Clarify the context limits of octave and duration in a repeated figure

[lpugin]

See full description of the problem here

[ahankinson]

From the linked issue:

I noticed this to be particularly "nasty" with the use of repetition. For example, if you do:

2'G4!G^B^''D!f/ you probably expect the last chord to be repeated like that:

However, this is what Verovio produces - keeping the last octave as current:

[ahankinson]

I would expect that the scoping of the repeat would implicitly re-scope the beginning octave, by the assumption that octave markings can be omitted by the encoder to be more concise, but they are still applied to the context.

So, if you take 2'G4!G^B^''D!f/ and add the implied octave markings and expand the repeat, it becomes 2'G4'G^'B^''D'G^'B^''D/ -- they are equivalent in this expansion. This is what makes the most sense to me, but I'm fully aware that this rule probably hasn't been checked against the other behaviours.

In other words, 2'G4!G^B^''D!f/ expanding to 2'G4'G^'B^''D''G^''B^'''G/ seems pretty weird to me, especially because it very easily can overrun the 4-octave limit if the octaves are additive on every repeat. Edit: I got that wrong...

[ahankinson]

Action: Add a note to the shortcuts section to say that the repeated figures are repeated as the musical values given in the first section, and are not continued in the repeat.

[ahankinson]

NB: related to (but not the same as?) #132

[BaMikusi]

That sounds all good to me (in keeping with our longer discussion in #132).