|
| | Application (Config config, Listener &listener) |
| |
| bool | canPublish () const noexcept |
| | What the last Gate said (H-16); lags the host by a hop.
|
| |
| bool | isLeading () const noexcept |
| |
| protocol::Publish | publish (const std::uint16_t payloadId, const std::uint8_t *payload, const std::uint16_t payloadLength) |
| | Offers one payload of this application's own to its own group (H-4).
|
| |
| protocol::Publish | publish (const std::uint16_t groupMask, const std::uint16_t payloadId, const std::uint8_t *payload, const std::uint16_t payloadLength) |
| | The same to the groups groupMask names as well as its own.
|
| |
| protocol::Publish | reply (const app::ReplyAddress &request, const std::uint16_t payloadId, const std::uint8_t *payload, const std::uint16_t payloadLength) |
| | The same on behalf of the producer that asked for it (H-11): the reply carries the request's sourceId, connectionId and groupMask.
|
| |
template<typename Listener>
class org::limitless::seqeron::hosted::Application< Listener >
The client stub of a hosted Application (doc/seqeron-protocol-spec.md §17): the surface of app::Application, with the façade itself running in a host process that this talks to over schema 218.
Its listener is the direct Application's, and a Payload reaches it as it would there.
What differs: a Published command can be lost at a gate edge (H-3, H-4), the queries mirror the host's last Gate (H-16), and a snapshot cut is answered from the listener the Config names. The Gateway's is hosted::Gateway.
Single-threaded: every method belongs to the caller's one duty-cycle thread.