https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128457
--- Comment #1 from Andrea Boido <andreaboido.asti at gmail dot com> --- Patch posted: https://gcc.gnu.org/pipermail/gcc-patches/2026-November/737631.html The cause is the special resolution in Resolve_Membership_Op (sem_res.adb) for a tested expression of a universal numeric type, which converts the bounds to the universal type instead of rounding a universal_real operand to the tested type. The same applies to a plain literal: 0.1 in F .. 1.0 also yields False.
