Rather than hardcoding assumptions or conclusions about what the machine is now, why not future-proof it?

Yes - that is certainly the plan.

It seems to me that if I need this functionality, someone else will as well. Or, at the very least, would benefit from a head-start in developing code for the next part of dealing with a GeoPackage. So some of the code will probably form a module for CPAN.

since you have not apparently tried sanity-checking your interpretation with another tool

I haven't yet as I have been away on training much of this week and haven't had the opportunity! I also feel that an external tool won't be very helpful for checking te envelope data. It will be the next step once I've decoded some fo the actual geometry data.