"Reset positions" on 2.x import should (?) also reset shape adjustments to slurs

• Dec 12, 2018 - 17:53
Reported version
P1 - High
S5 - Suggestion

OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit):, revision: dc5e1d7

Open the attached score (created in 2.x) and take the reset option. Now test all the slurs (inc. continuation slurs) by applying Ctrl + R to them one after the other.

Expected result: Nothing happens (the slurs have been reset already).
Actual result: Some of the slurs are being unexpectedly reset.

Attachment Size
chansonette_2.x.mscz 32.94 KB


Priority P1 - High

Curious, not sure what is happening, might have to do with the order in which things are reset on load? Or it might be that the positions are reset but the individual offsets of the grips (handles) are not. Seems this was reported previously...

Severity S3 - Major S4 - Minor

Looking at the code, it appears my initial guess is correct - the reset option resets positions but not shapes of slurs. And this is exactly what the dialog says will happens. So technically it seems to be "by design". I can see why this might be desirable, but I can also see why resetting shapes might sometimes be desirable. If there is consensus this should change, it certainly could. Let's get some more feedback (in the forum if necessary) and then either fix or close this.

Title Slurs in imported 2.x scores not properly reset "Reset positions" on 2.x import should (?) also reset shape adjustments to slurs
Severity S4 - Minor S5 - Suggestion
Severity S4 - Minor S5 - Suggestion

No, they are two different commands, both labeled correctly for what they actually do. The reset on import says it does positions only, and that's exactly what it does. The reset in the menu says it does shapes and positions, and that's exactly what it does. Two different commands, two different code paths, both doing their jobs as designed, currently.

The suggestion here is that maybe the import should do a reset of shapes as well as positions, just like the menu command (or Ctrl+R) does. As I said, I can see arguments both ways. The argument for the current behavior is that really, the changes in layout from 2.3.2 to 3.0 really are mostly about positions and not shapes, so manual adjustments to shapes are actually somewhat safer to preserve. Not totally, though, because if the endpoints change (and they have in some cases, then shape adjustments won't make sense. Plus the layout of the underlying measures may have changed. So that leads to the argument for: is if there is a consensus people find they do prefer the shapes reset as well based on actual experience using the feature (eg, doing the reset on import, then trying Ctrl+R and finding they like the new defaults results better than their re-interpreted manual adjustments.

It does have the advantage of allowing you to make the choice yourself, either individually for each slur or by doing the select all yourself. Still, it is worth considering using some real world examples which is more likely to give better results. Another possibility of course is to provide both a "positions" and "shapes" checkboxes in the dialog, or maybe just a single checkbox "also shapes"

Hmm, it's fine in 3.0.4 but not as nice in master. This is not related to 2.x import per se, it's a particular unusual case of a slur that has been set to appear between two different voices - not normally something that is recommended - confusing my code that tries to curve slurs more where needed. It wasn't actually needed here. It's kind of a corner case, but could you file this as a separate bug report? Doesn't matter if you use the same score or create a new one in 3.0 (master), result should be about the same.