https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118592
--- Comment #2 from Joseph S. Myers <jsm28 at gcc dot gnu.org> --- Actually, the _FloatN / _FloatNx functions are only reserved if the user defines __STDC_WANT_IEC_60559_TYPES_EXT__ so maybe DEF_EXT_LIB_FLOATN_NX_BUILTINS is right for them. Excluding built-in functions we already have (and also excluding encoding-conversion functions from Annex H, and DFP-only functions): Type-generic: iscanonical, issubnormal (bug 77925), iszero (bug 77926). >From TS 18661-1 and in C23: nextdown nextup llogb fromfp ufromfp fromfpx ufromfpx [note that the types of fromfp functions changed between TS 18661-1 and C23] canonicalize >From TS 18661-1 and in C23 Annex F (only reserved if __STDC_WANT_IEC_60559_EXT__ defined by the user, not otherwise reserved in C23): totalorder totalordermag getpayload setpayload setpayloadsig [last two may not have much that can usefully be done as built-in functions, and types of totalorder functions changed after the original TS] >From TS 18661-1 and -3: the narrowing function families for add, sub, mul, div, sqrt, fma (fadd, faddl, daddl, fMaddfN, fMaddfNx, fMxaddfN, fMxaddfNx etc.) Min/max: fmaximum fminimum fmaximum_num fminimum_num fmaximum_mag fminimum_mag fmaximum_mag_num fminimum_mag_num >From TS 18661-4 and in C23: acospi asinpi atanpi atan2pi cospi sinpi tanpi exp2m1 exp10m1 log2p1 log10p1 logp1 rsqrt compoundn rootn pown powr
