Bug #94981 for Proc-PID-File: DESTROY race condition causes warnings in forked children

When using Proc::PID::File, and then forking off child processes, then Proc::PID::File's DESTROY() method is called when each child exits. The problem is that if two children exit very close to each other, then there is a race condition where DESTROY() calls read(), which in turn calls flock(), and the flock can issue a warning such as: (in cleanup) pid "./jnk.pid" already locked: Resource temporarily unavailable It appears that you handled this case for threads in DESTROY already by skipping read() if this is not the main thread. I worked around it with the attached patch, which essentially does the same thing if the current PID is not the same as the PID that wrote the file. The attached patch eliminates the issue for me.