Glyphs and bounding boxes sometimes jump to the top left corner of the page

[JoyfulGen]

@PhoebeDM brought up this issue a while ago: glyphs and bounding boxes sometimes jump to the top left-hand corner of the page when she tries to move or resize them. They don't always want to go back where they first were. We still haven't figured out why this happens, BUT the issue finally happened for me (yay!!!) so I have a little bit more data.

In the screenshot below, the arrow is pointing to where the neume was and the circle shows where the neume jumped to:

In the MEI that corresponds to that screenshot, the syllable actually appears at the very beginning of the green staff, before the first divisio.

Here's the original MEI I got from Phoebe: 141v hai before.zip

And here's the MEI I downloaded after the neume jumped: 141v hai after.zip

In both cases, the jumping neume is called "hai."

EDIT: At the moment, I can't reproduce this issue. I've been trying to move the same neume around and nothing is happening! Frustrating.

That's all I know for now! @PhoebeDM and @annamorphism please let me know if this matches what you've noticed.

[annamorphism]

I can't reproduce this on these files either! So why is it sometimes a problem and sometimes not....???

[annamorphism]

Random observations: 1) There seems to be a divisio in the "after" that wasn't in the "before", but few other differences of consequence (so maybe the divisio is doing something?)
2) The BBox for "hai" keeps the same coordinates in both files, but while in "before" it appears in the file between the other boxes for "euouae", in the "after" it appears before any of the other elements in that staff.
3) Obviously the coordinates for "hai"'s neume change drastically, though not to the origin (it's 957, 182, 912, 151--not values shared by anything else, so far as I can tell, and not obviously linearly related to the original values or anything.) Strangely, it is still a b--but instead of octave=2, it's now oct=7 (!).