When changing to C-Major/a-minor after multimeasure rest, naturals are not shown/saved

• May 24, 2018 - 12:52
Reported version
2.2
Type
Functional
Severity
S3 - Major
Status
closed
Project

Create a score which some non C-Major/a-minor key sig, add a C-Major keysig in a later measure and some notes, and yet another keysig even later, enable multimeasure rests, result:
test-1.png
I.e. the naturals are shown.
Save, close, re-open (or take attached sample score), result:
test-2.png
I.e. the naturals are lost.

Works as expected in 2.1, so is a regression in 2.2, and still exists in the current 2.3

Attachment Size
test.mscz 3.24 KB

Comments

Hmm. This commit also seems to be responsible for another regression, a case of one score that did not have a layout shift problem in 2.1 but does in 2.2.1. Thus, the change actually caused the very problem it was trying to fix in that particular case, as per my comment in https://musescore.org/en/node/267602#comment-831071. And that particular fix was 2.x-specific - it was never merged for master, because master was sufficiently broken with respect to key signatures (naturals not displaying at all) that it was impossible to test in any way. Not sure if that's still the case. I wouldn't expect the same test cases to necessarily show the problem on master because it is so dependent on the exact widths of things, and that surely has changed a lot.

Anyhow, I think we should consider reverting that particular change, then going to master and getting key signatures with naturals to work again (if they don't), and then readdressing #267602: Measure shifts back and forth with each layout in light of these regressions, so we come up with a fix that doesn't also break those cases.

I have dealt with quite a few of these sorts of layout-shifting issues over the years, and I can say they are no fun to debug. If someone else wants to take a crack, I'm more than happy to share my insights, maybe we can discuss over IRC?

In reply to by Marc Sabatella

Severity S3 - Major S5 - Suggestion
Status (old) fixed patch (code needs review)
Status fixed  

Hey there :)

I have very little understanding of coding and software development, so I though I'd just ask it here instead of reading the code in the link below this comment.
Is there an estimated time we can expect an update on the naturals in multimeasure rests problem? I'd like to export a piece of mine, but I'll wait until this problem is fixed because multiple parts experience this problem.

Thank you in forward! :)

Greetings

Status (old) patch (code needs review) fixed
Status fixed

As the comment right above yours indicates, this is "Fixed in branch 2.3" - which is planned for release "any day now". Feel free to downlaod a nightly build using the "Download" menu above if you'd like to help test today!