Default Int32 coordinates type overflow issue

[zjans]

The default data type for X/Y coordinates leads to an integer overflow in some cases.
For some datasets, the projected coordinates are scaled by a factor, and they don't fit into Int32.

Would it be possible to add a check and change the type accordingly?

[evetion]

Thanks for creating an issue! Can you describe exactly (i.e. with code) what you're trying to do?

LAS/LAZ files store the coordinates as offsetted and scaled Int32. So on writing (if that's what you're trying to do), we can't change the type, but have to change the offset and/or scale.

At the moment, when reading you have to do the inverse operations yourself, (so .X * scale + offset) to get the correct floating point coordinate back.

[zjans]

Of course! Thanks, that fixed it. I'm working with two different datasets which should've been scaled in the same way – but aren't. The header included the correct scale+offset.