[Bug c++/108243] Missed optimization for static const std::string_view(const char*)

[redi at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108243

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
This is caused by a change in libstdc++ headers:

@@ -230,9 +230,15 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
     static _GLIBCXX_ALWAYS_INLINE constexpr bool
     __constant_string_p(const _CharT* __s)
     {
+#ifdef _GLIBCXX_HAVE_BUILTIN_IS_CONSTANT_EVALUATED
+      (void) __s;
+      // In constexpr contexts all strings should be constant.
+      return __builtin_is_constant_evaluated();
+#else
       while (__builtin_constant_p(*__s) && *__s)
        __s++;
       return __builtin_constant_p(*__s);
+#endif
     }

   /**

That affects the definition of std::char_traits<char>::length:

      static constexpr size_t
      length(const char_type* __s)
      {
        if (__constant_string_p(__s))
          return __gnu_cxx::char_traits<char_type>::length(__s);

        return __builtin_strlen(__s);
      }

So since GCC 9.1.0 we only treat the length as a constant expression if the
string_view object is constexpr, otherwise we call strlen, which requires the
object to exist.

I think the compiler _should_ be able to optimize this case anyway, but maybe
we need to partially revert the libstdc++ changes.