Declarations that are impossible without typedef

[Peter]

I read somewhere that typedef keyword is not merely
a syntactic sugar, but there are, in fact,
valid declarations which would be impossible to express
without using typedef due to language syntax constraints.
If that's true, could you give me an example of
such declaration (the simpler, the better)?

 

[Öö Tiib]

I read somewhere that typedef keyword is not merely
a syntactic sugar, but there are, in fact,
valid declarations which would be impossible to express
without using typedef due to language syntax constraints.

Yes you can't do much template metaprogramming without
typedef. I think one should avoid template metaprogramming
regardless if it is clear how to use typedef there or not.

Better stick to using typedef for convenient aliases for
long types (involving mix of arrays, templates and/or
functions).

If that's true, could you give me an example of
such declaration (the simpler, the better)?

Even simplest meta-manipulation of types without typedef
is impossible. For example how you would write things
like 'std::add_const' or 'std::remove_reference' without
it? These can look something like (I didn't test it):

template<typename T>
struct remove_pointer
{
typedef T type;
};

template<typename T>
struct remove_pointer<T*>
{
typedef T type;
};

 

[Stefan Ram]

how you would write things
like 'std::add_const' or 'std::remove_reference' without

[typdef]

using type = T;

 

[Öö Tiib]

how you would write things
like 'std::add_const' or 'std::remove_reference' without

[typdef]

using type = T;

Correct but that does not matter.

(7.1.3.2): A typedef-name can also be introduced by an alias-declaration.
The identifier following the using keyword becomes a typedef-name and
the optional attribute-specifier-seq following the identifier appertains
to that typedef-name. It has the same semantics as if it were introduced
by the typedef specifier.

So that particular form of 'using' is just a syntactic sugar defined in
terms of typedef.

 

[Marc]

I read somewhere that typedef keyword is not merely
a syntactic sugar, but there are, in fact,
valid declarations which would be impossible to express
without using typedef due to language syntax constraints.
If that's true, could you give me an example of
such declaration (the simpler, the better)?

You won't be able to test this one with most compilers, but try writing
a function with C++ linkage that takes as argument a pointer to a
function with C linkage.