Erratic behavior of lines in Edit mode

• May 15, 2020 - 13:05
Reported version
3.x-dev
Priority
P1 - High
Type
Functional
Frequency
Once
Severity
S3 - Major
Reproducibility
Always
Status
closed
Regression
Yes
Workaround
Yes
Project

OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.5.0.11614, revision: 1ee2fe3

1) Load this test file (default score with a few lines) : lines handles test.mscz
2) Drag slightly a line (no matter, but says Down)
3) Select the middle handle, and use now the Up arrow to rectify the position

Result: the line continues to go Down

(same behaviour with current 3.X)

NB: no better with other handles. Eg: reduce the hairpin size with the mouse. Select right handle and use the opposite direction arrow to rectify: same inexpected behaviour


Comments

Workaround No Yes
Priority P1 - High

I know there were deliberate changes to the drag behavior, so that it will actually change anchors. And I don't fully understand how it's supposed to work. But I agree this seems like a bug.

Workaround is to use the Inspector.

Status PR created fixed

Fixed in branch 3.x, commit 81cd95f4d5

_fix #305477: fix arrow keys behavior for moving elements

Fixup for PR #5763, especially for changes in ad6c8581d79e675f33c1481bcfe920742d7ecf8b_

Just downloaded latest 3.5Alpha AppImage and then added a crescendo to a simple measure, shift+right arrowed to extend the crescendo hairpin a few notes further, and then used left arrow (up and down also) and the line started to move to the right instead of the left etc. Not sure if this should be actually "opened" again or a new issue should be created, but it sounds similar to the OP bug.

Maybe the AppImage is old still since it's only been one day?

To be clear, there was only one 3.5 Alpha and it’s several weeks old. If you want to test the latest code, you need to download a Nightly build.

P.S. After checking, unfortunately as compared to previous versions there's a "lagging" feel to [ctrl+left/right arrow] with a crescendo hairpin. For most people it more than likely can be considered a compromise due to the new algorithm for dealing with anchor-point changes, so it can be 'dealt' with, but it wouldn't hurt to get that sped up a bit if there's some inefficiency going on in the code somehow. I'm sure it can be refined so that it is just as snappy as earlier versions.

Fix version
3.5.0