2.1.6 - wierd polyphony output

[Hans Forbrich]

On Lilypond 2.1.16 (self compiled).  Originally noticed on 2.0.1

Not sure whether these are interrelated or should be reported
separately.  The attached .ly and .pdf/.dvi) are the original test
case(s).  Hopefully I've put in slightly more concise test cases in line
here.

1: Chorded polyphony with dotted notes in voice 1 misplaces dots (as
shown in attached output bar 7)

\score { 
  \notes \relative c'' { 
      time 4/4 r1
      << { <a c>2 s4 } \\ { <b d>4. <b e>4. } >> r4
  }
  \paper {} 
}   

2: Collision avoidance is different in non-chorded vs chorded polyphony
( shown in bars (6 vs 7) compared to (9 vs 10)).  I believe the chord
case should use the same decision as the single-note case.

\score { 
  \notes \relative c'' { 
      time 4/4 r1
      << { a2 s4 } \\ { b4 c d } >> r4
      << { <a c>2 s4 } \\ { <b d>4. <b e>4. } >> r4 \break
% same notes, switch voices
      << { b4 c d } \\ { a2 s4 } >> r4
      << { <b d>4. <b e>4. } \\ { <a c>2 s4 } >> r4 
  }
  \paper {} 
}   


3: Starting a score with polyphony seems to generate unnecessary new
contexts.  (I discovererd this while eliminating the 'r1' at the
beginning of each system, trying to simplify the test cases - not
bothering me, but I thought I'd better report it.)

remove the 'r1' after 4/4 from either of the above.


Thanks
/Hans

lilytest.pdf
Description: Adobe PDF document

lilytest.ly
Description: application/unknown-content-type-ly_auto_file

lilytest.dvi
Description: TeX dvi file