On 05/06/2017 06:00 PM, Andrew Parsloe wrote: > I have finally got to grips with l3fp's exception handling. (The main > difficulty was realising that an exception like division_by_zero is a > different creature from the flag fp_division_by_zero.) However, I was > surprised to find that \fp_eval:n { ln(1) } raises the underflow flag: > \flag_height:n { fp_underflow } displays 1. > > Andrew Thanks for reporting. This is a bug. Somewhat known to me at least at some point since I wrote in the source the following comment %^^A todo: ln(1) should be "exact zero", not "underflow" I need to revisit l3fp soonish, but sadly a realistic estimate of when that'll be fixed is 1 year. I've opened issue #364 (https://github.com/latex3/latex3/issues/364) Bruno