Or should I say 'rather' than 'hardly'..

 Meaning that, possibly,  `C-code datediff` does not result in better processing of date ranges than a fair comparison with the `datediff.sh` script.

Cheers!