Avoid hidden <climits> dependency in <thread>

[joerg]

Summary

sleep_for tries to convert arbitrary durations toward nanosecond resolution in a way that avoids truncation. The nanosecond timer is declared using long long, but the minimum required size of 64bit signed is enough for all computation needs. Avoid doing a long double constexpr computation here as older GCC versions on PowerPC have problems folding the double-double type and also avoid the additional header include at the same time. The constant equals about 290 years and should be good enough.

libcxx/include/thread

 
 template <class _Rep, class _Period>
 void
 sleep_for(const chrono::duration<_Rep, _Period>& __d)
 {
     using namespace chrono;
     if (__d > duration<_Rep, _Period>::zero())
     {
-#if defined(_LIBCPP_COMPILER_GCC) && (__powerpc__ || __POWERPC__)
-    //  GCC's long double const folding is incomplete for IBM128 long doubles.
-        _LIBCPP_CONSTEXPR duration<long double> _Max = nanoseconds::max();
-#else
-        _LIBCPP_CONSTEXPR duration<long double> _Max = duration<long double>(ULLONG_MAX/1000000000ULL) ;
-#endif
+        // The standard guarantees a 64bit signed integer resolution for nanoseconds,
+        // so use INT64_MAX / 1e9 as cut-off point. Use a constant to avoid <climits>
+        // and issues with long double folding on PowerPC with GCC.
+        _LIBCPP_CONSTEXPR duration<long double> _Max = duration<long double>(9223372036.0L) ;
         nanoseconds __ns;
         if (__d < _Max)
         {
             __ns = duration_cast<nanoseconds>(__d);
             if (__ns < __d)
                 ++__ns;
         }
         else