I've noted some odd behavior with \partcombine when used in combination
with \partial, though I'm not entirely sure that \partial is to blame.
In the output, note that the r2 in the 2nd measure does not appear, and
the part-combiner has analyzed this as "Solo". In the 3rd measure, the
last beat of the measure has been analyzed as "Solo II". If the pickup
and \partial are eliminated, the 2nd measure still analyzes as "Solo"
and does not print the half-rest, but the last measure is analyzed
I don't see where (or if) determine-split-list() makes judgments based
measure position. No doubt I'm missing something, but I'm still
clueless after several hours of study. Any hints for where to look? (or
better, how to fix?)