Struct AsyncDiskStorage 

pub struct AsyncDiskStorage { /* private fields */ }

High-performance async disk writer.

Writes are sent over an MPSC channel and flushed to disk every 50 ms by a background Tokio task. The write path never blocks the caller.

Implementations

impl AsyncDiskStorage

pub fn new(path: &str) -> Result<Self, DbError>

Open (or create) the log file at path and spawn the background writer task.

Trait Implementations

impl Drop for AsyncDiskStorage

fn drop(&mut self)

On drop, close the sender (signals the writer task to exit) then block until the task has drained its queue and flushed everything to disk.

impl StorageBackend for AsyncDiskStorage

fn write_entry(&self, entry: &LogEntry) -> Result<(), DbError>

Serialize entry to a JSON string and send it to the background writer. This call returns immediately — it never blocks waiting for disk I/O.

fn read_log(&self) -> Result<Vec<LogEntry>, DbError>

Read all entries from the log file into a Vec. Used by EncryptedStorage which needs the full list to decrypt.

fn compact(&self, entries: Vec<LogEntry>) -> Result<(), DbError>

Compact the log: write a binary snapshot, rewrite the log to be empty, then signal the background task to swap the file.

fn read_at(&self, offset: u64, length: u32) -> Result<Vec<u8>, DbError>

Read exactly length bytes from the log at offset.

fn stream_log_into( &self, f: &mut dyn FnMut(LogEntry, u32) -> ControlFlow<(), ()>, ) -> Result<u64, DbError>

Stream log entries into state using snapshot + delta replay.

Fast path (after first compaction):

1. Load binary snapshot → apply all entries in it.
2. Stream only the log lines written AFTER the snapshot (the “delta”).

Slow path (first run, no snapshot): Stream the entire log file line-by-line. No full Vec in RAM.

fn get_size(&self) -> Result<u64, DbError>

Return the current size of the persistent log file in bytes.