A server can take a Follow with 202 and drop it afterwards, as Pleroma does
while it cannot fetch our actor; the request then stayed pending for good.
Following again now sends an unanswered request once more, the same
activity, at most once an hour (a delivery's `again` key).
accounts/search takes following=true: only accounts the persona follows,
by the start of their name, display name or server, never resolved; a
client fills a list with it. "already take" becomes "already taken".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LsXgEaXee4GCU1hwYgPJXw
Mastodon answers 422 when two first contacts from one actor race to
create its account (ActiveRecord::RecordInvalid on the unique uri), and
409 while another worker holds its lock. A persona that followed two
Mastodon accounts at once had one Follow refused that way; the job died
on its first attempt and the persona waited on "requested" forever.
Both answers are now retried twice on the usual backoff before they
count as refusals.
Found by the town (a village of 23 accounts on seven servers).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LsXgEaXee4GCU1hwYgPJXw
DeliveryJobHandler records each attempt as an 'out' event: the activity, object type and
audience read from the body (ActivityShape, shared with the inbox side), the receiving
server, the status, the time the POST took, the wait since it was queued, the bytes, the
attempt number and the signer's kind (none for circles or private traffic). The outcome is
ok, deferred (429 or 503 with Retry-After, or an open breaker: host-unavailable), retry
(5xx, 408, timeout, network), or dead: another 4xx, private-address, not-deliverable,
signer-gone, or a retry at the last attempt. Until now none of this outlived the job's
seven-day TTL, and the last error was overwritten on success.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
- Dislike and its Undo are kept as downvotes (Lemmy, PieFed, Mbin, Friendica) and shown with favourites as votes.
- ChatMessage (Lemmy 0.19, Mbin, PieFed to those two) arrives as a direct message.
- A deleted object's id is remembered for 90 days, so a Create that arrives after its Delete cannot bring the post
back; the post's ObjectRecord goes with it.
- A Join of one of our objects is answered with Ignore, as FEP-8a8e asks of a server without RSVP.
- A 503 with Retry-After is waited out like a 429 instead of counting as a failure of the host (GoToSocial throttling,
Mastodon's temporary key failures).
- Status.privapub carries what a Mastodon Status cannot: object type, title, excerpt, cover, the author's source,
link, video, audio and event details, and up/down votes, with every media URL proxied.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
The single serial DeliveryWorker is replaced by Infrastructure/Jobs:
- Job rows are leased with one FindOneAndUpdate (oldest RunAt first, a
two-minute lease) and a reaper returns expired leases every 30 s;
- enqueueing wakes the workers, which otherwise poll every five seconds;
- delivery runs eight at a time with at most two per host, so a slow or
dead server holds two slots, not the queue;
- a failure waits n^4 + 15 + jitter seconds (Mastodon's curve) for up to
16 attempts; a 4xx other than 408/429 is final, a 429 honours
Retry-After;
- RemoteInstance is a per-host circuit breaker: ten consecutive failures
quarantine a host for an hour, doubling to a week, and its jobs wait
without spending attempts;
- a delivery is queued once per activity and inbox (unique DedupeKey), and
finished jobs expire after seven days (TTL on FinishedAt).
Migration _004 moves pending Delivery rows into jobs and marks them
abandoned, so a rollback to the old worker cannot send them twice.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
Infrastructure/Http adds the client the roadmap's S3 and S4 ask for:
- the connect callback resolves the name itself and refuses loopback,
private, link-local, CGNAT, documentation, multicast, ULA, NAT64, 6to4,
Teredo and IPv4-mapped/compatible forms, then connects to the vetted
address, so DNS rebinding cannot swap it afterwards;
- redirects are followed by hand, at most three, each one re-checked;
- bodies are capped at 1 MB after decompression, only JSON media types are
read, every request has a 15 s budget, and a refused URL is not asked
again for five minutes.
Actor and WebFinger fetches and inbox deliveries all use it. Test networks
can switch on Federation:AllowPrivateNetworks/AllowPlainHttp; startup
refuses both in Production.
PrivaPub.Tests (xUnit v3) starts with the address table and the fetcher's
limits against an in-process peer; build.yml and deploy.yml run it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
Services/Federation and Controllers/ServerToServer become
Federation/{Actors,Signing,Inbox,Outbox,Rendering,Controllers}, the first step
of the roadmap's layout. No type, route or behaviour changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
thepraandClaude Opus 5.5