Voices 2-4 not padded with rests when lengthening a measure
Ubuntu 14.04, GIT commit: f1355b1
1) empty 4/4 score
2) enter a quarter note into any measure, both in voices 1 & 2
3) measure properties
4) change actual duration to 6/4
5) OK
Result: voice 1 is padded wikth rests to add up to 6 beats, but voice 2 is not. It probably should be.
Comments
In voices 2 to 4, it is possible to delete 'trailing' rests, leaving the end of the measure unfilled. I have no idea if this is a 'feature' or not, but it is there.
What you describe is on the same line, I believe. So, I'm not sure voices 2 to 4 need to be padded.
(Of course, assuming this does not preclude adding stuff in the unfilled portion, should the need arise).
M.
Yes, it is related to the idea that rests can be deleted, and I still think we eventually need a better story there too. The problem is that once the rests are gone, you can't get them back except by exchanging the voice with voice 1 (and then back). So while it *possible* to create these "holes", they aren't good. In the case of a user deleting a rest, we could simply tell people "don't do that" - mark them invisible unstead. And besides, if they delete, they presumably don't *want* to enter anything there (unless they deleted the rest by mistake).
Whereas in the case of lengthening a measure, there is no way to do this *without* creating "holes". And worse, the whole reason you lengthened the measure is probably to add more notes - and now that is impossible, except for voice 1, or if you know the exchange trick.
So while "need" is a strong word, I do think we *should* pad any non-empty voices when lengthening a measure, even though it is indeed possible to delete rests and create similar "holes".
Ok, I never tried deleting those trailing rests. This being the case, I agree with you:
1) Fill all the voices while lengthening a measure and
2) do not allow trailing rests deletion.
M.