Concept-check bug ?

Gabriel Dos Reis gdr@integrable-solutions.net
Fri Jun 17 12:51:00 GMT 2005


Jonathan Wakely <cow@compsoc.man.ac.uk> writes:

| On Fri, Jun 17, 2005 at 03:12:40PM +0300, Peter Dimov wrote:
| 
| > Jonathan Wakely wrote:
| > >Hi y'all,
| > >
| > >   #define _GLIBCXX_CONCEPT_CHECKS
| > >   #include <vector>
| > >
| > >   struct A {
| > >       std::vector<A> v;
| > >   };
| > >
| > >This produces concept-check errors because A is incomplete at the
| > >point of declaring v.
| > >
| > >Is this code really illegal, or just a flaw in the concept-checks ?
| > 
| > It's illegal (17.4.3.6/2 last bullet), but useful.
| 
| Thanks, Peter (I should have known that).
| 
| Presumably the truly legal way to do it is with a (smart) pointer to the

Probably "legal", but surely overkill and obfuscatory.  

Do you really believe that

  template<class T>
    struct Vector {
        T* data;
        int size;
        // ...
    };

   struct A {
      Vector<A> v;
   };

shall be expressed as:

|     struct A {
|         std::tr1::shared_ptr<std::vector<A> > v;
| 
|         A() : v(new std::vector<A>) { }
|     };

?  

Ugh.

-- Gaby



More information about the Libstdc++ mailing list