[cfe-commits] [libcxx] r132129 - in /libcxx/trunk: include/exception include/initializer_list include/limits include/new include/typeinfo src/exception.cpp src/new.cpp src/typeinfo.cpp

Jonathan Sauer jonathan.sauer at gmx.de
Thu May 26 11:42:48 PDT 2011


Hello,

>     // access functions
> -    void rethrow_nested [[noreturn]] () const;
> -    exception_ptr nested_ptr() const;
> +    [[noreturn]] void rethrow_nested() const;
> +    exception_ptr nested_ptr() const noexcept;
> };
> 
> -template <class T> void throw_with_nested [[noreturn]] (T&& t);
> +template <class T> [[noreturn]] void throw_with_nested(T&& t);

Is there a particular reason why you are using "[[noreturn]]" instead of "_ATTRIBUTE(noreturn)", as you
do below with <unexpected> and <terminate>? Especially since the latter supports compilation with
compilers that don't support attributes (yet).

Although I don't know if these compilers are relevant anymore.

Also, above you are using "noexcept", whereas below sometimes "_NOEXCEPT" is used? This as well as the
above seems inconsistent to me.

> [...]
> typedef void (*unexpected_handler)();
> -_LIBCPP_VISIBLE unexpected_handler set_unexpected(unexpected_handler) throw();
> -_LIBCPP_VISIBLE unexpected_handler get_unexpected() throw();
> +_LIBCPP_VISIBLE unexpected_handler set_unexpected(unexpected_handler) _NOEXCEPT;
> +_LIBCPP_VISIBLE unexpected_handler get_unexpected() _NOEXCEPT;
> _ATTRIBUTE(noreturn) _LIBCPP_VISIBLE void unexpected();
> [...]
> -    nested_exception();
> -//     nested_exception(const nested_exception&) throw() = default;
> -//     nested_exception& operator=(const nested_exception&) throw() = default;

In __config, _LIBCPP_HAS_NO_DEFAULTED_FUNCTIONS is always defined. Shouldn't the "= default" only be used if
defaulted functions are actually supported by the compiler?

Also in __config, there is (lines 94ff):

| //#if !__has_feature(cxx_alias_templates)
| #define _LIBCPP_HAS_NO_TEMPLATE_ALIASES
| //#endif

even though the check for feature "cxx_alias_templates" is now supported.


Jonathan





More information about the cfe-commits mailing list