fix(remote): don't hold control DB Mutex across network I/O

2 months ago

6549fae

Authored

Davy7/20/2026, 10:51:25 PM

The publish protocol and retry executor were holding the shared
control_db Mutex lock for the entire duration of execute_publish,
which includes network I/O. This blocked all other operations that
need the control DB for potentially minutes.

Open a dedicated SQLite connection for the publish execution instead.
WAL mode allows concurrent readers/writers, so the dedicated connection
can write operation state updates without blocking the shared Mutex.

Parent446932d

2 files changed

src-tauri/src/remote/recovery.rs+9−4
src-tauri/src/remote/sync/publish.rs+9−6

dev / openkara

fix(remote): don't hold control DB Mutex across network I/O