Interface ReentryGuard

All Known Implementing Classes:

ReentryGuard.NOPRentryGuard, ReentryGuard.ReentryGuardImpl

public interface ReentryGuard

Guard used to prevent re-entrant (recursive) appender invocations on a per-thread basis.

Implementations are used by appenders and other components that must avoid recursively calling back into themselves (for example when an error causes logging while handling a logging event). Typical usage: check isLocked() before proceeding and call lock() / unlock() around the guarded region.

Concurrency: guards operate on a per-thread basis; callers should treat the guard as thread-local state. Implementations must document their semantics; the provided ReentryGuard.ReentryGuardImpl uses a ThreadLocal to track the locked state for the current thread.

Since:

1.5.21

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static class	ReentryGuard.NOPRentryGuard	No-op implementation that never locks.
static class	ReentryGuard.ReentryGuardImpl	Default per-thread implementation backed by a ThreadLocal<Boolean>.

Method Summary

Modifier and Type	Method	Description
boolean	isLocked()	Return true if the current thread holds the guard (i.e.
void	lock()	Mark the guard as locked for the current thread.
void	unlock()	Release the guard for the current thread.

Method Details

isLocked

boolean isLocked()

Return true if the current thread holds the guard (i.e. is inside a guarded region).

Implementations typically return false if the current thread has not previously called lock() or if the stored value is null.

Returns:

true if the guard is locked for the current thread, false otherwise

lock

void lock()

Mark the guard as locked for the current thread.

Callers must ensure unlock() is invoked in a finally block to avoid leaving the guard permanently locked for the thread.

unlock

void unlock()

Release the guard for the current thread.

After calling unlock() the isLocked() should return false for the current thread (unless lock() is called again).