hyperhive/nix/host-modules/hive-matrix.nix

{
  pkgs,
  lib,
  config,
  ...
}:
let
  cfg = config.services.hyperhive.swarm.matrix;
  networkCfg = config.services.hyperhive.network;
  swarmDomain = config.services.hyperhive.swarm.domain;
  # Falls back to the SWARM domain: a swarm runs one homeserver, so its
  # identifier belongs to the swarm rather than to whichever hive happens
  # to host it — otherwise moving the container between hives would look
  # like a different homeserver.
  #
  # ⚠️ Changing this default is a BREAKING change in a way that moving
  # `gatewayHost` was not: `serverName` is baked irrevocably into every
  # user and room id, so a deployment that rebuilds onto a new one is a
  # *different homeserver*, not a renamed one. Existing hives pin the old
  # value explicitly (see the option's description); the default is what
  # a fresh swarm gets.
  #
  # Total on a null swarm domain, deliberately: the required-domain
  # assertion in hive-network.nix is what should fire, not a coercion
  # error from an unrelated option interpolating null.
  effectiveServerName =
    if cfg.serverName != null then
      cfg.serverName
    else if swarmDomain != null then
      swarmDomain
    else
      "invalid";

  # fluffychat-web build fixes: nixpkgs's `flutter341.buildFlutterApplication`
  # skips the dart web-worker compile + the emscripten native_imaging
  # build. Two derivations below cover both. Full rationale (why
  # passthru.pubspecLock.dependencySources, why `dontConfigure`, why
  # `make -C js`, why build-CWD-relative dart path): docs/matrix.md::
  # fluffychat-web build fixes.

  fluffychat-web-imaging = pkgs.stdenv.mkDerivation {
    pname = "fluffychat-web-imaging";
    version = pkgs.fluffychat-web.passthru.pubspecLock.dependencyVersions.native_imaging;
    src = pkgs.fluffychat-web.passthru.pubspecLock.dependencySources.native_imaging;

    nativeBuildInputs = with pkgs; [
      emscripten
      cmake
      gnumake
      jq
    ];

    # cmake runs inside js/Makefile via `emcmake cmake`; the default
    # configurePhase would invoke cmake at the package root (no
    # CMakeLists) and fail.
    dontConfigure = true;

    buildPhase = ''
      runHook preBuild
      # emscripten on-demand sysroot build needs writable HOME + cache.
      export HOME=$TMPDIR
      export EM_CACHE=$TMPDIR/.emscriptencache
      mkdir -p $EM_CACHE
      # `make -C js` keeps pwd at source root for the installPhase.
      make -C js Imaging.js Imaging.wasm
      runHook postBuild
    '';

    installPhase = ''
      runHook preInstall
      mkdir -p $out
      install -m 644 js/Imaging.js  $out/Imaging.js
      install -m 644 js/Imaging.wasm $out/Imaging.wasm
      runHook postInstall
    '';

    meta = with pkgs.lib; {
      description = "Imaging.js + Imaging.wasm built from the native_imaging dart package for fluffychat-web";
      homepage = "https://pub.dev/packages/native_imaging";
      license = licenses.agpl3Plus;
    };
  };

  fluffychat-web-fixed = pkgs.fluffychat-web.overrideAttrs (old: {
    # dart from the flutter341 closure (already pulled, no incremental
    # cost) to compile the web-worker entry point.
    nativeBuildInputs = (old.nativeBuildInputs or [ ]) ++ [ pkgs.flutter341.dart ];

    postInstall = (old.postInstall or "") + ''
      # `web/...` is BUILD-CWD-relative (not `$src/...`) so dart's
      # package_config walk-up hits buildFlutterApplication's
      # pub-get output `.dart_tool/`.
      ${pkgs.flutter341.dart}/bin/dart compile js \
        -o $out/native_executor.js \
        web/native_executor.dart

      install -m 644 ${fluffychat-web-imaging}/Imaging.js  $out/Imaging.js
      install -m 644 ${fluffychat-web-imaging}/Imaging.wasm $out/Imaging.wasm
    '';
  });
in
{
  # Private matrix-tuwunel homeserver wrapped in a nixos-container,
  # optional fluffychat-web client at matrix.<hive>/. Container shape,
  # serverName vs gatewayHost split, provisioning flow (registration
  # token + LoadCredential), assertion rationale, initial rollout
  # settings: docs/matrix.md. Vhost map + discovery flow + tuning
  # knobs: docs/gateway.md.

  # Matrix moved under `swarm` when the swarm-global services were
  # consolidated. One rename for the namespace: the subtree comes with it,
  # so existing hives keep evaluating and get one warning naming both paths.
  imports = [
    (lib.mkRenamedOptionModule
      [ "services" "hyperhive" "matrix" ]
      [ "services" "hyperhive" "swarm" "matrix" ]
    )
  ];

  options.services.hyperhive.swarm.matrix = {
    enable = lib.mkOption {
      type = lib.types.bool;
      default = false;
      description = ''
        Run hive-matrix — a private matrix-tuwunel homeserver (in a
        nixos-container) for hyperhive agents.

        Matrix is a swarm-wide service — one homeserver, not one per
        hive — so `services.hyperhive.swarm.enableRequiredServices`
        turns this on as part of saying the swarm's shared services live
        on this host. Set it here directly to run the homeserver
        somewhere other than the host that holds the rest of them.
      '';
    };

    package = lib.mkOption {
      type = lib.types.package;
      default = pkgs.matrix-tuwunel;
      defaultText = lib.literalExpression "pkgs.matrix-tuwunel";
      description = ''
        matrix-tuwunel package to run inside the container. Defaults
        to nixpkgs's `pkgs.matrix-tuwunel`. Override to pin a
        specific upstream if you need an unreleased feature.
      '';
    };

    serverName = lib.mkOption {
      type = lib.types.nullOr lib.types.str;
      default = null;
      example = "chat.example.org";
      description = ''
        Matrix `server_name` — the host part of every user ID
        (`@argus:<server_name>`) and room ID minted on this
        homeserver. CRITICAL: must be stable from day one because
        it's embedded irrevocably in the identifiers.

        Defaults to `services.hyperhive.swarm.domain` (the bare swarm
        domain), because **a swarm runs one homeserver** — tying its
        identity to a single hive's domain would make relocating the
        container between hives look like a different homeserver.
        Combined with the `.well-known/matrix/{client,server}` routes
        the gateway serves at that domain, clients auto-discover the
        actual matrix endpoint without needing a subdomain. Override
        here only if you need a different server_name shape (e.g.
        `chat.example.org` for a bespoke hostname).

        **Breaking change, and the one on this page that cannot be
        undone by rebuilding.** This default has now moved twice — from
        `matrix.''${services.hyperhive.domain}`, then to the bare hive
        domain, and now to the swarm domain. Every existing homeserver
        must pin whichever value it already minted ids under, e.g.

        ```nix
        services.hyperhive.swarm.matrix.serverName =
          config.services.hyperhive.domain;   # or "matrix.''${…domain}"
        ```

        before rebuilding. Adopting a new `server_name` does not rename
        the old users and rooms — it strands them, because their ids
        still name a homeserver that no longer answers.
      '';
    };

    httpPort = lib.mkOption {
      type = lib.types.port;
      default = 8008;
      description = ''
        TCP port tuwunel serves the matrix client-server API on.
        Default 8008 is the matrix-spec well-known port. Sits
        outside hyperhive's claimed ranges (dashboard 7000, every
        agent in 8100..8999 via FNV-1a hash). Federation listens on
        `federationPort` separately.
      '';
    };

    apiUrl = lib.mkOption {
      type = lib.types.nullOr lib.types.str;
      default = if cfg.enable then "http://127.0.0.1:${toString cfg.httpPort}" else null;
      defaultText = lib.literalExpression ''
        if services.hyperhive.swarm.matrix.enable
        then "http://127.0.0.1:''${toString services.hyperhive.swarm.matrix.httpPort}"
        else null
      '';
      example = "https://matrix.example.com";
      description = ''
        Client-server API base URL **hive-c0re itself** uses to
        provision matrix (register agent users, create the hive space
        and chat room, invite members). Distinct from the agent-facing
        `hyperhive.matrix.url`, which is the gateway vhost handed to
        each agent's `hive-matrix-daemon`.

        Defaults to the loopback listener **only when this module is the
        thing running tuwunel** — in that case the address is not a
        guess, it is where this module just put the container. Set it
        explicitly (with `enable = false`) when the homeserver runs on
        another machine; "everything on one host" is a special case of
        the full deployment, not the assumption.

        `null` means hive-c0re has no homeserver to provision against
        and matrix provisioning no-ops. There is deliberately no
        fallback compiled into the daemon: an address baked into the
        binary is one that builds fine and then talks to the wrong
        machine.
      '';
    };

    gatewayHost = lib.mkOption {
      type = lib.types.nullOr lib.types.str;
      # `chat.` under the SWARM domain — both halves change: a swarm runs
      # one homeserver, and the label follows the service rather than the
      # protocol.
      #
      # Total on a null swarm domain so the required-domain assertion in
      # hive-network.nix is the thing that fires; see the comment there.
      default = if swarmDomain == null then "chat.invalid" else "chat.${swarmDomain}";
      defaultText = lib.literalExpression ''"chat.''${services.hyperhive.swarm.domain}"'';
      example = "matrix.example.com";
      description = ''
        Public hostname for the matrix homeserver behind the gateway.
        Defaults to `chat.''${services.hyperhive.swarm.domain}` — the
        swarm's domain, because a swarm runs **one** homeserver. Set to
        `null` to skip the gateway vhost (tuwunel stays direct on
        `httpPort`). See `docs/gateway.md` for the vhost map + matrix
        discovery flow, and the federation port-8448 caveat at the
        bottom of that doc.

        ⚠️ **`gatewayHost` and `serverName` are different things, and
        they carry very different costs.** `gatewayHost` is the API
        listener hostname (where nginx proxies `/_matrix/*`) and is
        free to change: it is a routing detail clients rediscover
        through `.well-known`. `serverName` is the matrix-identifier
        domain embedded **irrevocably** in every user and room id —
        adopting a new one is a different homeserver, not a rename.
        Both defaults now sit under the swarm domain, but only this
        one is safe to move on a running deployment.

        A deployment that was running before this moved keeps its
        current name by pinning
        `matrix.''${services.hyperhive.domain}` here — exactly what the
        old default rendered.
      '';
    };

    openFirewall = lib.mkOption {
      type = lib.types.bool;
      default = false;
      example = true;
      description = ''
        Open `httpPort` in the host firewall. Off by default
        (secure-by-default): the host reaches the homeserver on
        loopback, and agent containers reach it at `matrix.<domain>`
        via the gateway — so the firewall open only matters for
        access from outside the host. Flip to `true` when announcing
        the homeserver to other hives or when an external matrix
        client needs to reach the client-server API directly.

        **Breaking change**: this used to default to `true`. If you
        relied on the old default for external reach, add
        `services.hyperhive.swarm.matrix.openFirewall = true;` to your host
        config before rebuilding.

        Note: federation (the matrix-spec well-known port 8448) is
        intentionally not opened here. tuwunel serves the federation
        API on the same `httpPort` as the client-server API by
        default; reaching it on 8448 requires either binding tuwunel
        to that port explicitly OR a reverse-proxy + `.well-known/
        matrix/server` delegation, neither of which lives in this
        module. Add that proxy config alongside whatever serves your
        dashboard or forge on 443.
      '';
    };

    trustedServers = lib.mkOption {
      type = lib.types.listOf lib.types.str;
      default = [ ];
      example = [ "matrix.org" ];
      description = ''
        List of trusted matrix servers (homeservers whose signing
        keys this server will fetch identity-server-style). Empty
        by default — federation is enabled at the protocol level
        but no peer is trusted until listed here, so the homeserver
        is effectively closed until the operator declares hive
        peers explicitly.
      '';
    };

    maxRequestSize = lib.mkOption {
      type = lib.types.ints.positive;
      default = 20000000;
      description = ''
        Maximum size in bytes of a single matrix client request body.
        Default 20 MB matches the matrix-spec recommendation for
        media uploads + the upstream tuwunel default.
      '';
    };

    registrationTokenFile = lib.mkOption {
      type = lib.types.path;
      default = "/var/lib/hyperhive/matrix-register-token";
      description = ''
        Host path to a file containing the matrix registration token
        tuwunel reads to authorise new-account creation. The token is
        generated automatically by `hive-c0re` on first boot (32-byte
        random hex, mode 0600) and is bind-mounted read-only into the
        tuwunel container at the same path. Agents never see this
        token — hive-c0re uses it to provision per-agent accounts
        and the agent only receives the resulting `access_token`.
        Override only when integrating with externally-managed
        registration tokens.
      '';
    };

    allowEncryption = lib.mkOption {
      type = lib.types.bool;
      default = false;
      description = ''
        Server-side switch for matrix end-to-end encryption — sets
        tuwunel's `allow_encryption`. Off by default: on the hive-internal
        homeserver the operator already controls the transport, so server
        E2EE adds key-management overhead (cross-signing, device
        verification, undecryptable-message recovery) without a clear
        threat-model win for the common single-hive case. Turn on when
        agents join encrypted rooms on external / federated homeservers,
        or when the operator wants message contents opaque to the
        homeserver admin. Independent of the agent matrix client, which
        always supports decryption so it can read encrypted rooms it is
        invited to regardless of this flag; this option only governs
        whether THIS homeserver permits room encryption.
      '';
    };

    gui = {
      enable = lib.mkOption {
        type = lib.types.bool;
        default = cfg.enable;
        defaultText = lib.literalExpression "config.services.hyperhive.swarm.matrix.enable";
        description = ''
          Serve a matrix web client at `matrix.''${services.hyperhive.domain}/`.
          Requires `matrix.gatewayHost != null` (default `matrix.<hive>`
          when hive-domain set); the gateway itself always runs. When
          off, the dashboard's `M4TR1X →` tab is hidden. See
          `docs/gateway.md` for the discovery flow that lets clients
          auto-find the sub-domain.
        '';
      };

      package = lib.mkOption {
        type = lib.types.package;
        default = fluffychat-web-fixed;
        defaultText = lib.literalMD ''
          `pkgs.fluffychat-web` with a `postInstall` patch that adds
          the three files `flutter341.buildFlutterApplication` skips.
        '';
        description = ''
          Static web client dist served at `matrix.<hive>/`. Override
          to swap fluffychat for hydrogen-web, cinny, element-web, or
          an out-of-tree dist — any replacement is mounted at the
          sub-domain root with the upstream-default `<base href "/">`,
          no sub-path gymnastics needed.
        '';
      };
    };
  };

  config = lib.mkIf cfg.enable {
    # `serverName` is irrevocably embedded in user/room IDs; it derives
    # from `services.hyperhive.domain` (required, asserted in
    # hive-network.nix) when not set explicitly, so no separate
    # domain/serverName assertion is needed here. gatewayHost may not be
    # "" (same footgun as forge.domain — nginx rejects an empty
    # server_name). docs/matrix.md::Assertion rationale.
    assertions = [
      {
        assertion = cfg.gatewayHost == null || cfg.gatewayHost != "";
        message = ''
          services.hyperhive.swarm.matrix.gatewayHost = "" is rejected. The
          rendered URLs would be invalid (nginx wildcard catch-all for
          an empty server_name, /etc/hosts rejects empty entries).
          Use `null` to disable the gateway vhost entirely (tuwunel
          stays direct on httpPort), or set a non-empty hostname like
          "matrix.example.com" or "homeserver.internal".
        '';
      }
    ];

    # Activation-time token generation — without this the bind-mount
    # would hand tuwunel an empty file on first boot and break every
    # registration until restart. Idempotent;
    # docs/matrix.md::Provisioning flow.
    system.activationScripts.hive-matrix-register-token = lib.stringAfter [ "var" ] ''
      tokenFile=${lib.escapeShellArg (toString cfg.registrationTokenFile)}
      if [ ! -s "$tokenFile" ]; then
        mkdir -p "$(dirname "$tokenFile")"
        head -c 32 /dev/urandom | od -An -tx1 | tr -d ' \n' > "$tokenFile"
        echo >> "$tokenFile"
        echo "hive-matrix: generated registration token at $tokenFile"
      fi
      # Re-apply 0600 (normalises any pre-LoadCredential carry-over).
      chmod 0600 "$tokenFile"
    '';

    containers.hive-matrix = {
      autoStart = true;
      ephemeral = false;
      # Shared host netns — agents reach tuwunel at localhost:<port>.
      privateNetwork = false;
      # Read-only bind of the host-managed registration token; tuwunel
      # reads it via systemd LoadCredential below (not directly).
      bindMounts.${cfg.registrationTokenFile} = {
        hostPath = cfg.registrationTokenFile;
        isReadOnly = true;
      };
      config =
        { ... }:
        {
          system.stateVersion = "26.05";

          # Shared host netns: this container's own firewall.service
          # would rewrite the HOST ruleset (flush nixos-fw, drop the
          # host's nixos-nat-* chains) at every boot — killing the
          # bridge DHCP/DNS holes and agent NAT. The host firewall owns
          # all filtering; never run one in here.
          networking.firewall.enable = false;

          # ⚠️ This container trusts no swarm-internal CA. It used to take
          # per-hive root CAs, hand-pinned as nix paths, so tuwunel could
          # validate *federation* TLS from a self-signed peer hive; that
          # field is gone, and the swarm root that replaces it cannot be
          # substituted here. `security.pki.certificateFiles` is read when
          # the system is BUILT, and the swarm root is a runtime file
          # (`swarm.ca.stateDir`) precisely because its key must never
          # reach the store — so there is nothing build-time to name.
          #
          # Giving the container the swarm root therefore needs a runtime
          # mechanism (bind-mount + a bundle assembled at start), which is
          # a different shape than this line and is tracked as its own
          # issue. Federation with a peer whose cert chains to the swarm
          # root does not validate until then.

          # tuwunel hard-fails to boot if `/etc/resolv.conf` has no
          # `nameserver` line (`Failed to configure DNS resolver ... no
          # nameservers found in config` → exit 1). This declarative
          # nixos-container comes up with an EMPTY resolv.conf even with
          # `networking.nameservers` set: the nixos-container default
          # `useHostResolvConf = true` puts in-container resolvconf in
          # host-tracking mode (ignores `networking.nameservers`, and never
          # gets the host file across the shared-netns boundary), so it
          # regenerates an empty file and tuwunel dies at boot.
          #
          # Trusting resolvconf to honour `networking.nameservers` doesn't
          # work either — that's a RUNTIME resolvconf behaviour, not
          # verifiable at eval time, and it still comes up empty in
          # practice. So take resolvconf out of the loop entirely and
          # write a STATIC `/etc/resolv.conf` from `bridgeIp` that nothing
          # regenerates. Eval-proven: the generated
          # `environment.etc."resolv.conf".text` is `nameserver <bridgeIp>`.
          # This container always shares the host netns
          # (`privateNetwork = false`), so it reaches `bridgeIp` regardless
          # of agent-container isolation. See `docs/network.md`.
          networking = {
            # resolvconf is taken out of the loop entirely; the static
            # `environment.etc."resolv.conf"` below is the sole source of
            # the resolver file (no `nameservers` — nothing would read it).
            useHostResolvConf = lib.mkForce false;
            resolvconf.enable = lib.mkForce false;
          };

          # resolvconf is disabled above, so write the static resolver file
          # explicitly — NixOS won't synthesise one from `nameservers` once
          # resolvconf is off, and this is the file tuwunel parses at boot.
          environment.etc."resolv.conf".text = ''
            nameserver ${networkCfg.bridgeIp}
            options edns0
          '';

          services.matrix-tuwunel = {
            enable = true;
            package = cfg.package;
            settings.global = {
              server_name = effectiveServerName;
              # `address` + `port` are upstream `listOf` — wrap singles.
              address = [ "0.0.0.0" ];
              port = [ cfg.httpPort ];
              max_request_size = cfg.maxRequestSize;
              # Federation enabled at the protocol level; empty
              # trustedServers keeps it effectively closed.
              allow_federation = true;
              trusted_servers = cfg.trustedServers;
              # Token-gated registration. The absent
              # `yes_i_am_very_very_sure_…_open_registration_…` flag
              # keeps the server closed to anyone without the token.
              allow_registration = true;
              # LoadCredential below copies the host file into a
              # 0400 dynamic-user-owned path; tuwunel reads from there.
              registration_token_file = "/run/credentials/tuwunel.service/registration_token";
              # Server-side E2EE is opt-in (default off); the agent matrix
              # client always supports decryption regardless.
              allow_encryption = cfg.allowEncryption;
              # Tuwunel's default suffix is " 💕" — suppress it so agent
              # display names are clean (just the agent name, no emoji).
              new_user_displayname_suffix = "";
            };
          };
          # Keeps DynamicUser=true + PrivateUsers=true intact — no
          # host-side chown :tuwunel / GID-pin gymnastics needed.
          # See `man systemd.exec` → LoadCredential.
          systemd.services.tuwunel.serviceConfig.LoadCredential = [
            "registration_token:${toString cfg.registrationTokenFile}"
          ];
          environment.systemPackages = [ cfg.package ];
        };
    };

    networking.firewall = lib.mkIf cfg.openFirewall {
      allowedTCPPorts = [
        cfg.httpPort
      ];
    };

    # The matrix container's resolver is the dnsmasq that runs in the
    # gateway container (bound at `bridgeIp`). Order the matrix
    # container start after the gateway container so the resolver is up
    # before tuwunel's first federation lookups. tuwunel boots fine
    # without this — it configures the resolver from `/etc/resolv.conf`
    # at startup and only queries on-demand (the boot failure this
    # module guards against is an *empty* resolv.conf, a parse error,
    # not a connectivity one) — so this is robustness, not a boot
    # requirement. Soft `after` ordering (not `requires`) keeps the
    # matrix container's lifecycle decoupled from the gateway's. The
    # gateway always runs alongside hyperhive, so the gateway container
    # unit always exists here. (Declarative `containers.<n>` →
    # `container@<n>.service` — the nspawn template NixOS generates.)
    systemd.services."container@hive-matrix".after = [
      "container@hive-gateway.service"
    ];
  };
}