Le 17 nov. 2009 à 19:00, Carl Sorensen a écrit : > David, > > I appreciate your work on this. > > However, I am *not* in favor of moving in this direction to solve the > problems you correctly identified. > > In my mind, the *last* thing we need is another opaque interface in > LilyPond, where in the markup command we don't know whether a certain > property is to be looked up in props or to be assigned a value in a > let-binding from the define-internal-markup-command macro. > > I think it's *much* better to pass default values as appended values to the > tail of props. Then we can go ahead and use a props lookup in the code.
I disagree. The patch is consistent with all other markup command definitions, and inforces a common coding style. Changing the value of `props' as you suggest is not much better: changing the value of a function argument behind the scene is not what I would call a fine design. Moreover, seeing the chain-assoc-get calls ever and ever on the very same argument calls for a simplification: this is the lisp way of things. That's why I introduced this property binding mechanism. The interface is opaque, you're right. Then it shall be better documented: please wait for a few days until I work on that. Nicolas _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org http://lists.gnu.org/mailman/listinfo/lilypond-devel
