stripTies() now sets StreamStatus.beams = False

[jacobtylerwalls]

• Fixes Beaming corrections for stripTies() #994: stripTies() sets StreamStatus.beams = False
• for that to have effect, refactored fixupNotationMeasured() in musicxml export so that it doesn't use a dummy stream as an iterator and then query that dummy stream for its beam status
• called splitAtDurations() earlier in musicxml export so that it happens before fixupNotationMeasured remakes beams
• moved full-measure rest splitting logic out of musicxml exporter and into splitAtDurations()

[coveralls]

Coverage increased (+0.006%) to 92.267% when pulling bfb6e97 on jacobtylerwalls:stripTies-beams into a8ae14a on cuthbertLab:master.

[mscuthbert]

Other than that Q on dot groups, looks good to me! And probably worth merging first since the actual problem solved is so much bigger.

[jacobtylerwalls]

Super, is there a review you need to submit so I can see the dot groups stuff?

[mscuthbert]

Ah! Here's the review!

[mscuthbert]

Can you explain why this can be removed? Are we doing it elsewhere? Ah, I see below. Interesting! Does it work with dotGroups, since they're only split above?

Perhaps dotGroups should also be split in splitAtDurations (as an optional parameter) since they're something most software, including the musicxml spec, can't handle.

[jacobtylerwalls]

I think both this and the 'inexpresible' bit above show that there is still a need to call splitAtDurations() in these edge cases. I'll restore the else: objList = [obj] bits so that all we skip here is the full-measure rest logic.

[jacobtylerwalls]

For anyone else: the reason this was moved into splitAtDurations() was because it needed to be moved somewhere earlier during musicxml parsing so that beams could be made afterward (fixupNotationMeasured), and it seemed convenient to just make it splitAtDuration()'s responsibility.

[mscuthbert]

Sorry about that -- mentioned the review now.

n =  note.Note()
n.duration.dotGroups = [1, 1]
n.show()

if this works, we're fine.

[jacobtylerwalls]

if this works, we're fine.

checked this against the most recent commit on this branch and get correct output (i.e. same as master)

[jacobtylerwalls]

Thanks for catching the multiple StreamIterators; I had just internalized that first() was faster because it short-circuits and have bee using it all over the place without thinking, but certainly faster to just get something once.