Comment #2 on issue 2682 by [email protected]: Collision of notehead and beam in beam-collision-basic.ly regression test
http://code.google.com/p/lilypond/issues/detail?id=2682
I don't think that we should change systematic regtests to cover only what LilyPond does well since we then miss when it changes something to the better.
Instead changing the description to "Collisions between manual beams and notes are minimized." seems more useful.
