Thanks for the help. That sounds like a reasonable explanation to me.

In the meantime I've built a construction using PID files.

On script start it saves it's PID to a file
When script ends it removes it's PID file.

When you start the script you check on PID file, if it exists and a proces uses this PID you just exit and wait for the next TASK event.
Not 100% solid, but 99,999999% I guess :)

Looking at http://cpansearch.perl.org/src/ELIZABETH/Sys-RunAlone-0.12/t/RunAlone.t its written in non-portable manner (like a shell script)

Looking at http://cpansearch.perl.org/src/ELIZABETH/Sys-RunAlone-0.12/lib/Sys/RunAlone.pm its a very simple module based around flocking the DATA filehandle -- it will work on windows even if the test suite is broken (should have used Proc::Background or IPC::System::Simple )

However I would not use it because the test suite is still broken and the author knows how to code for portability

I would use File::HomeDir with File::Lockfile ( lock $my_data/appname/.lockfile )

See also only allow single instance of program running, Ensuring only one copy of a perl script is running at a time, Wx::SingleInstanceChecker

See also How can i check my script is already running?, and a curiosity, Paranoid::Lockfile