Re: [boost] [proto] design of expr<>

From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2008-03-10 12:18:55

AMDG

Markus Werle wrote:
> Steven Watanabe <watanabesj <at> gmail.com> writes:
>
>   
>> Markus Werle wrote:
>> [...]
>>     
>>> OTOH the third template argument contains redundant information: 
>>> The docs say: "Proto expression nodes are valid Fusion random-access 
>>> sequences of their children nodes."
>>> So the information about the number of arguments is accessible via
>>> some size<T>::type metafunction e.g. depending on
>>>
>>> result_of::distance<result_of::begin<S>::type
>>> 		result_of::end<S>::type>::type?
>>>
>>>       
>> [...]
>> I believe that it is needed to allow partial specialization.
>> expr<Tag, Args, 0>, expr<Tag, Args, 1>, ...
>>     
>
> Sorry if I do not get that one.
>   

The definition of expr is something like:

template<class Tag, class Args, int Arity = Args::arity>
struct expr;

template<class Tag, class Args>
struct expr<Tag, Args, 0> {
   typename Args::arg0 arg0;
};
template<class Tag, class Args>
struct expr<Tag, Args, 1> {
   typename Args::arg0 arg0;
};
template<class Tag, class Args>
struct expr<Tag, Args, 2> {
   typename Args::arg0 arg0;
   typename Args::arg1 arg1;
};

...

>> And users don't need to care about it, since it has a default.
>>     
>
> "User" is a very ambiguous word for proto.
> Actually most "users" of proto will be "library writers" 
> seeking to boost (sic!) their productivity.
> So even the nitty-gritty details matter. 
> I reject any "do-not-mind-it's-only-deep-inside" arguments
> here. Everybody writing a compile time algorithm acting on
> proto structures for whatever reason will be affected by this
> "detail" and the implications might only become clear later,
> when somebody is in need to do some compile-time-magic.
>
> >From Daixtrose I know that the number of template arguments
> can be reciprocal to readability, maintainability and extensibility.
>   

You can (and should) write expr<terminal_tag, args0<int> > rather
than expr<terminal_tag, args0<int>, 0>.  Library writers are not
somehow special in this respect.  When I said "user" I meant anyone who
is not the author or maintainer of proto.The only problems that I can
see are that mpl::apply<expr<mpl::_, mpl::_>, terminal_tag, args0<int> >
won't work and that mpl::quote2<expr> will fail. 

In Christ,
Steven Watanabe