It would probably be easier/better to just call the same code that myip -l would run versus creating a subprocess to call the actual binary again. I.e. your error handling logic calls a shard method that would be the same as option parser would do when passed the -l option.