Re: too many warnings from Bison CVS for Pike

[Hans Aberg]

On 30 Jan 2006, at 21:34, Joel E. Denny wrote:

> Thanks for that information. I'd be interested to know what C++  
> features
> are no longer supported.

It's easier to answer what is supported: none, as far as compile C as  
C++. There is some C++ skeleton, using C++ containers, in the work by  
Akim Demaille, and some may already using it.

> Fortunately for my work, support for compiling C as C++ has not been
> dropped entirely, or perhaps it's just been reinstated  
> unofficially.  For
> example, make maintainer-check runs all test cases with a C++  
> compiler.
> Also, in recent months, I've seen effort to cast the return of malloc
> because C++ doesn't allow assigning a void* to a non-void* type.   
> On the
> other hand, this cast is actually considered by some to be bad  
> style in C:
> if you forget to #include the prototype for malloc, the cast  
> prevents the
> compiler from warning you that the assumed int return type of  
> malloc is
> unreasonable.

It is Paul Eggert that is doing the C skeleton, and I doubt he does  
anything with C++ in mind. Note that the C parser stack does not  
invoke copy constructors when extending. And a strict C++ compiler  
will complain when using names such as "malloc", as the C++ name is  
"std::malloc". Is your C++ compiler old?

  Hans Aberg