daveyarwood
commented
4 years ago I do appreciate these issues - please keep them coming!
I've never intended for markers to be used inside of repeated figures. The idea is that a marker represents a single moment in time. I suppose it might make sense to make it a compile-time error if a marker is encountered in a context like this, I just haven't bothered to do so.
Out of curiosity, I tried running your example using the WIP ground-up rewrite of Alda (Alda v2), and it actually happens to work the way you're expecting it to work! I didn't plan it that way, it just happens to work. But I guess I have some thinking to do about whether that's the way I really want it to work.
Something that makes this especially interesting is that I did have a thought, a while back, about using markers as a way to synchronize parts during a live-coding session. In a setting like that, it would actually be really convenient to have a named point in time that a new part could "latch onto" so that it starts exactly in sync with the existing parts. Imagine that you had an infinitely looped bass line, something like:
This "infinite loop" syntax is imaginary. It'll happen someday in the not-so-distant future!
electric-bass:
[ %sync o1 g4. > d4. f4 |~4. d8 f4 g ] **Then, while the bass line is looping, let's say you want to add a drum pattern, but you want the first note to line up exactly with the first note of some repetition of the bass line. You don't care which repetition, it's just important that it lines up exactly with the %sync marker, the next time it rolls around in the loop. So, maybe you could add the drum pattern like this:
midi-percussion:
@sync [ ... notes go here ... ] **That's just an off-the-cuff idea from a while back. My point is, it's quite possible that at some point in the future, it will make sense to use markers inside of repeated figures. For now, it's not something that Alda officially supports, and I don't think it's very useful in the absence of the live coding feature that I described above.
In the case of your example, you could use voices instead, like this:
cello:
[V1: o3 c d e f g a b > c
V2: o3 c < b a g f e d c]*2
dhpiggott
It isn't to me clear from the documentation for markers and sequences/repetitions whether they are supported in combination, and if so what the semantics are, but given the documentation at https://github.com/alda-lang/alda/blob/master/doc/markers.md#markers my best guess would be to expect both repetitions of the following to render in the same way:
The first repetition renders as I would expect it to (the two scales play simultaneously), but the second repetition renders as though the marker reference isn't there (the ascending scale plays first, followed by the descending scale).
I realise this may be a particularly pedantic issue to raise - so if you'd prefer I didn't, please tell me and I'll hold fire on any more!
A bit more context for how/why I found this and the others that I've opened: I started experimentally building something for programmatic music generation in Scala a few months ago and, happily, discovered Alda Lang soon after starting. Rather than attempting to reinvent the wheel by defining my own DSL, I'm effectively trying to write a clone of (parts of) Alda in Scala. I'm mostly using the docs in https://github.com/alda-lang/alda/blob/master/doc (as opposed to reading the source), which puts me in a mindset where I find ambiguities/edge cases!