beam breaking in tuplets

[Victor Eijkhout]

Is this a bug or a feature? (2.9.27, PPC)

\relative c'' {
\time 3/4
\times 3/4 { r4 g8 a16 bes a8 bes16 c d8 c16 bes } |
#(override-auto-beam-setting '(end 1 16 * *) 2 4)
r4 g8 a16 bes a8 bes16 c  |
}

I was expecting the beam in the tuplet to be broken the same way as  
in the measure after.

If I was expecting wrong, I'd appreciate a hint as to how to force  
this behaviour.

Victor.