Re: C99 in Bison

[Akim Demaille]

Le 28 août 09 à 18:55, Joel E. Denny a écrit :

> Hi Akim.

Hi Joel,

> On Fri, 28 Aug 2009, Akim Demaille wrote:
>
> > My quotation does not make it clear, but yysyntax_error is always  
> > defined (and
> > used, even for "simple" messages), and always with this signature.   
> > So that
> > one is really wrong.
>
> yysyntax_error is disabled if YYERROR_VERBOSE is off:
>
>  % echo '%% start: ;' > tmp.y
>  % ./tests/bison tmp.y
>  % gcc -E tmp.tab.c | grep yysyntax_error

Bummer.  Thanks.  I was looking at lalr1.cc's, which has the  
ERROR_VERBOSE dependency inside it, instead of leaving it on the call  
site.


> That gives no output unless I add %error-verbose to the grammar.
>
> > so I guess we can afford dropping knr in the generated skeletons,  
> > without
> > polling in the NEWS (but we can inform in the NEWS of 2.4.2 that  
> > 2.5 will
> > completely drop KnR support in yacc.c, which was accidentally  
> > broken anyway
> > since Bison 2.1).  After all, old tarballs are still there.
>
> I'd be willing to take that approach anyway.  Like you said, if  
> someone
> needs K&R, old Bisons are lying around:
>
>  % aptitude search -F %p bison | grep -v ++
>  bison
>  bison-1.35
>  bison-doc
>
> Alternatively, we could poll at help-bison and might get an answer  
> faster.

Yes, but what's the size of the poll?

> > As for C90 in bison itself, I'm fine with it.
>
> You mean C99, right?

Yes, sorry.

> > Yet C++03 would be better ;)
>
> Why don't we discuss that again when at least 2.5 is released?  Some
> cherry picks are already painful.

Well, I would also like to release 2.6 not too late, so I would wait  
for 2.6 to be released first.

> Then again, I believe C++03 doesn't support all C99 constructs, such  
> as
> variable length arrays.

It has std::vector.

>  So, if we're not careful about our C99 usage, we
> might be making more work for whenever we switch to C++03.

I'm not sure it would require a lot of work.