Re: [PATCH] %define boolean variables

[Joel E. Denny]

On Mon, 29 Jan 2007, Paolo Bonzini wrote:

> > As a compromise, if you keep the current behavior of empty being true, and
> > `%define FOO' aliasing `%define FOO ""', it's ok for me to complain on
> > anything but empty/"true"/"false", including making "0" or "1" invalid.
> 
> Here's my concrete proposal.

Thanks.

>                        [m4_fatal([invalid %define variable passed to 
> b4_percent_define_flag_if: ]$1)])])

Shouldn't skeleton authors also benefit from translation?  That is, 
shouldn't this be b4_fatal?  Just a thought.

>  # b4_public_if(TRUE, FALSE)
>  # -------------------------
> -b4_percent_define_default([public], 0)
> +b4_percent_define_default([public], [false])
>  m4_define([b4_public_if],
>  [b4_percent_define_flag_if([public], [$1], [$2])])

[public] and [false] are underquoted.  Imagine if a skeleton author 
m4_define's public or false for some reason.  Maybe they shouldn't, but we 
should be robust, I think.
 
>  # b4_single_class_if(TRUE, FALSE)
>  # -------------------------------
> -b4_percent_define_default([single_class], 0)
> +b4_percent_define_default([single_class], [false])
>  m4_define([b4_single_class_if],
>  [b4_percent_define_flag_if([single_class], [$1], [$2])])

Likewise.
 
> Index: doc/bison.texinfo
> ===================================================================
> RCS file: /sources/bison/bison/doc/bison.texinfo,v
> retrieving revision 1.225
> diff -u -r1.225 bison.texinfo
> --- doc/bison.texinfo 29 Jan 2007 10:54:42 -0000      1.225
> +++ doc/bison.texinfo 29 Jan 2007 12:15:59 -0000
> @@ -4696,10 +4696,10 @@
>  Define a variable to adjust Bison's behavior.
>  The list of available variables and their meanings depends on the selected

"variables" should probably be "DEFINE-VARIABLEs" especially since the 
remaining text now does that.  My mistake.  Would you fix this for me 
since you're already working on this?

Actually, isn't "DEFINE-VARIABLE" a little redundant here?  Maybe just 
"VARIABLE"?  Just a thought, but I don't care much.

More importantly, we need quotes in the synopsis here.  So, I'm
thinking:

  %define VARIABLE
  %define VARIABLE "VALUE"

> -The @var{value} can be omitted for boolean variables; for
> -boolean variables, the skeletons will treat a @var{value} of @samp{0}
> -or @samp{false} as the boolean variable being false, and anything else
> -as true.
> +Some @var{define-variable}s may be used as boolean values: in this case,
> +the skeleton will conventionally treat a @var{value} of @samp{false}
> +as the boolean variable being false; a value of @samp{true}, or @var{value}

`a value of' --> `a @var{value} of'

> +%define public "maybe"

:-)