fix #100656: C key signature not transposed on change instrument

[MarcSabatella]

A change I made in #2228 had an unintended side effect. I added code to transpositionChanged() to optionally only transpose a given range (for use with the instrument change text), but since we support both tick == 0 and tick == -1 to indicate start of score with respect to instrument definitions (the actual instrument defined for the part has a tick of -1, whereas an instrument change on the first chord has a tick of 0), I needed to make transposeKeys() aware of this.

Since #9352 was only merged for master, this issue only affects the master, so this current PR is needed only for master (but would actually be harmless in 2.0.3, FWIW).

[Jojo-Schmitz]

Shouldn't #2228 get merged into 2.0.3 too?

[MarcSabatella]

@Jojo-Schmitz : I have mixed feelings on that. The current situation with respect to transposition and instrument change is actually pretty messed up - notes entered before the instrument changes is added are treated differently than notes entered afterwards, and TPC corruption can also ensue. I'd love to see that fixed. But if you've managed to use workarounds to get your score looking and/or sounding the way you want (eg, using hidden staves etc), then this change could break that, and this feels like maybe something we'd only want to do when we are making a bigger change. I guess we're already at this point doing enough things in 2.0.3 that could potentially change existing scores that maybe it is OK.

[lasconic]

I guess we're already at this point doing enough things in 2.0.3 that could potentially change existing scores that maybe it is OK.
Nope. I disagree. I started a list here and it's not that long. I don't want to make it longer except if it's really really worth it. (actually I was on the edge of reverting the cresc. one...)

[MarcSabatella]

Fair enough.