[flang-commits] [flang] [llvm] [Flang] Introduce common::numeric_limits (PR #228108)

[Michael Kruse via flang-commits]

Meinersbur wrote:

NB: `std::numeric_limits<__float128>` has been specialized by libstdc++ [only very recently](https://github.com/gcc-mirror/gcc/commit/f150a084e25eaa98afaf25e2211001047ecc8d44) and libc++ doesn't specialize it at all. `std::numeric_limits<__float128>::infinity()` is being used by flang-rt
https://github.com/llvm/llvm-project/blob/b11606a912ae297cc4b9db6a7c59f44b0e326aa6/flang-rt/lib/runtime/transformational.cpp#L322-L323
 which on an unspecialized template returns a value-initialized value (0.0).

https://github.com/llvm/llvm-project/pull/228108