Bug deleting duplicate time signature on MuseScore 2.0.3 on OS X 10.12.2
The relevant bars (at least where the bug presents) are 121 to 127, but the bug really affects all bars past 121 in the attached file.
I put in a 1/4 time signature in 122 in a passage that was otherwise 3/4 while trying to figure out the right time signatures for that passage. The 1/4 was actually a bit further down originally. I put in an extra 3/4 in order to prevent bars further down from being messed up when I was inserting the original 1/4. Then I decided to move the 1/4 up to its current position, and I was left with a redundant 3/4 - as in, a change to 3/4 (bar 123) and another 3/4 a few bars down (bar 127).
Bugs:
1. When I delete the 1/4 measure and then delete the 3/4 in bar 123, rests for Voice 2 that I deleted earlier reappear and stem directions change.
2. When I delete the 3/4 in bar 127, the rest of the piece gets messed up completely. Deleted rests reappearing, stem directions changing, spacers disappearing, dynamics disappearing. I give a demonstration of this in this video: //youtu.be/uOhx_6pQMH8 .
I have also tried deleting bars 129 and 130, then deleting the redundant 3/4, and then inserting two new measures and manually populating 129 and 130. The score looked fine when I did that, but the crescendos stopped working during playback...
| Attachment | Size |
|---|---|
| Sonata in C Minor real Rendezvous revised - MS2 - current.mscz | 90.84 KB |
Comments
1) Can you be more specific about what problem you are seeing? Which rest specifically becomes visible, which stem direction specifically changes? I don't notice any problems offhand if I follow those instructions.
2) Same question. I delete that bar and don't offhand notice any problems. Can you be specific about what problem you are seeing.
Your video for 2) doesn't seem to show the same steps you described. You talked about deleting *measure* 127, but your video shows you deleting a *time signature*. It is normal that a time signature change will result in measures needing to be rewritten and thus losing some of the things formerly attached to the old measures, like line breaks and spacers.
In reply to 1) Can you be more specific by Marc Sabatella
I do in fact talk about deleting time signatures in the two bugs. For instance, when I say, "When I delete the 1/4 measure and then delete the 3/4 in bar 123," what I mean is that I delete the measure containing 1/4 and then delete the *time signature* 3/4 in bar 123. Similarly, when I say, "When I delete the 3/4 in bar 127," I refer to the *time signature* 3/4. As a result, the video is a demonstration of what I describe in bug 2.
Among lots of other instances, deleted rests in bar 131 in the lower staff reappear, after the E-flat in the upper voice.
Thank you for responding.
In reply to I do in fact talk about by Kaiser W.
You are right, you do mention both deleting time signatures and deleting measures - it just wasn't totally clear how your video matched the steps you gave. Maybe I misinterpreted something. But I don't see any 3/4 time signature in bar 127. Not in the original score, not after following the steps in #1. Hence my confusion. Beyond that, it isn't always clear when you refer to measure numbers if you mean the *original* numbers, or the new numbers after the deletion of measures. nor is it clear if the steps in #2 are meant to apply *after* the steps in #1, or if they are *alternate* steps.
So I was having trouble figuring out how to reproduce what you are seeing, but after some more trial and error, I think I have it - the clue to look at measure 131 was the key.
You are absolutely right that upon deleting a time signature, some things in the score after that change. This is exactly as I said before - the measures need to be rewritten, and this means you lose things attached to the measures themselves as opposed to specific notes (eg, breaks, spacers). It also means any "holes" created by deleted rests get re-filled. So this much is perfectly normal, which I hadn't specifically noticed before but makes sense given how MuseScore works internally - it generally doesn't like these holes and prefers dealing with invisible rests over deleted ones.
Anyhow, there are an outstanding issues in the tracker suggesting we add smarts to the algorithm to detect when we can avoid the measure rewrite - see for example #34076: Don't eliminate layout elements upon some cases of time signature change and #51176: Change the time signature remove the score elements. What you are seeing with the "holes" being filled is just another effect of that same phenomenon.