Struct LoopLock 

pub struct LoopLock;

The loop lock mechanism.

Uses flock() to provide advisory locking on .ralph/loop.lock. The lock is automatically released when the process exits (even on crash).

Implementations

impl LoopLock

pub const LOCK_FILE: &'static str = ".ralph/loop.lock"

The relative path to the lock file within the workspace.

pub fn try_acquire( workspace_root: impl AsRef<Path>, prompt: &str, ) -> Result<LockGuard, LockError>

Try to acquire the loop lock (non-blocking).

Arguments

• workspace_root - Root directory of the workspace
• prompt - The prompt/task being executed (stored in lock metadata)

Returns

• Ok(LockGuard) - Lock acquired successfully
• Err(LockError::AlreadyLocked(metadata)) - Another process holds the lock
• Err(LockError::Io(_)) - IO error

pub fn acquire_blocking( workspace_root: impl AsRef<Path>, prompt: &str, ) -> Result<LockGuard, LockError>

Acquire the loop lock, blocking until available.

This should be used with the --exclusive flag to wait for the primary loop slot instead of spawning into a worktree.

Arguments

• workspace_root - Root directory of the workspace
• prompt - The prompt/task being executed

Returns

• Ok(LockGuard) - Lock acquired successfully
• Err(LockError::Io(_)) - IO error

pub fn read_existing( workspace_root: impl AsRef<Path>, ) -> Result<Option<LockMetadata>, LockError>

Read the metadata from an existing lock file.

This can be used to check who holds the lock without acquiring it.

pub fn is_locked(workspace_root: impl AsRef<Path>) -> Result<bool, LockError>

Check if the lock is currently held (without acquiring it).

Returns true if another process holds the lock.