With these new time signature fixes coming, I tested again a project from couple years ago when we were wondering with LCipher the funny behaviour.
This is a time signature change for one measure, no tempo changes. Results are still as interesting as they were back then, so here's the old pic:
http://img684.**************/img684/8...rechanges1.gif
To my mind, the change should happen within measure 5 and everything from measure 6 forward should stay as is. For example, cut away the new odd measure 6, done. What do you think?
Workaround is of course to split the item at markers, but there's still the odd measure 6 created when changing the ts at M5.
EDIT: This anomaly is FIXED in 4.25. Really appreciate all the related fixes there. Thanks!