Owner decisions (2026-10-04, recorded in docs/ROADMAP.md): production runs everything that is built, and nothing waits
on a person running a command.
- Geolocation updates itself. GeoUpdater, a hosted service, checks daily whether each DB-IP Lite database was built this
month. If not, it fetches this month's, or last month's early in the month. It installs a file only once it opens as
the right kind of database, then swaps it in atomically, and the locator reloads at once. Lookups now run under the
lock, so a reload can no longer dispose a reader mid-lookup. The systemd timer, its script and their setup.sh lines
are gone: the root step they needed never happened, and none is needed now. /stargazing names the database in use.
- The admin CLI runs after the app is built, with every service and nothing started.
- `create-root <login> [--admin]` takes the password on stdin; it is how the first login is made while sign-up is
closed.
- `smoke <persona>` keeps the root `deploy-smoke` and an undiscoverable persona, and gives the root a new password
on every run.
- The deploy signs in as @thepra. It runs the CLI, gets a token through the real OAuth flow (tools/smoke/oauth.sh,
moved out of the pasture's privapub_token, which now uses it), checks the signed-in API and that @thepra is
undiscoverable, then revokes the token. PRIVAPUB_SMOKE_TOKEN is gone.
- The deploy also fails when:
- NodeInfo and the instance API disagree about registrations;
- /stargazing does not say the crawler is on;
- the geolocation databases are missing or more than 40 days old.
- The crawler is on in production, seeded with ten large servers of different kinds. FEDERATION.md now describes it
and how to opt out.
- One registrations switch (Registrations:Mode, default Invitations; Open in tests and the pasture). It is read by
open sign-up (403 when closed), NodeInfo `openRegistrations`, and v1 and v2 of the instance API, so they can no longer
disagree. Before, NodeInfo said open and the instance API said closed. Group invitations always work, so
invites_enabled is true.
- A persona edit through /clientapi no longer resets what the Mastodon API set (discoverable, locked, quote policy…):
the theme is merged into the settings instead of replacing them.
650 tests pass. The deploy's smoke step was rehearsed against the pasture's PrivaPub.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
tools/ci/with-test-mongod.sh starts mongod (or podman's mongo:8) on a random localhost port
with a temporary data directory, runs the command, and removes both. build.yml and deploy.yml
test through it, so the ~85 integration tests stop being skipped in CI. MongoFixture refuses
production's port and data directory, and in CI anything but the wrapper's mongod; with
PRIVAPUB_TEST_REQUIRE_MONGOD=1 a missing mongod fails instead of skipping.
The deploy now passes the PRIVAPUB_SMOKE_TOKEN secret to the Mastodon smoke check, so its
signed-in half runs once the owner creates the persona and the secret.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
- PersonaSeparationTests: two personas of one login follow the same
account and post; nothing the API maps for one contains the other's id,
username or the root id.
- tools/smoke/mastodon-api.sh checks what a client meets first (instance
v1/v2, discovery, app registration, client credentials, an app token
refused by a user endpoint, the public timeline, revocation) and, given
a persona token, verify_credentials, home and notifications. deploy.yml
runs it after every deploy.
- OAuthPruner removes invalid tokens and authorizations older than two
weeks, every six hours.
- The consent page no longer sets form-action, which browsers apply to the
redirect back to the client after the form is posted.
CLAUDE.md gains the Mastodon API layout and invariants.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
- Sign-up, login, the invitation flows and password recovery allow ten
requests a minute per client address; the inboxes give each sending
origin (the keyId's) a bucket of 300 that refills at 300 a minute, and
answer 429 beyond it, which peers retry.
- deploy.yml dumps the PrivaPub database to /var/backups before it stops
the service (the last seven are kept), and after the swap checks that
Swagger answers 404 and that the shared inbox answers junk with 400 and
an unsigned activity with 401.
- ActivityPubClient and PostBoost, never used, are gone.
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
The tree had not compiled since its first commit: Group, DmGroup and
IGroupUsersService were referenced and never written, an IDE rename had
turned the user-settings DTO into the ViewAvatarServer enum, and the settings
were saved as an entity they no longer were.
Built now:
- Group (an ActivityPub Group actor with its own keys, members, invitation
code and optional password) and DmGroup (a conversation), with
/clientapi/group/{list,insert,update,join,leave,approve}.
- Posts and DMs: /clientapi/post/{list,insert,delete}, /clientapi/dm/{list,insert};
DM recipients are local usernames or user@host handles resolved by WebFinger.
- Invitation sign-up and login against the group's invitation code, checking
the password before any account is created.
- Federation: WebFinger, NodeInfo 2.0, actors at /peasants/{name} (Person,
Group, and an Application instance actor) with SPKI keys, draft-cavage
RSA-SHA256 HTTP signatures both ways, an inbox handling Follow (+Accept),
Undo, Create, Delete and Update, an outbox, notes at /posts/{id}, and a
persisted, retried, signed delivery queue. A post to a group is announced
by the group to its followers (FEP-1b12).
- The unused, broken typed ActivityPub models are replaced by a renderer;
NSign's HMAC setup, which could not federate, is gone.
Upgrade: net10.0, MongoDB.Entities 25.1 (instance DB API, Standard GUIDs),
Swashbuckle 10 / OpenApi 2, Serilog.AspNetCore 10, MailKit 4.18,
PasswordGenerator 3. The JWT keys are 64 bytes (IdentityModel 8 refuses
shorter for HS512). Production runs its own mongod on 127.0.0.1:27022, as
Sintopia's apps do, and deploys to privapub.thepra.dev from the build runner.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
thepraandClaude Opus 5.5