Certain Formatting Changes/Resets Do Not Save in Parts

• Jan 20, 2021 - 05:29
Reported version
3.6
Priority
P1 - High
Type
Functional
Frequency
Few
Severity
S3 - Major
Reproducibility
Always
Status
closed
Regression
Yes
Workaround
No
Project

I've come across multiple similar issues in 3.6 formatting defaults/saves (some already posted in the issue tracker; sorry for the redundancy) but I'd like to collect them here.
For similar issues, see:
#315815: Header and Footer Text is not identical to 3.5 scores when opened in 3.6
#315781: Measure Number Offset Value Changes Everytime Opening the File

Summary of problems: parts do not save changes (some don't save if they are set to style/default) in position/offset for hairpin, staff text, measure numbers; font size for header and footer; disabling of indentation on first systems; and deletion of top-right page numbers. I have a good feeling some of these can appear in main scores without parts, but I haven't found how to recreate that yet...

OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.6.0.487915773, revision: 1977cb3

How to recreate the problems.
1) Open the score and observe that the main score has these desired settings:
- Header in 10 pt. font (3.6 default is 11 pt.)
- Footer in 8 pt. font (3.6 default is 9 pt.)
- Page number on the center-bottom of all pages except the first (which doesn't have one)
- Hairpins at Y: 3.50 sp.
- Staff text at Y: -2.00 sp.
- Measure numbers with center alignment

2) Open any part and observe that it has these undesired settings: (note that Part 1 has staff text in a different position than the other parts due to autoplacement)
- Enabled indentation on first system
- Header in 11 pt. font
- Footer in 9 pt. font
- Page number also on the top-right of all pages except the first
- Hairpins at Y: 2.50 sp.
- Staff text at Y: -1.00 sp.
- Measure numbers with left alignment

3) In a part, make these edits (you can only do one/a few of these changes and can apply to all similar elements, if applicable; you must set as style, when applicable) and then save:
- Disable indentation of first system
- Header to 10 pt. font
- Footer to 8 pt. font
- Delete top-right page numbers
- Set hairpins back to Y: 3.50 sp. (set as style)
- Set staff text back to Y: -2.00 sp. (set as style)
- Set measure numbers back to center alignment (set as style)

4) Close and re-open file and observe that the score and parts regressed to what I noted in 1) and 2). You do not need to apply changes to all parts and you can only pick one of the steps; the re-opened score regresses to the original setting.

I made this sample score as a copy of one that originated last fall. I'm using the new fonts but my Styles palette has some custom settings (Ex. minimum measure width at 25.00 sp.). I figure plenty of you know this way better than me; I'm just here to report what I've found. Hope this helps!

Attachment Size
3.6 Test.mscz 28.95 KB

Comments

In addition -
- Composer, Lyricist, Style and Part Name text sizes default to 3.6 sizes / are not updated in projects loaded from 3.5
- Bar numbers italicised by default

Regression No Yes

I have a similar issue, not sure what it comes under but here goes -

1) Open a 3.5 score (not sure how relevant version is though) and select a part.
2) Open up style and adjust a parameter (e.g remove the "bold" option under text for the header)
3) Click "Apply to all parts"

Expected - this issue is fixed and paper size stays the same
Actual Result - the issue is fixed (good) and paper size is the same as the part selected when the modification was made

(N.B. Reproducing the issue, I clicked on the Keys 1 / Conductor / Full Score part before making the modifications to style)

Attachment Size
DEC NEW ORCH FIX 3.6 ADJUSTED.mscz 39.02 KB

I can reproduce some but not all of this given the original score. I see different current settings on loading the score, and some of those settings do survive save/reload. But some do not. And I am unable to reproduce that behavior in a score created from scratch, or in one I import from 3.5 myself.

So I'm guessing it has something to do with the specific series of steps used to get the score to the state it is in now. Can you retrace that and describe in more detail?

Hmm, I take it back, I did manage to reproduce it in a score imported from 3.5. Let me describe what I think is happening, you tell me if this sounds like what you are experiencing. I'll take hairpins as an example.

1) create a score in 3.5.2 using all defaults
2) generate parts
3) verify the default offset below the staff is 3.5 sp in both score and parts
4) save the score
5) import it into 3.6, choose "keep old style"

Due to a known bug with a pending fix (#315530: "Keep old style" still updates style in parts, the "keep old style" doesn't take for the parts. Among other things, the hairpin offset is now at 3.5 for the score (old default) but 2.0 (new default) for the part.

Something noteworthy I see already is that the 3.5 sp offset in the score shows as being the default - the reset button is disabled. I guess that's OK since this is the default for old scores, but I think this is where things start to go wrong. You'll also note that the 2.0 sp in the part does have a reset button by it.

Now:

7) change the 2.0 sp in the part to 3.5 sp. you'll note the reset button disables, which is a bad sign
8) save (actually, in my test, I did a save as, after first flipping back to viewing the score)
9) reload

Result: hairpin is back to 2.0 sp for the part. Not surprising; it wasn't written to the file because 3.5 sp is supposedly the default, even though it isn't really the default.

Somewhere we're getting confused about what the default really is here I think. But that's why this afrects some setting and not others - only ones whose default values has changed seem affected.

It's possible the PR for #315530: "Keep old style" still updates style in parts will fix this too, and in any case, probably not worth further investigation until that code is merged.

I think this is tied to the "usePre_3_6_defaults" tag that gets written to the file upon import. I guess the intent is to remember that we are trying to use the older defaults for these styles. And that's why the 3.5 sp for the hairpin actually does show as the default. I guess that's actually pretty clever. But somehow it isn't being honored everywhere it needs to be, in particular upon reading the score later. The part does not have the 3.5 sp setting written to it, not surprising because we think it's the default. So on read, we should be setting this to the default, but we're setting it to the new default, not the old one.

This gives me some hope that it is in fact going to be fixed along with #315530: "Keep old style" still updates style in parts - basically it's another manifestation of the issue of "keep old style" not applying to parts.

In reply to by Marc Sabatella

If it helps, this is how I made the sample file:
The original I created from a 3.5.x version from last fall (or whatever version was up-to-date back then) and worked on it until now. When the 3.6 update came about, I enabled both new fonts and the changes but I did not reset the Styles palette as I had some defaults I wanted to keep. To make parts, I made a copy of the score (I needed to do quite a bit of editing for parts so to avoid messing up the main score I made a copy just for parts) and when engraving the parts, I noticed certain things don't save. To create the sample score, I made a copy of this "parts only" score and deleted the music so all that remains is just some basic settings and elements.

Priority P1 - High

Thanks for the report and detailed info! We're trying to fix all of the most critical issues discovered in 3.6.1 and this is definitely the sort of thing that means.

OK. I suspect the same applies even if you don't click keep old style, because really that button is a bit misnamed - it really just means "keep old fonts". A little more to it than that - there are a smaller number of style settings that are automatically associated with the font, like stem width etc. But things like hairpin offset, that's not connected to the font, that's just a change made for 3.6 to improve overall results, keeping them closer to the close they are connected to. Really, the old default only made sense for piano scores, to more or less center them between the staves.

Anyhow, bottom line, I can reproduce the problem in 3.6 with our without keep old style, and the PR mentioned above fixes it either way. So, 3.6.1 should be good to go here.

Meanwhile, I won't even bother marking this as a workaround, but - even in 3.6.0 currently, you can actually set that hairpin to anything except 3.5 sp. So, 3.51 sp is fine.

In reply to by Jojo-Schmitz

I tried the nightly build 202101270430. Opening a MS 3.5 file with parts and 'Keep old style' gives good results. As far as I could detect, only the position of the bar numbers is reset.
But when I save that file and reopen it, then lots of styles (music font, text styles) are reset (to the default?), even in the score itself. So the fix still needs some fixing.
OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.6.0.514063336, revision: 379a44d

Fix version
3.6.1