[MU4] exclude accidentals etc. from chord shape if autoplace disabled

[MarcSabatella]

There are a number of spacing issues that stem (pun not intended)
from the inclusion of things other than the notes themselves
in chord shapes:
- despite the best intentions of "smart layout" changes for 3.0,
accidentals can't tuck under the previous note
- accidentals in a measure result in increased note spacing,
even if the measure ie wider enough to not need extra space
- for the latter, similar for arpeggios, fingering, etc

There has been good discussion and proposals for ways of fixing this,
but they require relatively significant changes to the layout,
probably require new options to support, etc.
And the tucking issue might not be solvable at all for beamed notes
without changes to the beaming algorithm as well.

This commit provides a manual workaround:
allowing the user to disable autoplace for a note,
which results in the accidentals and other attached symbols
being excluded from the chord shape.
The note itself is still added to the chord shape,
so the note is spaced as it normally would be
if it didn't have accidentals etc.
That means you will get collisions if spacing is tight,
so don't disable autoplace unless there is definitely room.
And in that case, the spacing automatically improves,
plus accidentals will tuck under previous notes where appropriate.

I'm not calling this a fix because eventually this should be automatic.
But see:

https://musescore.org/en/node/289446
https://musescore.org/en/node/299741
https://musescore.org/en/node/285591

I have not yet created vtests because I'm not sure we're ready for that, I expect the possibility of further changes based on review. I'm therefore marking this WIP, and not necessarily recommending this be included in 3.5 unless there is strong agreement it makes sense to do this now while we await a good autoamtic solution.

[MarcSabatella]

Some images to give an idea of what this does. First, showing the before/after regarding tucking:

Before:

After disabling autoplace for the G#:

[MarcSabatella]

Actually, regarding the uneven spacing, it helps, but not as much as I'd like, so I'm not bothering with a picture yet.

[MarcSabatella]

See also #4784. I guess in the end I'd say, neither is ready for merge. It will require more work to make something really useful out of them, I'm afraid. But this one to me is worth keeping open as WIP I think.

[Jojo-Schmitz]

Any further plans on this? At least a rebase is needed.

[MarcSabatella]

This definitely feels worth revisiting post-4.0 whenever the next big "engraving push" happens. There would need to be a significant update of the spacing algorithm required for this approach to be viable. And I understand this is being looked at.

Meanwhile, this particular hack is "harmless" enough as it won't affect scores unless you explicitly disable autoplace for a given accidental. On the other hand, who knows how many scores have it disabled because people tried it in an attempt to solve some other problem, found it had no effect, and just left it disabled. So there is that risk to assess. Personally I'd rather see the effort going into a real fix.

[Jojo-Schmitz]

Interestingly when rebses onto my PR #9000 some mtests fail with autoplace issues on copy/paste, see https://github.com/musescore/MuseScore/pull/9000/checks?check_run_id=3589269967#step:5:698

Actually this PR here seems to have never run the mtests in the first place?

[Jojo-Schmitz]

Rebase needed

[MarcSabatella]

Unless I'm missing something, this shouldn't be needed anymore, now accidentals tuck automatically, as of @mike-spa 's most recent set of changes.

[Jojo-Schmitz]

Then this one should get closed I guess

[MarcSabatella]

Probably so, I’m unless there is a use case for which it could still make sense. I haven’t had an opportunity to play with the latest builds yet.