Re: experimental features in Bison 2.3a+

[Joel E. Denny]

On Fri, 1 Dec 2006, Paul Eggert wrote:

> "Joel E. Denny" <address@hidden> writes:
> 
> > It seems to me that the following features ought to be documented as 
> > experimental since they have been somewhat controversial even among Bison 
> > developers:
> >
> > 1. Default %destructor's and %printer's.  These are currently declared 
> > with <*> and <>.
> >
> > 2. The prologue alternatives: %code, %requires, %provides, and %code-top.
> 
> Yes.  Though to be honest I can't even remember what all of them are
> for....

I just remembered that %code and %requires are used in the C++ examples in 
the documentation.  Without a rewrite of those examples, %requires is 
necessary because there's no other directive that inserts user code into 
the parser header file.

I'm not going to rewrite the C++ examples.  Is it still ok to mark the 
prologue alternatives as experimental?

Also, I'm thinking of commenting out any mention of Java from the 
documentation on the prologue alternatives.  I doubt the Java skeleton 
will be complete in the next release, and we might forget to remove the 
documentation before then.  Agreed?