use accidental width to space key signatures

[Nick-Mazuk]

Resolves: #9379

This PR uses the accidental width to calculate the spacing of key signatures. Before, all accidentals were spaced 1 space apart, regardless of font or symbol. Now, the key signature calculates each symbol's width to space them out appropriately, and the gap between them is accidentalDistance, the same style value used for accidentals on chords.

This will provide greater consistency and clarity for accidentals.

Notes

• This PR may not work for custom key signatures. As of today, it looks like it's not possible to create custom key signatures in v4, so I was not able to test it.
• Unlike for chords, there does not seem to be much space remove extra space with symbol cutouts. This is because key signatures have well defined patterns (like they move in 4ths or 5ths), and there's just not the vertical clearance in any of the fonts. The only places I could see making the case for adding this is sharps in Peteluma or naturals cancelling out the sharps. If this is desired, I can add that in (though it would increase the scope of the PR quite a bit).

[Jojo-Schmitz]

Compare with this

using Leland, it does seem to use cutouts and looks narrower than your 1st example. same with Bravura, Emmentaler, Gonville, MuseJazz and Petaluma too, very clearly (and possible a tad too much) for the latter two in 3.6.2, MuseJazz:

Petaluma:

Not with flats though.
So IMHO those cutouts should be used.

For more comparison see #9390's (faling) vtests (a backport of this PR on top of the 3.x branch, as the vtests on the master branch just don't work)

We can get to a denser spacing for sharp key signatures, one similar to that of MuseScore 3(.6.2) though by reducing (the default for) Sid::accidentalDistance from 0.22 to 0.1.
That in turn would make flat key signatures narrower than before though.

On the up-side: Even reducing that value to 0 does not result in that too dense appearance of MuseJazz and Petaluma, although it could be made denser still (using cutouts).

[oktophonie]

Using the cutouts would be helpful, particularly for naturals-cancelling-sharps, as you mention, since they look rather wonky otherwise:

I also think it would be good to have the spacing as its own setting (independent of accidentalDistance).

[Nick-Mazuk]

Three updates to the PR:

1. Rebased to make sure it works with the most recent version of master
2. The accidental distance is now defined by its own style parameter named keysigAccidentalDistance It currently has the same value as accidentalDistance.
3. Accidental cutouts are now honored

Unless there's anything that needs fixing, I'm considering this PR and ready for merge (upon approval).

[Nick-Mazuk]

Looks like the vtests are failing, but the exact failure is expected with this change.

[Jojo-Schmitz]

Before:

After (esp. see the naturals being much closer):

Seems a vtests label is in order ;-)

[oktophonie]

The kerning-via-cutouts is excellent in principle, but the naturals in flat-formation are too close together. Indeed everything is a bit too tight so probably the basic spacing value should be increased slightly.

[Nick-Mazuk]

Quick update, this PR is not ready for merge. For some reason, if the gap size style property for key signature is updated, the score does not visually update.

[RobFog]

You can mark the pull request as a "draft".

[Nick-Mazuk]

Ok, assuming the tests pass, I believe the PR is ready for merging. I've updated the exact spacing values after a discussion with @oktophonie.