Tie LilyPond, lexer and parser together more type-safely.

Tie LilyPond, lexer and parser together more type-safely.

Obviously, this would be followed by a separate commit that replaces
"PARSER" in parser.yy with "parser" and removes the respective #define in order to keep
changes in this manually created commit small and reviewable.

[Carl, 2011-12-21 17:18:36]

Looks excellent to me.  Thanks!

[Reinhold, 2011-12-21 17:22:44]

LGTM (looks very much like a code simplification)

[hanwenn, 2011-12-21 21:48:12]

LGTM.

Does this require a required version bump in configure for flex/bison?

On Wed, Dec 21, 2011 at 3:18 PM,  <Carl.D.Sorensen@gmail.com> wrote:
> Looks excellent to me.  Thanks!
>
> http://codereview.appspot.com/5501056/
>
> _______________________________________________
> lilypond-devel mailing list
> lilypond-devel@gnu.org
> https://lists.gnu.org/mailman/listinfo/lilypond-devel



-- 
Han-Wen Nienhuys - hanwen@xs4all.nl - http://www.xs4all.nl/~hanwen

[dak, 2011-12-21 22:29:20]

Han-Wen Nienhuys <hanwenn@gmail.com> writes:

> LGTM.
>
> Does this require a required version bump in configure for flex/bison?

Good question.

Looking at /usr/share/doc/bison/changelog.gz, I see for example

2003-04-28  Tim Van Holder <tim.van.holder@pandora.be>

        * data/yacc.c: (b4_lex_param): Corrected for the case where
        %lex-param is provided and %pure-parser isn't.

That makes it likely that we won't be hearing lots of hollering from
users if the version is not entirely accurate in configure.  I would
guess that this is about the time from which on the %lex-param and
%parse-param have been usable in the current form.

Flex should not actually be involved in that change.

-- 
David Kastrup

[dak, 2011-12-22 09:10:27]

On 2011/12/21 22:29:20, dak wrote:
> Han-Wen Nienhuys <mailto:hanwenn@gmail.com> writes:
> 
> > Does this require a required version bump in configure for flex/bison?
> 
> Good question.
> 
> Looking at /usr/share/doc/bison/changelog.gz, I see for example
> 
> 2003-04-28  Tim Van Holder <mailto:tim.van.holder@pandora.be>
> 
>         * data/yacc.c: (b4_lex_param): Corrected for the case where
>         %lex-param is provided and %pure-parser isn't.
> 
> That makes it likely that we won't be hearing lots of hollering from
> users if the version is not entirely accurate in configure.

In case that has not been obvious: I think it is not trivial to acquire accurate
version information and I don't expect to see problems in practice.  If and when
we get a version-related report, we can use that information for updating the
required versions, but I don't think it's worth actively hunting for it.