RepeatList should use measureIndex instead of measure->no()

Having a jump marker (such as Fine) in a measure where the consecutive measure has measure number -1 offset will result in the second measure being included into the jump when it shouldn't.

Report, start of analysis and example score in https://musescore.org/en/node/285671

This score is affected by using no() where measureIndex() should be used. See
https://github.com/musescore/MuseScore/blob/master/libmscore/repeatlist…
Using index() should work as well and be a little more performant.

Likely this change should happen for the whole of the RepeatList class internal logic…