don't merge rests for cmdJoinMeasure or cmdSplitMeasure. Keep original rest elements.
If I take two measures of quarter rests:
and join them in 2.0.3, I get
whereby it combined all 8 quarter rests into a "Measure Rest" of duration 8/4. Which I don't like (and mike320 seems to agree with me) because have "lost" information about how the space was divided up. I believe MuseScore should presume that the individual rest elements were deliberate.
Another example with notes in those measures:
The result of join measure again merges those rests together, producing something that is not nice to read:
If user actually wanted to merge those rests, then after joining the measures they can do that as a separate step by selecting those rests and deleting them. The semantics of "Join Measure" shouldn't mean anything more than join measure.
(Note, I'm not quite sure best thing to do when joining two measures containing only measure rests. I find the 2.0.3 behavior of merging them into a larger measure rest confusing because the result is a measure rest for a measure whose actual duration is different than it's time signature, so just from reading the ink, user would not know what .)
Similarly for cmdSplitMeasure, if take measure of quarter rests:
And then I split in the middle, the result in 2.0.3 is to produce a measure rest in each new measure:
This is especially confusing because "measure rests" look the same as "whole rests". Again, I think the "Split Measure" command semantics shouldn't mean anything more than to split the measure.
Note that for the case if the measure being split was already a measure rest, there is a separate feature request #166096: Make individual beat rests when splitting a measure which I agree with.
I'm starting to work on this...I haven't heard any objections.
This isn't an objection, just an observation:
One reason where the current behavior would actually be desirable would be if you started with a 4/4 measure, split it into two measures of 1/4 and 3/4, then later rejoined them. You'd want to get back where you started, and this would possibly involve combining rests. I'm guessing this sort of use case is why it works as it does. But I don't know that this is a strong reason.
As discussed in #166096: Make individual beat rests when splitting a measure (#7). The plan is to make measure rests if beat count equals time signture.
Thanks for that observation marc. But as mike320 mentioned in that link on splitting measures, I was thinking about checking if the duration of the resulting measure equals the nominal duration of the measure. So if I were to also to that test when joining measures, then than means performing either:
- split of a measure rest measure and then a join.
- join of two measures of measure rest and then a split in the middle.
would both result in starting and ending with measure rests.
I'm going to wait to fix this...I want to hear more opinions before I try to "fix" something and then find out people don't like my "fix". Or at least an approval from someone high up.
At least take this +1 from someone down below :)
This has bothered me repeatedly when transcribing chants, in which I often start with a rather high time signature and often used the split command. It's inconvenient that I can't immediately see how many beats are available in the cutoff measure.
This strikes me as something not many people would care about, but it sounds good to me.