Add support for f16, f64, f128. (#5952)

Generalize the f64 support to support other sizes. Also provide interop
support for `float`, `_Float16`, and `__float128`.

Also lay some groundwork for non-standard floating-point types, though
we don't have any syntax to name them yet.
This commit is contained in:
Richard Smith
2025-08-14 01:14:40 +00:00
committed by GitHub
parent cfe5599144
commit b851e8c423
27 changed files with 826 additions and 379 deletions
+3 -2
View File
@@ -1882,8 +1882,9 @@ static auto TryResolveTypedInst(ImportRefResolver& resolver,
}
return ResolveAsDeduplicated<SemIR::FloatType>(
resolver,
{.type_id = SemIR::TypeType::TypeId, .bit_width_id = bit_width_id});
resolver, {.type_id = SemIR::TypeType::TypeId,
.bit_width_id = bit_width_id,
.float_kind = inst.float_kind});
}
static auto TryResolveTypedInst(ImportRefResolver& resolver,