Implement AudioStreamSynchronized::get_bar_beats(), fix division by zero

[colinator27]

Fixes #92453, and tested using its MRP. get_bar_beats() was unimplemented on AudioStreamSynchronized, so this implements it similarly to other methods like get_bpm(). It could also be implemented more similarly to get_beat_count(), which gets the maximum beat count (not just the first), although it's a bit unclear what the intended behavior would be if there are different bar beats for each sub-stream.

The underlying issue is that get_bar_beats() is used to calculate the end time of the current bar (in AudioStreamInteractive), and without an implementation, it returns zero. This leads to a division by zero, resulting in NaN, which completely breaks timings. Thus, this PR also adds a guard which will check if bar beats is > 0 (if not, the transition will instead occur immediately), since this can probably occur with other stream types that still do not implement get_bar_beats().

(As a side note, this division by zero guard may also need to be implemented in the new code introduced in #99188.)

[adamscott]

Approving the PR. @colinator27, next time though, please create 2 PRs instead of cramming two features into one. ;)

[adamscott]

So AudioStreamSynchronized::get_bar_beats() returns the first bpm available. Is this something we want though?

[colinator27]

Approving the PR. @colinator27, next time though, please create 2 PRs instead of cramming two features into one. ;)

Noted, thanks!

So AudioStreamSynchronized::get_bar_beats() returns the first bpm available. Is this something we want though?

Yeah, I'm not certain on this. At least in the given MRP for the original issue, it's better than returning 0, although it could be defined to not work with synchronized streams. Besides always being 0 or returning the first BPM, the only other options I can think of is returning a minimum or maximum BPM, but it seems pretty arbitrary either way. I'm not sure if it should be defined what happens if there's multiple different BPMs, especially as they're meant to be synchronized, but maybe there's a case to be made.

[Repiteo]

Thanks! Congratulations on your first merged contribution! 🎉

[timoschwarzer]

I rebased #99188 onto the latest master branch so it contains this fix, thanks for mentioning!

[adamscott]

@colinator27 I just saw that the other methods in modules/interactive_music/audio_stream_synchronized.cpp use the same exact algorithm that the one you implemented. So no worries, it's OK then.