Re: Support articulations, slurs and breaths in MIDI (issue 26470047)

[Devon Schudy]

Janek Warchoł wrote:
> i don't understand why would someone not want to have at
> least tremolo repeats unfolded

I think everyone does want tremolo and percent unfolded; only volta
needs customization.

The global context already knows (via its Output_def) whether it's
doing layout or midi. So there are some simple ways to implement
automatic unfolding:
 1) Make *_repeat_iterator act like Unfolded_repeat_iterator when unfolding.
 2) Pass the context to the iterator-ctor functions, and have those
for repeats create an Unfolded_repeat_iterator instead when unfolding.
This has the advantage of putting the unfolding decision in Scheme
where it's easy to customize.

These have the same problem as \unfoldRepeats where voltas in one
voice don't unfold the other voices, so they get out of sync and
produce a useless midi. I guess this is a reason to keep voltas folded
by default.