2 Commits
Author SHA1 Message Date
thepraandClaude Opus 5.5 c8a305ae92 M8: every server we touch is described, located and snapshotted weekly
- Touches: the ledger marks a server as touched when it sends us a verified activity, when
  we exchange activities with it, or when we read its actors, keys, objects or WebFinger.
  It upserts RemoteInstance.Seen, FirstSeenAt and LastSeenAt at most hourly per server, and
  queues one DescribeInstance a week with the same dedupe key ObjectRecords uses. Suspended
  servers and pages behind link previews are never described. Migration _010 marks the
  servers already known as touched, with their dates.
- InstanceDescriber.Describe(host, crawled, allowed) reads:
  - NodeInfo 2.2/2.1/2.0, now with its published user counts, posts, comments,
    description, languages and schema version;
  - for software with a Mastodon API, /api/v2/instance falling back to v1: title,
    languages, registration mode, character limit, API version, source URL.

  It never keeps a contact as a field; the raw document is kept for the admin only. It
  locates the server from the address our connection reached (DB-IP Lite city and ASN, the
  CDN named when fronted) and writes a RemoteInstanceSnapshot per ISO week, unreachable
  weeks included. A crawled server is upserted as crawled only on insert, so it never
  downgrades a touched one, and robots.txt can deny any path.
- PublicGeo.Project is the only public form of a location: a CDN-fronted server shows its
  CDN only, a server reporting at least ten users shows its city, coordinates and network,
  any other only its country.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
2026-10-03 12:07:57 +02:00
thepraandClaude Opus 5.5 2cfea7b60c T7: the federation surface over HTTP
Tests through the real routes of PeasantsController, WellKnownController and
UsersController, on the whole server under test (34 new tests):

- FederationGetTests: the actor document (activity+json, SPKI key at
  #main-key owned by the actor, sharedInbox, published = PublishedOn's day,
  no creation date, no root); ld+json; browsers sent to /@name; Vary: Accept;
  /users 301; the outbox's totalItems and ?page=true&max_id paging across
  the 20-item boundary, boosts as Announces, and no followers-only, direct,
  located, federated-copy or deleted post; /groupies and /stalking naming
  nobody; /trophies (public pins, newest first) and /tattoos; /scribbles
  (public and unlisted 200, browsers redirected, followers-only, direct and
  located 404, deleted 410 Tombstone); a circle post only for a signed member
  or its instance actor; a circle's /groupies, /flock and /wardens only for
  members; /grunts create- and announce- ids; /parrot-licences 200, revoked
  410, wrong author 404; secure mode's 401 for every unsigned GET but the
  instance actor's.
- WellKnownTests: WebFinger by acct:, @-prefixed, bare, upper-case and actor
  URI; other domains, unknown names, a root's login name and no resource;
  the instance actor, a community, a circle (answered: current behaviour);
  NodeInfo links, 2.0 and 2.1 naming no root, unknown versions 404; usage
  counting only public, unlisted, non-boost local posts (Exclusive).
- InboxRouteTests: all three inboxes accept a signed delivery and refuse
  junk (400), unsigned (401), a bad Digest, a two-hour-old Date, a signature
  for another host and a swapped body (401); an unknown persona's /mouth is
  404; ld+json with the ActivityStreams profile is accepted; a signer whose
  actor answers 503 gets 503 with Retry-After; the 301st unsigned POST from
  one address is 429 while a signed server from it is not.
- PersonaSeparationHttpTests: with a sibling persona and its community on
  the same login, every GET under /api (filled with the persona's ids) plus
  search, lookup and relationships, and a crawl of everything federation
  publishes about the persona (actor, outbox pages, collections, scribbles,
  grunts, WebFinger, NodeInfo, /@ pages), never name the sibling, its
  community or the login.

Fixed:
- A circle's /groupies told anyone how many followers (members) it has,
  while its /flock and /wardens were already for members only; it now
  answers 404 to anyone but a signed member or a member's instance actor.
- WebFinger answered 404 to a bare user@domain or @user@domain resource,
  which Mastodon, GoToSocial and Pleroma all accept; it now treats them as
  acct: (noted in docs/INTEROP.md).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
2026-10-03 11:53:23 +02:00