Support measure boundary/barline locations

[joeberkovitz]

As raised by in #251 (comment) we sometimes need a way to refer to the locations of the start points or endpoints of bars, in order to position directions, slurs, ties, and other objects relative to them. Ordinarily these horizontal positions are occupied by barlines. But sometimes barlines are absent from these positions — for example, at the start of a system — so let's call these positions measure boundaries for clarity.

So far, the measure location data type has been limited to a metrical position within a measure: some fractional number of beats relative to the first event in the bar. Boundaries, however, do not occur at a metrical position. The start of a bar is separated from its first note or event by some amount of horizontal space. And the end of a bar may not occur exactly where the next note would have been placed if the bar had been longer.

[joeberkovitz]

Proposal: measure locations being used for positioning purposes may employ the special keywords start or end to refer to the position of a measure boundary. If used standalone, these refer to the placement of the start or end boundaries of the containing measure. If preceded by a measure index (like 4:end or 6:start by analogy with metrical positions), they refer to the boundaries of the measure identified by that index.

As a special case, the measure location corresponding to the full duration of a measure (e.g. 1 in a 4/4 bar) is defined to be the same as end, to avoid ambiguity. Although a barline and an imaginary "next note" might not have occupied the same exact position, there is in fact no next note.

[notator]

👍