Re: missing dots with merge-differently-dotted

[Matthias Kilian]

On Sat, Jan 31, 2004 at 01:14:32PM +0100, Jan Nieuwenhuizen wrote:
> You found a bug.  The dot-collision/merging code was rewritten in 2.1.
> It turned out that merging removed any dots were removed from the
> lower nothead:
[...]
> Fixed in CVS.

HEAD works fine for me. However, I just ran into the next problem (I
hope that's a newbie error rather a bug):

\version "2.1.15"
\score {
  \notes \relative c {
    \time 1/4
    \key d \minor
    \clef bass
    << <cis a' cis>4 \\ {g'8. bes16} >>
  }
  \paper {}
}

This will awfully misplace the dotted g (stem-down) left of the chord
(stem-up), with the dot running straight into the chord. Breitkopf &
Härtel put it right of the chord. It seems that LilyPond isn't aware of
the g's dot when calculating the optimal spacing.

Ciao,
        Kili

ps: may be that Busoni's transcription of Bach's d minor Chaconne is a
good torture test for LilyPond. I've two printouts of that tune (one
from gdr, early 80's, one from brd, mid 90s), both of equally (bad)
quality. If you're interested, I'll send an exemplar to you (or to
Han-Wen). However, LilyPonding it seems to be non-trivial :-/