone with an account and it can be fetched by anyone: it says no more than PACC already says in public. Identifiers are derived from the site and the payload rather than drawn at random, because iOS replaces a profile whose ids it knows and stacks a duplicate when they differ. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015GeH3y8fwTaTudaD88P57o	2026-10-07 11:18:46 +02:00
Bendik Aagaard LynghaugandClaude Opus 5	214d4eb85c	Sessions outlive the deploy that interrupted them ... A MemoryStore meant every restart signed out every desk on the site being deployed. Bearable while a desk is three people, and not the kind of thing that gets better on its own. They live in NATS KV now: the JetStream already under the events, the projections, the content reload and the access questions. No new service, no new store, and the same bucket a second instance of a site would read if there ever is one. `create` draws a fresh id rather than writing over a session already at that key, the bucket's `max_age` collects what nobody comes back for, and a record's own expiry is what answers a load, whatever the bucket still holds. Proved the only way it can be: a store opened twice against one JetStream, which is what two runs of the binary are. Also fixed, while there: the integration test asserted a record's projected state was one that `aggregates::transition` never writes - it appends events, the desk handler updates the projection - so it failed whenever it was actually run. It now asserts what it says it does, that a mailed reply leaves the state alone. test.yml has been red since 0.3.47 while publish.yml stayed green, which is how v0.4.0 went out over it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-28 16:38:05 +02:00