Re: proposal: simplify prologue alternatives into %code

[Hans Aberg]

On 9 Jan 2007, at 16:10, Paolo Bonzini wrote:

> > > Not a big case of interface mismatch.
> > If you think that, you are of course free to fix it your own way.
> > You idea is to let the C++ polymorphic class implementation to  
> > emulate the implementation of a C 'union', and such ties between  
> > unrelated features in different languages does not seem wise to me.
>
> Absolutely not.  I would be in favor of allowing to specify a type  
> within %type<> (using static_cast instead of accessing a union).

This, even admitting dynamic_cast as well, may not be enough, in  
amore general setting. There are also the boost:variant's suggested  
by Akim. I do not recall their syntax.

> What I meant, is that requiring the yystype to implement operator* 
> () is not likely to put a lot of burden on somebody who has to  
> resolve the interface mismatch.

Then one may have to convert the object constructed by the Bison  
generated parser to the ones used in the rest of the program. I would  
not want to program that way, but if you do, fine with me, as long as  
I do not have to do it. :-)

  Hans Aberg