Re: [PATCH] c++ testsuite: Add test for PR107703

[Jason Merrill via Gcc-patches]

On 3/10/23 10:43, Jakub Jelinek wrote:

Hi!

This is on top of the
https://gcc.gnu.org/pipermail/gcc-patches/2022-November/606398.html
and
https://gcc.gnu.org/pipermail/gcc-patches/2023-March/613724.html
patches (to be precise, the latter isn't essential for it), I've
realized that for the PR107703 bugfix in the first patch I haven't
added some test coverage that the extended floating vs. integral
or vice versa conversions work correctly.

This new testcase adds such checks.  And when writing it I've
found that in ext-floating.h header in the testsuite I forgot back
in November to remove #undef __STDCPP_BFLOAT16_T__ which was left
there because the bfloat16 support wasn't in yet.

The new testcase (and all older ext-floating*.C tests too) passes
on vanilla trunk without the ext-floating.h change (x86_64-linux
-m32/-m64) and with the PR107703 fix also with the ext-floating.h
change.

Ok for trunk?

OK.

2023-03-10  Jakub Jelinek  <ja...@redhat.com>

        PR target/107703
        * g++.dg/cpp23/ext-floating.h (__STDCPP_BFLOAT16_T__): Don't undefine
        it.
        (std::bfloat16_t): Use decltype (0.0bf16) like libstdc++, rather than
        __bf16.
        * g++.dg/cpp23/ext-floating14.C: New test.

--- gcc/testsuite/g++.dg/cpp23/ext-floating.h.jj        2022-09-27 
08:03:27.118982749 +0200
+++ gcc/testsuite/g++.dg/cpp23/ext-floating.h   2023-03-10 15:04:01.647824767 
+0100
@@ -14,9 +14,8 @@ namespace std
    #ifdef __STDCPP_FLOAT128_T__
    using float128_t = _Float128;
    #endif
-  #undef __STDCPP_BFLOAT16_T__
    #ifdef __STDCPP_BFLOAT16_T__
-  using bfloat16_t = __bf16; // ???
+  using bfloat16_t = decltype (0.0bf16);
    #endif
    template<typename T, T v> struct integral_constant {
      static constexpr T value = v;
--- gcc/testsuite/g++.dg/cpp23/ext-floating14.C.jj      2023-03-10 
14:12:17.658925358 +0100
+++ gcc/testsuite/g++.dg/cpp23/ext-floating14.C 2023-03-10 15:32:26.912057825 
+0100
@@ -0,0 +1,585 @@
+// P1467R9 - Extended floating-point types and standard names.
+// PR target/107703
+// { dg-do run { target c++23 } }
+// { dg-options "-fexcess-precision=standard" }
+
+#include "ext-floating.h"
+
+#ifdef __SIZEOF_INT128__
+#define INT128_MAX ((signed __int128) ((~(unsigned __int128) 0) >> 1))
+#endif
+
+template <typename T, typename F>
+[[gnu::noipa]] T cvt (F f)
+{
+  return T (F (f));
+}
+
+int
+main ()
+{
+  // __FLT32_MAX_EXP__ is 128, so make sure all unsigned long long and 
unsigned __int128
+  // values fit into it.  __FLT16_MAX__ is 65504.0f16, so we need to be
+  // careful for that.
+#if __SIZEOF_LONG_LONG__ * __CHAR_BIT__ <= 128
+#if !defined(__SIZEOF_INT128__) || __SIZEOF_INT128__ * __CHAR_BIT__ == 128
+#ifdef __STDCPP_FLOAT16_T__
+  if (cvt <std::float16_t, signed char> (42) != (std::float16_t) 42
+      || cvt <std::float16_t, signed char> (-42) != (std::float16_t) -42
+#if __SCHAR_MAX__ < 65504
+      || cvt <std::float16_t, signed char> (__SCHAR_MAX__) != (std::float16_t) 
__SCHAR_MAX__
+      || cvt <std::float16_t, signed char> (-__SCHAR_MAX__ - 1) != 
(std::float16_t) (-__SCHAR_MAX__ - 1)
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned char> (42) != (std::float16_t) 42
+#if __SCHAR_MAX__ * 2 + 1 < 65504
+      || cvt <std::float16_t, unsigned char> ((unsigned char) ~0) != 
(std::float16_t) ((unsigned char) ~0)
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <std::float16_t, signed short> (42) != (std::float16_t) 42
+      || cvt <std::float16_t, signed short> (-42) != (std::float16_t) -42
+#if __SHRT_MAX__ < 65504
+      || cvt <std::float16_t, signed short> (__SHRT_MAX__) != (std::float16_t) 
__SHRT_MAX__
+      || cvt <std::float16_t, signed short> (-__SHRT_MAX__ - 1) != 
(std::float16_t) (-__SHRT_MAX__ - 1)
+#else
+      || cvt <std::float16_t, signed short> (65504) != (std::float16_t) 65504
+      || cvt <std::float16_t, signed short> (-65504) != (std::float16_t) -65504
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned short> (42) != (std::float16_t) 42
+      || cvt <std::float16_t, unsigned short> (65504U) != (std::float16_t) 
65504U)
+    __builtin_abort ();
+  if (cvt <std::float16_t, signed int> (42) != (std::float16_t) 42
+      || cvt <std::float16_t, signed int> (-42) != (std::float16_t) -42
+#if __INT_MAX__ < 65504
+      || cvt <std::float16_t, signed short> (__INT_MAX__) != (std::float16_t) 
__INT_MAX__
+      || cvt <std::float16_t, signed short> (-__INT_MAX__ - 1) != 
(std::float16_t) (-__INT_MAX__ - 1)
+#else
+      || cvt <std::float16_t, signed int> (65504) != (std::float16_t) 65504
+      || cvt <std::float16_t, signed int> (-65504) != (std::float16_t) -65504
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned int> (42) != (std::float16_t) 42U
+      || cvt <std::float16_t, unsigned int> (65504U) != (std::float16_t) 
65504U)
+    __builtin_abort ();
+  if (cvt <std::float16_t, signed long int> (42L) != (std::float16_t) 42L
+      || cvt <std::float16_t, signed long int> (-42L) != (std::float16_t) -42L
+      || cvt <std::float16_t, signed long int> (65504L) != (std::float16_t) 
65504L
+      || cvt <std::float16_t, signed long int> (-65504L) != (std::float16_t) 
-65504L)
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned long int> (42UL) != (std::float16_t) 42UL
+      || cvt <std::float16_t, unsigned long int> (65504UL) != (std::float16_t) 
65504UL)
+    __builtin_abort ();
+  if (cvt <std::float16_t, signed long long int> (42LL) != (std::float16_t) 
42LL
+      || cvt <std::float16_t, signed long long int> (-42LL) != 
(std::float16_t) -42LL
+      || cvt <std::float16_t, signed long long int> (65504LL) != 
(std::float16_t) 65504LL
+      || cvt <std::float16_t, signed long long int> (-65504LL) != 
(std::float16_t) -65504LL)
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned long long int> (42ULL) != (std::float16_t) 
42ULL
+      || cvt <std::float16_t, unsigned long long int> (65504ULL) != 
(std::float16_t) 65504ULL)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <std::float16_t, signed __int128> (42LL) != (std::float16_t) (signed 
__int128) 42LL
+      || cvt <std::float16_t, signed __int128> (-42LL) != (std::float16_t) 
(signed __int128) -42LL
+      || cvt <std::float16_t, signed __int128> (65504LL) != (std::float16_t) 
(signed __int128) 65504LL
+      || cvt <std::float16_t, signed __int128> (-65504LL) != (std::float16_t) 
(signed __int128) -65504LL)
+    __builtin_abort ();
+  if (cvt <std::float16_t, unsigned __int128> (42ULL) != (std::float16_t) 
(unsigned __int128) 42ULL
+      || cvt <std::float16_t, unsigned __int128> (65504ULL) != 
(std::float16_t) (unsigned __int128) 65504ULL)
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_BFLOAT16_T__
+  if (cvt <std::bfloat16_t, signed char> (42) != (std::bfloat16_t) 42
+      || cvt <std::bfloat16_t, signed char> (-42) != (std::bfloat16_t) -42
+      || cvt <std::bfloat16_t, signed char> (__SCHAR_MAX__) != 
(std::bfloat16_t) __SCHAR_MAX__
+      || cvt <std::bfloat16_t, signed char> (-__SCHAR_MAX__ - 1) != 
(std::bfloat16_t) (-__SCHAR_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned char> (42) != (std::bfloat16_t) 42
+      || cvt <std::bfloat16_t, unsigned char> ((unsigned char) ~0) != 
(std::bfloat16_t) ((unsigned char) ~0))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, signed short> (42) != (std::bfloat16_t) 42
+      || cvt <std::bfloat16_t, signed short> (-42) != (std::bfloat16_t) -42
+      || cvt <std::bfloat16_t, signed short> (__SHRT_MAX__) != 
(std::bfloat16_t) __SHRT_MAX__
+      || cvt <std::bfloat16_t, signed short> (-__SHRT_MAX__ - 1) != 
(std::bfloat16_t) (-__SHRT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned short> (42) != (std::bfloat16_t) 42
+      || cvt <std::bfloat16_t, unsigned short> ((unsigned short) ~0) != 
(std::bfloat16_t) ((unsigned short) ~0))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, signed int> (42) != (std::bfloat16_t) 42
+      || cvt <std::bfloat16_t, signed int> (-42) != (std::bfloat16_t) -42
+      || cvt <std::bfloat16_t, signed int> (__INT_MAX__) != (std::bfloat16_t) 
__INT_MAX__
+      || cvt <std::bfloat16_t, signed int> (-__INT_MAX__ - 1) != 
(std::bfloat16_t) (-__INT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned int> (42) != (std::bfloat16_t) 42U
+      || cvt <std::bfloat16_t, unsigned int> (~0U) != (std::bfloat16_t) ~0U)
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, signed long int> (42L) != (std::bfloat16_t) 42L
+      || cvt <std::bfloat16_t, signed long int> (-42L) != (std::bfloat16_t) 
-42L
+      || cvt <std::bfloat16_t, signed long int> (__LONG_MAX__) != 
(std::bfloat16_t) __LONG_MAX__
+      || cvt <std::bfloat16_t, signed long int> (-__LONG_MAX__ - 1) != 
(std::bfloat16_t) (-__LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned long int> (42UL) != (std::bfloat16_t) 42UL
+      || cvt <std::bfloat16_t, unsigned long int> (~0UL) != (std::bfloat16_t) 
~0UL)
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, signed long long int> (42LL) != (std::bfloat16_t) 
42LL
+      || cvt <std::bfloat16_t, signed long long int> (-42LL) != 
(std::bfloat16_t) -42LL
+      || cvt <std::bfloat16_t, signed long long int> (__LONG_LONG_MAX__) != 
(std::bfloat16_t) __LONG_LONG_MAX__
+      || cvt <std::bfloat16_t, signed long long int> (-__LONG_LONG_MAX__ - 1) 
!= (std::bfloat16_t) (-__LONG_LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned long long int> (42ULL) != 
(std::bfloat16_t) 42ULL
+      || cvt <std::bfloat16_t, unsigned long long int> (~0ULL) != 
(std::bfloat16_t) ~0ULL)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <std::bfloat16_t, signed __int128> (42LL) != (std::bfloat16_t) 
(signed __int128) 42LL
+      || cvt <std::bfloat16_t, signed __int128> (-42LL) != (std::bfloat16_t) 
(signed __int128) -42LL
+      || cvt <std::bfloat16_t, signed __int128> (INT128_MAX) != 
(std::bfloat16_t) INT128_MAX
+      || cvt <std::bfloat16_t, signed __int128> (-INT128_MAX - 1) != 
(std::bfloat16_t) (-INT128_MAX - 1))
+    __builtin_abort ();
+  if (cvt <std::bfloat16_t, unsigned __int128> (42ULL) != (std::bfloat16_t) 
(unsigned __int128) 42ULL
+      || cvt <std::bfloat16_t, unsigned __int128> (~(unsigned __int128) 0) != 
(std::bfloat16_t) (~(unsigned __int128) 0))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT32_T__
+  if (cvt <std::float32_t, signed char> (42) != (std::float32_t) 42
+      || cvt <std::float32_t, signed char> (-42) != (std::float32_t) -42
+      || cvt <std::float32_t, signed char> (__SCHAR_MAX__) != (std::float32_t) 
__SCHAR_MAX__
+      || cvt <std::float32_t, signed char> (-__SCHAR_MAX__ - 1) != 
(std::float32_t) (-__SCHAR_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned char> (42) != (std::float32_t) 42
+      || cvt <std::float32_t, unsigned char> ((unsigned char) ~0) != 
(std::float32_t) ((unsigned char) ~0))
+    __builtin_abort ();
+  if (cvt <std::float32_t, signed short> (42) != (std::float32_t) 42
+      || cvt <std::float32_t, signed short> (-42) != (std::float32_t) -42
+      || cvt <std::float32_t, signed short> (__SHRT_MAX__) != (std::float32_t) 
__SHRT_MAX__
+      || cvt <std::float32_t, signed short> (-__SHRT_MAX__ - 1) != 
(std::float32_t) (-__SHRT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned short> (42) != (std::float32_t) 42
+      || cvt <std::float32_t, unsigned short> ((unsigned short) ~0) != 
(std::float32_t) ((unsigned short) ~0))
+    __builtin_abort ();
+  if (cvt <std::float32_t, signed int> (42) != (std::float32_t) 42
+      || cvt <std::float32_t, signed int> (-42) != (std::float32_t) -42
+      || cvt <std::float32_t, signed int> (__INT_MAX__) != (std::float32_t) 
__INT_MAX__
+      || cvt <std::float32_t, signed int> (-__INT_MAX__ - 1) != 
(std::float32_t) (-__INT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned int> (42) != (std::float32_t) 42U
+      || cvt <std::float32_t, unsigned int> (~0U) != (std::float32_t) ~0U)
+    __builtin_abort ();
+  if (cvt <std::float32_t, signed long int> (42L) != (std::float32_t) 42L
+      || cvt <std::float32_t, signed long int> (-42L) != (std::float32_t) -42L
+      || cvt <std::float32_t, signed long int> (__LONG_MAX__) != 
(std::float32_t) __LONG_MAX__
+      || cvt <std::float32_t, signed long int> (-__LONG_MAX__ - 1) != 
(std::float32_t) (-__LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned long int> (42UL) != (std::float32_t) 42UL
+      || cvt <std::float32_t, unsigned long int> (~0UL) != (std::float32_t) 
~0UL)
+    __builtin_abort ();
+  if (cvt <std::float32_t, signed long long int> (42LL) != (std::float32_t) 
42LL
+      || cvt <std::float32_t, signed long long int> (-42LL) != 
(std::float32_t) -42LL
+      || cvt <std::float32_t, signed long long int> (__LONG_LONG_MAX__) != 
(std::float32_t) __LONG_LONG_MAX__
+      || cvt <std::float32_t, signed long long int> (-__LONG_LONG_MAX__ - 1) 
!= (std::float32_t) (-__LONG_LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned long long int> (42ULL) != (std::float32_t) 
42ULL
+      || cvt <std::float32_t, unsigned long long int> (~0ULL) != 
(std::float32_t) ~0ULL)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <std::float32_t, signed __int128> (42LL) != (std::float32_t) (signed 
__int128) 42LL
+      || cvt <std::float32_t, signed __int128> (-42LL) != (std::float32_t) 
(signed __int128) -42LL
+      || cvt <std::float32_t, signed __int128> (INT128_MAX) != 
(std::float32_t) INT128_MAX
+      || cvt <std::float32_t, signed __int128> (-INT128_MAX - 1) != 
(std::float32_t) (-INT128_MAX - 1))
+    __builtin_abort ();
+  if (cvt <std::float32_t, unsigned __int128> (42ULL) != (std::float32_t) 
(unsigned __int128) 42ULL
+      || cvt <std::float32_t, unsigned __int128> (~(unsigned __int128) 0) != 
(std::float32_t) (~(unsigned __int128) 0))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT64_T__
+  if (cvt <std::float64_t, signed char> (42) != (std::float64_t) 42
+      || cvt <std::float64_t, signed char> (-42) != (std::float64_t) -42
+      || cvt <std::float64_t, signed char> (__SCHAR_MAX__) != (std::float64_t) 
__SCHAR_MAX__
+      || cvt <std::float64_t, signed char> (-__SCHAR_MAX__ - 1) != 
(std::float64_t) (-__SCHAR_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned char> (42) != (std::float64_t) 42
+      || cvt <std::float64_t, unsigned char> ((unsigned char) ~0) != 
(std::float64_t) ((unsigned char) ~0))
+    __builtin_abort ();
+  if (cvt <std::float64_t, signed short> (42) != (std::float64_t) 42
+      || cvt <std::float64_t, signed short> (-42) != (std::float64_t) -42
+      || cvt <std::float64_t, signed short> (__SHRT_MAX__) != (std::float64_t) 
__SHRT_MAX__
+      || cvt <std::float64_t, signed short> (-__SHRT_MAX__ - 1) != 
(std::float64_t) (-__SHRT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned short> (42) != (std::float64_t) 42
+      || cvt <std::float64_t, unsigned short> ((unsigned short) ~0) != 
(std::float64_t) ((unsigned short) ~0))
+    __builtin_abort ();
+  if (cvt <std::float64_t, signed int> (42) != (std::float64_t) 42
+      || cvt <std::float64_t, signed int> (-42) != (std::float64_t) -42
+      || cvt <std::float64_t, signed int> (__INT_MAX__) != (std::float64_t) 
__INT_MAX__
+      || cvt <std::float64_t, signed int> (-__INT_MAX__ - 1) != 
(std::float64_t) (-__INT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned int> (42) != (std::float64_t) 42U
+      || cvt <std::float64_t, unsigned int> (~0U) != (std::float64_t) ~0U)
+    __builtin_abort ();
+  if (cvt <std::float64_t, signed long int> (42L) != (std::float64_t) 42L
+      || cvt <std::float64_t, signed long int> (-42L) != (std::float64_t) -42L
+      || cvt <std::float64_t, signed long int> (__LONG_MAX__) != 
(std::float64_t) __LONG_MAX__
+      || cvt <std::float64_t, signed long int> (-__LONG_MAX__ - 1) != 
(std::float64_t) (-__LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned long int> (42UL) != (std::float64_t) 42UL
+      || cvt <std::float64_t, unsigned long int> (~0UL) != (std::float64_t) 
~0UL)
+    __builtin_abort ();
+  if (cvt <std::float64_t, signed long long int> (42LL) != (std::float64_t) 
42LL
+      || cvt <std::float64_t, signed long long int> (-42LL) != 
(std::float64_t) -42LL
+      || cvt <std::float64_t, signed long long int> (__LONG_LONG_MAX__) != 
(std::float64_t) __LONG_LONG_MAX__
+      || cvt <std::float64_t, signed long long int> (-__LONG_LONG_MAX__ - 1) 
!= (std::float64_t) (-__LONG_LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned long long int> (42ULL) != (std::float64_t) 
42ULL
+      || cvt <std::float64_t, unsigned long long int> (~0ULL) != 
(std::float64_t) ~0ULL)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <std::float64_t, signed __int128> (42LL) != (std::float64_t) (signed 
__int128) 42LL
+      || cvt <std::float64_t, signed __int128> (-42LL) != (std::float64_t) 
(signed __int128) -42LL
+      || cvt <std::float64_t, signed __int128> (INT128_MAX) != 
(std::float64_t) INT128_MAX
+      || cvt <std::float64_t, signed __int128> (-INT128_MAX - 1) != 
(std::float64_t) (-INT128_MAX - 1))
+    __builtin_abort ();
+  if (cvt <std::float64_t, unsigned __int128> (42ULL) != (std::float64_t) 
(unsigned __int128) 42ULL
+      || cvt <std::float64_t, unsigned __int128> (~(unsigned __int128) 0) != 
(std::float64_t) (~(unsigned __int128) 0))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT128_T__
+  if (cvt <std::float128_t, signed char> (42) != (std::float128_t) 42
+      || cvt <std::float128_t, signed char> (-42) != (std::float128_t) -42
+      || cvt <std::float128_t, signed char> (__SCHAR_MAX__) != 
(std::float128_t) __SCHAR_MAX__
+      || cvt <std::float128_t, signed char> (-__SCHAR_MAX__ - 1) != 
(std::float128_t) (-__SCHAR_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned char> (42) != (std::float128_t) 42
+      || cvt <std::float128_t, unsigned char> ((unsigned char) ~0) != 
(std::float128_t) ((unsigned char) ~0))
+    __builtin_abort ();
+  if (cvt <std::float128_t, signed short> (42) != (std::float128_t) 42
+      || cvt <std::float128_t, signed short> (-42) != (std::float128_t) -42
+      || cvt <std::float128_t, signed short> (__SHRT_MAX__) != 
(std::float128_t) __SHRT_MAX__
+      || cvt <std::float128_t, signed short> (-__SHRT_MAX__ - 1) != 
(std::float128_t) (-__SHRT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned short> (42) != (std::float128_t) 42
+      || cvt <std::float128_t, unsigned short> ((unsigned short) ~0) != 
(std::float128_t) ((unsigned short) ~0))
+    __builtin_abort ();
+  if (cvt <std::float128_t, signed int> (42) != (std::float128_t) 42
+      || cvt <std::float128_t, signed int> (-42) != (std::float128_t) -42
+      || cvt <std::float128_t, signed int> (__INT_MAX__) != (std::float128_t) 
__INT_MAX__
+      || cvt <std::float128_t, signed int> (-__INT_MAX__ - 1) != 
(std::float128_t) (-__INT_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned int> (42) != (std::float128_t) 42U
+      || cvt <std::float128_t, unsigned int> (~0U) != (std::float128_t) ~0U)
+    __builtin_abort ();
+  if (cvt <std::float128_t, signed long int> (42L) != (std::float128_t) 42L
+      || cvt <std::float128_t, signed long int> (-42L) != (std::float128_t) 
-42L
+      || cvt <std::float128_t, signed long int> (__LONG_MAX__) != 
(std::float128_t) __LONG_MAX__
+      || cvt <std::float128_t, signed long int> (-__LONG_MAX__ - 1) != 
(std::float128_t) (-__LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned long int> (42UL) != (std::float128_t) 42UL
+      || cvt <std::float128_t, unsigned long int> (~0UL) != (std::float128_t) 
~0UL)
+    __builtin_abort ();
+  if (cvt <std::float128_t, signed long long int> (42LL) != (std::float128_t) 
42LL
+      || cvt <std::float128_t, signed long long int> (-42LL) != 
(std::float128_t) -42LL
+      || cvt <std::float128_t, signed long long int> (__LONG_LONG_MAX__) != 
(std::float128_t) __LONG_LONG_MAX__
+      || cvt <std::float128_t, signed long long int> (-__LONG_LONG_MAX__ - 1) 
!= (std::float128_t) (-__LONG_LONG_MAX__ - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned long long int> (42ULL) != 
(std::float128_t) 42ULL
+      || cvt <std::float128_t, unsigned long long int> (~0ULL) != 
(std::float128_t) ~0ULL)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <std::float128_t, signed __int128> (42LL) != (std::float128_t) 
(signed __int128) 42LL
+      || cvt <std::float128_t, signed __int128> (-42LL) != (std::float128_t) 
(signed __int128) -42LL
+      || cvt <std::float128_t, signed __int128> (INT128_MAX) != 
(std::float128_t) INT128_MAX
+      || cvt <std::float128_t, signed __int128> (-INT128_MAX - 1) != 
(std::float128_t) (-INT128_MAX - 1))
+    __builtin_abort ();
+  if (cvt <std::float128_t, unsigned __int128> (42ULL) != (std::float128_t) 
(unsigned __int128) 42ULL
+      || cvt <std::float128_t, unsigned __int128> (~(unsigned __int128) 0) != 
(std::float128_t) (~(unsigned __int128) 0))
+    __builtin_abort ();
+#endif
+#endif
+
+#ifdef __STDCPP_FLOAT16_T__
+  if (cvt <signed char, std::float16_t> (42.0f16) != (signed char) 
(std::float16_t) 42.0f16
+      || cvt <signed char, std::float16_t> (-42.0f16) != (signed char) 
(std::float16_t) -42.0f16
+#if __SCHAR_MAX__ < 65504
+      || cvt <signed char, std::float16_t> ((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2)) != (signed char) (std::float16_t) ((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))
+      || cvt <signed char, std::float16_t> (-((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2))) != (signed char) (std::float16_t) (-((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2)))
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <unsigned char, std::float16_t> (42.0f16) != (unsigned char) 
(std::float16_t) 42.0f16
+#if __SCHAR_MAX__ * 2 + 1 < 65504
+      || cvt <unsigned char, std::float16_t> ((unsigned char) 1 << (__CHAR_BIT__ * 
sizeof (unsigned char) - 1)) != (unsigned char) (std::float16_t) ((unsigned char) 1 << 
(__CHAR_BIT__ * sizeof (unsigned char) - 1))
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <signed short, std::float16_t> (42.0f16) != (signed short) 
(std::float16_t) 42.0f16
+      || cvt <signed short, std::float16_t> (-42.0f16) != (signed short) 
(std::float16_t) -42.0f16
+#if __SHRT_MAX__ < 65504
+      || cvt <signed short, std::float16_t> ((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2)) != (signed short) (std::float16_t) ((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))
+      || cvt <signed short, std::float16_t> (-((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2))) != (signed short) (std::float16_t) (-((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2)))
+#else
+      || cvt <signed short, std::float16_t> (65504.0f16) != (signed short) 
(std::float16_t) 65504.0f16
+      || cvt <signed short, std::float16_t> (-65504.0f16) != (signed short) 
(std::float16_t) -65504.0f16
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <unsigned short, std::float16_t> (42.0f16) != (unsigned short) 
(std::float16_t) 42.0f16
+      || cvt <unsigned short, std::float16_t> (65504.0f16) != (unsigned short) 
(std::float16_t) 65504.0f16)
+    __builtin_abort ();
+  if (cvt <signed int, std::float16_t> (42.0f16) != (signed int) 
(std::float16_t) 42.0f16
+      || cvt <signed int, std::float16_t> (-42.0f16) != (signed int) 
(std::float16_t) -42.0f16
+#if __INT_MAX__ < 65504
+      || cvt <signed int, std::float16_t> ((signed int) 1 << (__CHAR_BIT__ * sizeof 
(signed int) - 2)) != (signed int) (std::float16_t) ((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))
+      || cvt <signed int, std::float16_t> (-((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))) != (signed int) (std::float16_t) (-((signed int) 1 << 
(__CHAR_BIT__ * sizeof (signed int) - 2)))
+#else
+      || cvt <signed int, std::float16_t> (65504.0f16) != (signed int) 
(std::float16_t) 65504.0f16
+      || cvt <signed int, std::float16_t> (-65504.0f16) != (signed int) 
(std::float16_t) -65504.0f16
+#endif
+     )
+    __builtin_abort ();
+  if (cvt <unsigned int, std::float16_t> (42.0f16) != (unsigned int) 
(std::float16_t) 42.0f16
+      || cvt <unsigned int, std::float16_t> (65504.0f16) != (unsigned int) 
(std::float16_t) 65504.0f16)
+    __builtin_abort ();
+  if (cvt <signed long int, std::float16_t> (42.0f16) != (signed long int) 
(std::float16_t) 42.0f16
+      || cvt <signed long int, std::float16_t> (-42.0f16) != (signed long int) 
(std::float16_t) -42.0f16
+      || cvt <signed long int, std::float16_t> (65504.0f16) != (signed long 
int) (std::float16_t) 65504.0f16
+      || cvt <signed long int, std::float16_t> (-65504.0f16) != (signed long 
int) (std::float16_t) -65504.0f16)
+    __builtin_abort ();
+  if (cvt <unsigned long int, std::float16_t> (42.0f16) != (unsigned long int) 
(std::float16_t) 42.0f16
+      || cvt <unsigned long int, std::float16_t> (65504.0f16) != (unsigned 
long int) (std::float16_t) 65504.0f16)
+    __builtin_abort ();
+  if (cvt <signed long long int, std::float16_t> (42.0f16) != (signed long 
long int) (std::float16_t) 42.0f16
+      || cvt <signed long long int, std::float16_t> (-42.0f16) != (signed long 
long int) (std::float16_t) -42.0f16
+      || cvt <signed long long int, std::float16_t> (65504.0f16) != (signed 
long long int) (std::float16_t) 65504.0f16
+      || cvt <signed long long int, std::float16_t> (-65504.0f16) != (signed 
long long int) (std::float16_t) -65504.0f16)
+    __builtin_abort ();
+  if (cvt <unsigned long long int, std::float16_t> (42.0f16) != (unsigned long 
long int) (std::float16_t) 42.0f16
+      || cvt <unsigned long long int, std::float16_t> (65504.0f16) != 
(unsigned long long int) (std::float16_t) 65504.0f16)
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <signed __int128, std::float16_t> (42.0f16) != (signed __int128) 
(std::float16_t) 42.0f16
+      || cvt <signed __int128, std::float16_t> (-42.0f16) != (signed __int128) 
(std::float16_t) -42.0f16
+      || cvt <signed __int128, std::float16_t> (65504.0f16) != (signed 
__int128) (std::float16_t) 65504.0f16
+      || cvt <signed __int128, std::float16_t> (-65504.0f16) != (signed 
__int128) (std::float16_t) -65504.0f16)
+    __builtin_abort ();
+  if (cvt <unsigned __int128, std::float16_t> (42.0f16) != (unsigned __int128) 
(std::float16_t) 42.0f16
+      || cvt <unsigned __int128, std::float16_t> (65504.0f16) != (unsigned 
__int128) (std::float16_t) 65504.0f16)
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_BFLOAT16_T__
+  if (cvt <signed char, std::bfloat16_t> (42.0bf16) != (signed char) 
(std::bfloat16_t) 42.0bf16
+      || cvt <signed char, std::bfloat16_t> (-42.0bf16) != (signed char) 
(std::bfloat16_t) -42.0bf16
+      || cvt <signed char, std::bfloat16_t> ((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2)) != (signed char) (std::bfloat16_t) ((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))
+      || cvt <signed char, std::bfloat16_t> (-((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2))) != (signed char) (std::bfloat16_t) (-((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned char, std::bfloat16_t> (42.0bf16) != (unsigned char) 
(std::bfloat16_t) 42.0bf16
+      || cvt <unsigned char, std::bfloat16_t> ((unsigned char) 1 << (__CHAR_BIT__ * 
sizeof (unsigned char) - 1)) != (unsigned char) (std::bfloat16_t) ((unsigned char) 1 << 
(__CHAR_BIT__ * sizeof (unsigned char) - 1)))
+    __builtin_abort ();
+  if (cvt <signed short, std::bfloat16_t> (42.0bf16) != (signed short) 
(std::bfloat16_t) 42.0bf16
+      || cvt <signed short, std::bfloat16_t> (-42.0bf16) != (signed short) 
(std::bfloat16_t) -42.0bf16
+      || cvt <signed short, std::bfloat16_t> ((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2)) != (signed short) (std::bfloat16_t) ((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))
+      || cvt <signed short, std::bfloat16_t> (-((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2))) != (signed short) (std::bfloat16_t) (-((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned short, std::bfloat16_t> (42.0bf16) != (unsigned short) 
(std::bfloat16_t) 42.0bf16
+      || cvt <unsigned short, std::bfloat16_t> ((unsigned short) 1 << (__CHAR_BIT__ 
* sizeof (unsigned short) - 1)) != (unsigned short) (std::bfloat16_t) ((unsigned short) 1 
<< (__CHAR_BIT__ * sizeof (unsigned short) - 1)))
+    __builtin_abort ();
+  if (cvt <signed int, std::bfloat16_t> (42.0bf16) != (signed int) 
(std::bfloat16_t) 42.0bf16
+      || cvt <signed int, std::bfloat16_t> (-42.0bf16) != (signed int) 
(std::bfloat16_t) -42.0bf16
+      || cvt <signed int, std::bfloat16_t> ((signed int) 1 << (__CHAR_BIT__ * sizeof 
(signed int) - 2)) != (signed int) (std::bfloat16_t) ((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))
+      || cvt <signed int, std::bfloat16_t> (-((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))) != (signed int) (std::bfloat16_t) (-((signed int) 1 << 
(__CHAR_BIT__ * sizeof (signed int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned int, std::bfloat16_t> (42.0bf16) != (unsigned int) 
(std::bfloat16_t) 42.0bf16
+      || cvt <unsigned int, std::bfloat16_t> ((unsigned int) 1 << (__CHAR_BIT__ * 
sizeof (unsigned int) - 1)) != (unsigned int) (std::bfloat16_t) ((unsigned int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long int, std::bfloat16_t> (42.0bf16) != (signed long int) 
(std::bfloat16_t) 42.0bf16
+      || cvt <signed long int, std::bfloat16_t> (-42.0bf16) != (signed long 
int) (std::bfloat16_t) -42.0bf16
+      || cvt <signed long int, std::bfloat16_t> ((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2)) != (signed long int) (std::bfloat16_t) ((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))
+      || cvt <signed long int, std::bfloat16_t> (-((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2))) != (signed long int) (std::bfloat16_t) (-((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long int, std::bfloat16_t> (42.0bf16) != (unsigned long 
int) (std::bfloat16_t) 42.0bf16
+      || cvt <unsigned long int, std::bfloat16_t> ((unsigned long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long int) - 1)) != (unsigned long int) (std::bfloat16_t) 
((unsigned long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long long int, std::bfloat16_t> (42.0bf16) != (signed long 
long int) (std::bfloat16_t) 42.0bf16
+      || cvt <signed long long int, std::bfloat16_t> (-42.0bf16) != (signed 
long long int) (std::bfloat16_t) -42.0bf16
+      || cvt <signed long long int, std::bfloat16_t> ((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2)) != (signed long long int) (std::bfloat16_t) 
((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))
+      || cvt <signed long long int, std::bfloat16_t> (-((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2))) != (signed long long int) (std::bfloat16_t) 
(-((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long long int, std::bfloat16_t> (42.0bf16) != (unsigned 
long long int) (std::bfloat16_t) 42.0bf16
+      || cvt <unsigned long long int, std::bfloat16_t> ((unsigned long long int) 1 
<< (__CHAR_BIT__ * sizeof (unsigned long long int) - 1)) != (unsigned long long int) 
(std::bfloat16_t) ((unsigned long long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long long 
int) - 1)))
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <signed __int128, std::bfloat16_t> (42.0bf16) != (signed __int128) 
(std::bfloat16_t) 42.0bf16
+      || cvt <signed __int128, std::bfloat16_t> (-42.0bf16) != (signed 
__int128) (std::bfloat16_t) -42.0bf16
+      || cvt <signed __int128, std::bfloat16_t> ((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2)) != (signed __int128) (std::bfloat16_t) ((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))
+      || cvt <signed __int128, std::bfloat16_t> (-((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2))) != (signed __int128) (std::bfloat16_t) (-((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned __int128, std::bfloat16_t> (42.0bf16) != (unsigned 
__int128) (std::bfloat16_t) 42.0bf16
+      || cvt <unsigned __int128, std::bfloat16_t> ((unsigned __int128) 1 << 
(__CHAR_BIT__ * sizeof (unsigned __int128) - 1)) != (unsigned __int128) (std::bfloat16_t) 
((unsigned __int128) 1 << (__CHAR_BIT__ * sizeof (unsigned __int128) - 1)))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT32_T__
+  if (cvt <signed char, std::float32_t> (42.0f32) != (signed char) 
(std::float32_t) 42.0f32
+      || cvt <signed char, std::float32_t> (-42.0f32) != (signed char) 
(std::float32_t) -42.0f32
+      || cvt <signed char, std::float32_t> ((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2)) != (signed char) (std::float32_t) ((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))
+      || cvt <signed char, std::float32_t> (-((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2))) != (signed char) (std::float32_t) (-((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned char, std::float32_t> (42.0f32) != (unsigned char) 
(std::float32_t) 42.0f32
+      || cvt <unsigned char, std::float32_t> ((unsigned char) 1 << (__CHAR_BIT__ * 
sizeof (unsigned char) - 1)) != (unsigned char) (std::float32_t) ((unsigned char) 1 << 
(__CHAR_BIT__ * sizeof (unsigned char) - 1)))
+    __builtin_abort ();
+  if (cvt <signed short, std::float32_t> (42.0f32) != (signed short) 
(std::float32_t) 42.0f32
+      || cvt <signed short, std::float32_t> (-42.0f32) != (signed short) 
(std::float32_t) -42.0f32
+      || cvt <signed short, std::float32_t> ((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2)) != (signed short) (std::float32_t) ((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))
+      || cvt <signed short, std::float32_t> (-((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2))) != (signed short) (std::float32_t) (-((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned short, std::float32_t> (42.0f32) != (unsigned short) 
(std::float32_t) 42.0f32
+      || cvt <unsigned short, std::float32_t> ((unsigned short) 1 << (__CHAR_BIT__ * 
sizeof (unsigned short) - 1)) != (unsigned short) (std::float32_t) ((unsigned short) 1 << 
(__CHAR_BIT__ * sizeof (unsigned short) - 1)))
+    __builtin_abort ();
+  if (cvt <signed int, std::float32_t> (42.0f32) != (signed int) 
(std::float32_t) 42.0f32
+      || cvt <signed int, std::float32_t> (-42.0f32) != (signed int) 
(std::float32_t) -42.0f32
+      || cvt <signed int, std::float32_t> ((signed int) 1 << (__CHAR_BIT__ * sizeof 
(signed int) - 2)) != (signed int) (std::float32_t) ((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))
+      || cvt <signed int, std::float32_t> (-((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))) != (signed int) (std::float32_t) (-((signed int) 1 << 
(__CHAR_BIT__ * sizeof (signed int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned int, std::float32_t> (42.0f32) != (unsigned int) 
(std::float32_t) 42.0f32
+      || cvt <unsigned int, std::float32_t> ((unsigned int) 1 << (__CHAR_BIT__ * 
sizeof (unsigned int) - 1)) != (unsigned int) (std::float32_t) ((unsigned int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long int, std::float32_t> (42.0f32) != (signed long int) 
(std::float32_t) 42.0f32
+      || cvt <signed long int, std::float32_t> (-42.0f32) != (signed long int) 
(std::float32_t) -42.0f32
+      || cvt <signed long int, std::float32_t> ((signed long int) 1 << (__CHAR_BIT__ 
* sizeof (signed long int) - 2)) != (signed long int) (std::float32_t) ((signed long int) 1 
<< (__CHAR_BIT__ * sizeof (signed long int) - 2))
+      || cvt <signed long int, std::float32_t> (-((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2))) != (signed long int) (std::float32_t) (-((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long int, std::float32_t> (42.0f32) != (unsigned long int) 
(std::float32_t) 42.0f32
+      || cvt <unsigned long int, std::float32_t> ((unsigned long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long int) - 1)) != (unsigned long int) (std::float32_t) 
((unsigned long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long long int, std::float32_t> (42.0f32) != (signed long 
long int) (std::float32_t) 42.0f32
+      || cvt <signed long long int, std::float32_t> (-42.0f32) != (signed long 
long int) (std::float32_t) -42.0f32
+      || cvt <signed long long int, std::float32_t> ((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2)) != (signed long long int) (std::float32_t) 
((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))
+      || cvt <signed long long int, std::float32_t> (-((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2))) != (signed long long int) (std::float32_t) 
(-((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long long int, std::float32_t> (42.0f32) != (unsigned long 
long int) (std::float32_t) 42.0f32
+      || cvt <unsigned long long int, std::float32_t> ((unsigned long long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long long int) - 1)) != (unsigned long long int) 
(std::float32_t) ((unsigned long long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long long 
int) - 1)))
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <signed __int128, std::float32_t> (42.0f32) != (signed __int128) 
(std::float32_t) 42.0f32
+      || cvt <signed __int128, std::float32_t> (-42.0f32) != (signed __int128) 
(std::float32_t) -42.0f32
+      || cvt <signed __int128, std::float32_t> ((signed __int128) 1 << (__CHAR_BIT__ 
* sizeof (signed __int128) - 2)) != (signed __int128) (std::float32_t) ((signed __int128) 1 
<< (__CHAR_BIT__ * sizeof (signed __int128) - 2))
+      || cvt <signed __int128, std::float32_t> (-((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2))) != (signed __int128) (std::float32_t) (-((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned __int128, std::float32_t> (42.0f32) != (unsigned __int128) 
(std::float32_t) 42.0f32
+      || cvt <unsigned __int128, std::float32_t> ((unsigned __int128) 1 << 
(__CHAR_BIT__ * sizeof (unsigned __int128) - 1)) != (unsigned __int128) (std::float32_t) 
((unsigned __int128) 1 << (__CHAR_BIT__ * sizeof (unsigned __int128) - 1)))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT64_T__
+  if (cvt <signed char, std::float64_t> (42.0f64) != (signed char) 
(std::float64_t) 42.0f64
+      || cvt <signed char, std::float64_t> (-42.0f64) != (signed char) 
(std::float64_t) -42.0f64
+      || cvt <signed char, std::float64_t> ((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2)) != (signed char) (std::float64_t) ((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))
+      || cvt <signed char, std::float64_t> (-((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2))) != (signed char) (std::float64_t) (-((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned char, std::float64_t> (42.0f64) != (unsigned char) 
(std::float64_t) 42.0f64
+      || cvt <unsigned char, std::float64_t> ((unsigned char) 1 << (__CHAR_BIT__ * 
sizeof (unsigned char) - 1)) != (unsigned char) (std::float64_t) ((unsigned char) 1 << 
(__CHAR_BIT__ * sizeof (unsigned char) - 1)))
+    __builtin_abort ();
+  if (cvt <signed short, std::float64_t> (42.0f64) != (signed short) 
(std::float64_t) 42.0f64
+      || cvt <signed short, std::float64_t> (-42.0f64) != (signed short) 
(std::float64_t) -42.0f64
+      || cvt <signed short, std::float64_t> ((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2)) != (signed short) (std::float64_t) ((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))
+      || cvt <signed short, std::float64_t> (-((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2))) != (signed short) (std::float64_t) (-((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned short, std::float64_t> (42.0f64) != (unsigned short) 
(std::float64_t) 42.0f64
+      || cvt <unsigned short, std::float64_t> ((unsigned short) 1 << (__CHAR_BIT__ * 
sizeof (unsigned short) - 1)) != (unsigned short) (std::float64_t) ((unsigned short) 1 << 
(__CHAR_BIT__ * sizeof (unsigned short) - 1)))
+    __builtin_abort ();
+  if (cvt <signed int, std::float64_t> (42.0f64) != (signed int) 
(std::float64_t) 42.0f64
+      || cvt <signed int, std::float64_t> (-42.0f64) != (signed int) 
(std::float64_t) -42.0f64
+      || cvt <signed int, std::float64_t> ((signed int) 1 << (__CHAR_BIT__ * sizeof 
(signed int) - 2)) != (signed int) (std::float64_t) ((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))
+      || cvt <signed int, std::float64_t> (-((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))) != (signed int) (std::float64_t) (-((signed int) 1 << 
(__CHAR_BIT__ * sizeof (signed int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned int, std::float64_t> (42.0f64) != (unsigned int) 
(std::float64_t) 42.0f64
+      || cvt <unsigned int, std::float64_t> ((unsigned int) 1 << (__CHAR_BIT__ * 
sizeof (unsigned int) - 1)) != (unsigned int) (std::float64_t) ((unsigned int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long int, std::float64_t> (42.0f64) != (signed long int) 
(std::float64_t) 42.0f64
+      || cvt <signed long int, std::float64_t> (-42.0f64) != (signed long int) 
(std::float64_t) -42.0f64
+      || cvt <signed long int, std::float64_t> ((signed long int) 1 << (__CHAR_BIT__ 
* sizeof (signed long int) - 2)) != (signed long int) (std::float64_t) ((signed long int) 1 
<< (__CHAR_BIT__ * sizeof (signed long int) - 2))
+      || cvt <signed long int, std::float64_t> (-((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2))) != (signed long int) (std::float64_t) (-((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long int, std::float64_t> (42.0f64) != (unsigned long int) 
(std::float64_t) 42.0f64
+      || cvt <unsigned long int, std::float64_t> ((unsigned long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long int) - 1)) != (unsigned long int) (std::float64_t) 
((unsigned long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long long int, std::float64_t> (42.0f64) != (signed long 
long int) (std::float64_t) 42.0f64
+      || cvt <signed long long int, std::float64_t> (-42.0f64) != (signed long 
long int) (std::float64_t) -42.0f64
+      || cvt <signed long long int, std::float64_t> ((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2)) != (signed long long int) (std::float64_t) 
((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))
+      || cvt <signed long long int, std::float64_t> (-((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2))) != (signed long long int) (std::float64_t) 
(-((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long long int, std::float64_t> (42.0f64) != (unsigned long 
long int) (std::float64_t) 42.0f64
+      || cvt <unsigned long long int, std::float64_t> ((unsigned long long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long long int) - 1)) != (unsigned long long int) 
(std::float64_t) ((unsigned long long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long long 
int) - 1)))
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <signed __int128, std::float64_t> (42.0f64) != (signed __int128) 
(std::float64_t) 42.0f64
+      || cvt <signed __int128, std::float64_t> (-42.0f64) != (signed __int128) 
(std::float64_t) -42.0f64
+      || cvt <signed __int128, std::float64_t> ((signed __int128) 1 << (__CHAR_BIT__ 
* sizeof (signed __int128) - 2)) != (signed __int128) (std::float64_t) ((signed __int128) 1 
<< (__CHAR_BIT__ * sizeof (signed __int128) - 2))
+      || cvt <signed __int128, std::float64_t> (-((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2))) != (signed __int128) (std::float64_t) (-((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned __int128, std::float64_t> (42.0f64) != (unsigned __int128) 
(std::float64_t) 42.0f64
+      || cvt <unsigned __int128, std::float64_t> ((unsigned __int128) 1 << 
(__CHAR_BIT__ * sizeof (unsigned __int128) - 1)) != (unsigned __int128) (std::float64_t) 
((unsigned __int128) 1 << (__CHAR_BIT__ * sizeof (unsigned __int128) - 1)))
+    __builtin_abort ();
+#endif
+#endif
+#ifdef __STDCPP_FLOAT128_T__
+  if (cvt <signed char, std::float128_t> (42.0f128) != (signed char) 
(std::float128_t) 42.0f128
+      || cvt <signed char, std::float128_t> (-42.0f128) != (signed char) 
(std::float128_t) -42.0f128
+      || cvt <signed char, std::float128_t> ((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2)) != (signed char) (std::float128_t) ((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))
+      || cvt <signed char, std::float128_t> (-((signed char) 1 << (__CHAR_BIT__ * 
sizeof (signed char) - 2))) != (signed char) (std::float128_t) (-((signed char) 1 << 
(__CHAR_BIT__ * sizeof (signed char) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned char, std::float128_t> (42.0f128) != (unsigned char) 
(std::float128_t) 42.0f128
+      || cvt <unsigned char, std::float128_t> ((unsigned char) 1 << (__CHAR_BIT__ * 
sizeof (unsigned char) - 1)) != (unsigned char) (std::float128_t) ((unsigned char) 1 << 
(__CHAR_BIT__ * sizeof (unsigned char) - 1)))
+    __builtin_abort ();
+  if (cvt <signed short, std::float128_t> (42.0f128) != (signed short) 
(std::float128_t) 42.0f128
+      || cvt <signed short, std::float128_t> (-42.0f128) != (signed short) 
(std::float128_t) -42.0f128
+      || cvt <signed short, std::float128_t> ((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2)) != (signed short) (std::float128_t) ((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))
+      || cvt <signed short, std::float128_t> (-((signed short) 1 << (__CHAR_BIT__ * 
sizeof (signed short) - 2))) != (signed short) (std::float128_t) (-((signed short) 1 << 
(__CHAR_BIT__ * sizeof (signed short) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned short, std::float128_t> (42.0f128) != (unsigned short) 
(std::float128_t) 42.0f128
+      || cvt <unsigned short, std::float128_t> ((unsigned short) 1 << (__CHAR_BIT__ 
* sizeof (unsigned short) - 1)) != (unsigned short) (std::float128_t) ((unsigned short) 1 
<< (__CHAR_BIT__ * sizeof (unsigned short) - 1)))
+    __builtin_abort ();
+  if (cvt <signed int, std::float128_t> (42.0f128) != (signed int) 
(std::float128_t) 42.0f128
+      || cvt <signed int, std::float128_t> (-42.0f128) != (signed int) 
(std::float128_t) -42.0f128
+      || cvt <signed int, std::float128_t> ((signed int) 1 << (__CHAR_BIT__ * sizeof 
(signed int) - 2)) != (signed int) (std::float128_t) ((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))
+      || cvt <signed int, std::float128_t> (-((signed int) 1 << (__CHAR_BIT__ * 
sizeof (signed int) - 2))) != (signed int) (std::float128_t) (-((signed int) 1 << 
(__CHAR_BIT__ * sizeof (signed int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned int, std::float128_t> (42.0f128) != (unsigned int) 
(std::float128_t) 42.0f128
+      || cvt <unsigned int, std::float128_t> ((unsigned int) 1 << (__CHAR_BIT__ * 
sizeof (unsigned int) - 1)) != (unsigned int) (std::float128_t) ((unsigned int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long int, std::float128_t> (42.0f128) != (signed long int) 
(std::float128_t) 42.0f128
+      || cvt <signed long int, std::float128_t> (-42.0f128) != (signed long 
int) (std::float128_t) -42.0f128
+      || cvt <signed long int, std::float128_t> ((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2)) != (signed long int) (std::float128_t) ((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))
+      || cvt <signed long int, std::float128_t> (-((signed long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long int) - 2))) != (signed long int) (std::float128_t) (-((signed 
long int) 1 << (__CHAR_BIT__ * sizeof (signed long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long int, std::float128_t> (42.0f128) != (unsigned long 
int) (std::float128_t) 42.0f128
+      || cvt <unsigned long int, std::float128_t> ((unsigned long int) 1 << 
(__CHAR_BIT__ * sizeof (unsigned long int) - 1)) != (unsigned long int) (std::float128_t) 
((unsigned long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long int) - 1)))
+    __builtin_abort ();
+  if (cvt <signed long long int, std::float128_t> (42.0f128) != (signed long 
long int) (std::float128_t) 42.0f128
+      || cvt <signed long long int, std::float128_t> (-42.0f128) != (signed 
long long int) (std::float128_t) -42.0f128
+      || cvt <signed long long int, std::float128_t> ((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2)) != (signed long long int) (std::float128_t) 
((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))
+      || cvt <signed long long int, std::float128_t> (-((signed long long int) 1 << 
(__CHAR_BIT__ * sizeof (signed long long int) - 2))) != (signed long long int) (std::float128_t) 
(-((signed long long int) 1 << (__CHAR_BIT__ * sizeof (signed long long int) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned long long int, std::float128_t> (42.0f128) != (unsigned 
long long int) (std::float128_t) 42.0f128
+      || cvt <unsigned long long int, std::float128_t> ((unsigned long long int) 1 
<< (__CHAR_BIT__ * sizeof (unsigned long long int) - 1)) != (unsigned long long int) 
(std::float128_t) ((unsigned long long int) 1 << (__CHAR_BIT__ * sizeof (unsigned long long 
int) - 1)))
+    __builtin_abort ();
+#ifdef __SIZEOF_INT128__
+  if (cvt <signed __int128, std::float128_t> (42.0f128) != (signed __int128) 
(std::float128_t) 42.0f128
+      || cvt <signed __int128, std::float128_t> (-42.0f128) != (signed 
__int128) (std::float128_t) -42.0f128
+      || cvt <signed __int128, std::float128_t> ((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2)) != (signed __int128) (std::float128_t) ((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))
+      || cvt <signed __int128, std::float128_t> (-((signed __int128) 1 << 
(__CHAR_BIT__ * sizeof (signed __int128) - 2))) != (signed __int128) (std::float128_t) (-((signed 
__int128) 1 << (__CHAR_BIT__ * sizeof (signed __int128) - 2))))
+    __builtin_abort ();
+  if (cvt <unsigned __int128, std::float128_t> (42.0f128) != (unsigned 
__int128) (std::float128_t) 42.0f128
+      || cvt <unsigned __int128, std::float128_t> ((unsigned __int128) 1 << 
(__CHAR_BIT__ * sizeof (unsigned __int128) - 1)) != (unsigned __int128) (std::float128_t) 
((unsigned __int128) 1 << (__CHAR_BIT__ * sizeof (unsigned __int128) - 1)))
+    __builtin_abort ();
+#endif
+#endif
+#endif
+#endif
+}

        Jakub