Sigh. Dan has suggested I might want to provide feedback here. Now I'm not really much into the Midi code to be honest, so this more or less boils down to consulting a gut feeling and/or general policy.
The main thing this boils down to here is that Heikki feels that Dan changed one occurence of code and not a parallel one, and Dan is not really sure whether the second code path should even get encountered. I'm with Heikki here (if I understand the argument correctly) that it does not seem to make sense to leave the second code path different from the first: that way lies confusion for the next person encountering the code, and additional time will get spent analysing what may be dead code. Heikki seems more confident about what code might be dead here but Dan is currently proposing a patch. Can we get to some version of the code where the code paths supposed to be equivalent (is there agreement about that?) actually looks the same? If so, this would be a good starter for Heikki to eventually propose a cleanup that would result in a removal of the dead code or keeping it but adding a programming error. Something like that. This sounds like work either party has not bargained on doing. But it would likely be easier to get it done now that people have taken a bit of a look than at a time when the person puzzling over it has no indication about why stuff looks different. https://codereview.appspot.com/308890043/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
