Re: 1.4.12 stem direction bug

[Erik Sandberg]

I found a better example of the stem direction bug I found about a week ago:

\score{ \notes {[g'' <e'''8 g> g''] }}

The g'':s use stemDown, while the chord uses stemUp. The result is not nice, 
the g'':s get "mirrored".

It is sometimes OK to have a stem direction different from the others in the 
same beam, but only if the _highest_ note among the stemUp notes is enough 
much lower than the _lowest_ note among the stemDown notes. Looks like some 
bug in the code for finding the highest one (it seems that you 
consider only the note with the lowest instead of highest pitch in each 
chord).

The bug also works upside-down, i.e.:
\notes {[d' <f8 d'''> d']}
Unfortunately, I am not yet familiar enough with the internals & source to 
find & fix the bug.

Erik