Re: [PATCH] overquoting in implementation of %code and %defines

[Joel E. Denny]

On Mon, 15 Jan 2007, Paolo Bonzini wrote:

> 
> > > So another way to break the tie and not add the second parameter would be
> > > to
> > > change the default macro name to "b4_get_percent_define(NAME)" and have a
> > > "b4_define_default(NAME, FOO)" macro that defines this special-named
> > > macro.
> > 
> > Ok, how about this?
> > 
> >   b4_default_percent_define([[foo]], [[default value]])
> 
> Fine by me.

On second look, I think I misread what you proposed.  Let's make sure 
we're clear.  I'm proposing that:

1. b4_percent_define_foo is the value of foo.  However, skeleton authors 
shouldn't need to touch this.

2. b4_get_percent_define([[foo]]) retrieves the value of 
b4_percent_define_foo and records that it's used in the skeleton.

3. b4_default_percent_define([[foo]], [[default value]]) sets the value of 
b4_percent_define_foo if it's not already set.

Does that make sense?

We may eventually want a b4_is_defined_percent_define([[foo]]).  Both this 
and b4_default_percent_define can be handled by examining 
b4_percent_define_foo directly.  I wonder if this sort of data 
encapsulation is worthwhile in M4....  What do you think?