Struct Service 

pub struct Service {
    db: Data,
    mutex: MutexMap<OwnedUserId, ()>,
    services: Arc<OnceServices>,
}

Stores and mutates versioned room key backups.

Mutations take a per-user lock so concurrent update paths cannot interleave. Read operations stream directly from the versioned database prefixes.

Fields

db: Data

mutex: MutexMap<OwnedUserId, ()>

services: Arc<OnceServices>

Implementations

impl Service

pub async fn create_backup( &self, user_id: &UserId, backup_metadata: &Raw<BackupAlgorithm>, ) -> Result<String>

Creates an empty backup version and returns its identifier.

The version and initial change tag use separate global sequence numbers. Metadata and the change tag are committed together while the user’s backup lock is held.

Panics

Panics when dispatching either global sequence number fails.

impl Service

pub async fn delete_backup(&self, user_id: &UserId, version: &str)

Deletes a backup version and its stored session keys.

Metadata and the change tag are removed before the version’s key rows are scanned. Missing versions are accepted, and unreadable key rows are skipped during cleanup.

impl Service

pub async fn update_backup<'a>( &self, user_id: &UserId, version: &'a str, backup_metadata: &Raw<BackupAlgorithm>, ) -> Result<&'a str>

Replaces a backup version’s metadata and advances its change tag.

The version must already exist. Its metadata and fresh change tag are committed together while the user’s backup lock is held.

Panics

Panics when dispatching the new global sequence number fails.

impl Service

pub async fn get_latest_backup_version( &self, user_id: &UserId, ) -> Result<String>

Returns a user’s latest numeric backup version.

The greatest parseable version number is selected. Unreadable rows and nonnumeric version names are skipped, and an absent result is reported as not found.

impl Service

pub async fn get_latest_backup( &self, user_id: &UserId, ) -> Result<(String, Raw<BackupAlgorithm>)>

Returns a user’s latest backup version and its algorithm metadata.

The version is selected by Self::get_latest_backup_version. Missing or unreadable metadata is reported as a not-found request error.

impl Service

pub async fn get_backup( &self, user_id: &UserId, version: &str, ) -> Result<Raw<BackupAlgorithm>>

Loads the algorithm metadata for a backup version.

The raw value retains the stored JSON representation while constraining it to [BackupAlgorithm].

impl Service

pub async fn add_keys<'a, S>( &self, user_id: &UserId, version: &str, keys: S, ) -> Result<(usize, u64)>

where S: Stream<Item = (&'a RoomId, &'a str, &'a Raw<KeyBackupData>)> + Send,

Adds a stream of room keys to the latest backup version.

The stream is drained serially while this user’s backup mutations are locked. The returned count and etag describe the completed operation.

Panics

Panics when an accepted key requires a global sequence number that cannot be allocated or persisted.

impl Service

async fn check_backup_version(&self, user_id: &UserId, version: &str) -> Result

impl Service

async fn add_key( &self, user_id: &UserId, version: &str, room_id: &RoomId, session_id: &str, key_data: &Raw<KeyBackupData>, ) -> Result

impl Service

pub async fn get_count_etag( &self, user_id: &UserId, version: &str, ) -> Result<(usize, u64)>

Reads a backup’s current key count and etag.

Mutation callers keep the user lock across this method. Other callers receive two independent current observations without snapshot semantics.

impl Service

pub async fn count_keys(&self, user_id: &UserId, version: &str) -> usize

Counts the session keys stored in a backup version.

Every raw key row under the user’s version prefix contributes to the count.

impl Service

pub async fn get_etag(&self, user_id: &UserId, version: &str) -> Result<u64>

Returns the current change tag for a backup version.

The tag advances whenever accepted key data or metadata changes.

impl Service

pub async fn get_all( &self, user_id: &UserId, version: &str, ) -> BTreeMap<OwnedRoomId, RoomKeyBackup>

Loads every session key in a backup version, grouped by room.

Session identifiers map to their raw key backup data within each room. Unreadable rows are omitted from the result.

impl Service

pub async fn get_room( &self, user_id: &UserId, version: &str, room_id: &RoomId, ) -> BTreeMap<String, Raw<KeyBackupData>>

Loads every backed-up session key for one room.

The returned map is keyed by session ID and retains each value as raw key backup data. Unreadable rows are omitted.

impl Service

pub async fn get_session( &self, user_id: &UserId, version: &str, room_id: &RoomId, session_id: &str, ) -> Result<Raw<KeyBackupData>>

Loads one backed-up session key.

The key is selected by user, backup version, room, and session ID. The stored JSON is returned without eagerly deserializing the key data.

impl Service

pub async fn delete_all_keys( &self, user_id: &UserId, version: &str, ) -> Result<(usize, u64)>

Deletes session keys from a backup version.

The version must exist. Rows whose scan reports an error are skipped. Deletion runs under the user’s backup lock, advances the change tag, and reports a zero remaining count with the new tag.

Panics

Panics when dispatching the new change tag fails.

impl Service

async fn check_backup_exists(&self, user_id: &UserId, version: &str) -> Result

impl Service

fn bump_etag(&self, user_id: &UserId, version: &str) -> u64

impl Service

pub async fn delete_room_keys( &self, user_id: &UserId, version: &str, room_id: &RoomId, ) -> Result<(usize, u64)>

Deletes session keys for one room in a backup version.

The version must exist. Rows whose scan reports an error are skipped. Deletion runs under the user’s backup lock, advances the change tag, and returns the remaining key count from the exact-version scan.

Panics

Panics when dispatching the new change tag fails.

impl Service

pub async fn delete_room_key( &self, user_id: &UserId, version: &str, room_id: &RoomId, session_id: &str, ) -> Result<(usize, u64)>

Deletes one session key from a backup version.

The version must exist. The change tag advances even when the selected key was absent, and the returned count covers every remaining key in the version.

Panics

Panics when dispatching the new change tag fails.

Trait Implementations

impl Service for Service

fn build(args: &Args<'_>) -> Result<Arc<Self>>

Implement the construction of the service instance. Services are generally singletons so expect this to only be called once for a service type. Note that it may be called again after a server reload, but the prior instance will have been dropped first. Failure will shutdown the server with an error.

fn name(&self) -> &str

Return the name of the service. i.e. crate::service::make_name(std::module_path!())

fn worker<'async_trait>( self: Arc<Self>, ) -> Pin<Box<dyn Future<Output = Result> + Send + 'async_trait>>

where Self: 'async_trait,

Implement the service’s worker loop. The service manager spawns a task and calls this function after all services have been built.

fn interrupt<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Interrupt the service. This is sent to initiate a graceful shutdown. The service worker should return from its work loop.

fn clear_cache<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Clear any caches or similar runtime state.

fn memory_usage<'life0, 'life1, 'async_trait>( &'life0 self, _out: &'life1 mut (dyn Write + Send), ) -> Pin<Box<dyn Future<Output = Result> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Memory usage report in a markdown string.

fn unconstrained(&self) -> bool

Return true if the service worker opts out of the tokio cooperative budgeting. This can reduce tail latency at the risk of event loop starvation.