SoundsFromSound <soundsfromso...@gmail.com> writes: > Hi David, > > I'm a little confused. What problems was David talking about, I didn't > follow. So we shouldn't use this boxed notation approach, or we should but > with caveats?
One shouldn't a) tamper with internals b) expect the tampering to remain working in future versions Mike took some code that was _not_ made a public interface because it was not able to work according to the expectations of a public interface (namely that using it in one user-level file will not affect other, independent files), copied it and made a public interface from it. Now of course this does not work reliably in a multi-file scenario, and since it does not work reliably, whenever there _will_ be a public interface for that kind of thing, it will _necessarily_ have different interfaces. So the approach was broken, and it was _guaranteed_ to stop working the moment the underlying implementation was changed to something less broken. -- David Kastrup _______________________________________________ lilypond-user mailing list lilypond-user@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-user