Dear mutex subscribers.
During the typesetting of one of the scores in my GMD entry (Miraklet,
http://www.gmd.de/Misc/Music/scores/mondrup/mon.gif) I have experienced
a strange difference between the dvi output generated by MiKTeX (win32)
and TeTeX(linux) on the other hand.
I have made a condensed one bar excerpt from the score to illustrate the
problem. The musixtex result generated by mtx v. 0.52 and the not yet
released v. 1.42 of pmx goes like this:
\immediate\write10{M-Tx 0.52 (Music from TeXt) <29 October 1998>}
\let\:=\relax\input musixtex\:\sepbarrules\input pmx
\input musixtex
\input pmx
\input musixmad
\normalmusicsize%
\nopagenumbers
\tracingstats=2\relax
\hsize=524pt
\vsize740pt
\def\nbinstruments{1}
\setstaffs12
\setclef1{60}
\setname1{ }
\generalsignature{ 0}%
\generalmeter{\meterfrac{9}{8}}%
\parindent 0pt
\elemskip1pt\afterruleskip1.000pt\beforeruleskip0pt\relax
\stafftopmarg0pt\staffbotmarg5\Interligne\interstaff{10}\relax
\readmod{test}
\startmuflex\startpiece\addspace\afterruleskip%
\advance\barno by -1
\znotes|\zcharnote{16}{\titles{2.0}{}{0}{}{0}{}{0}}\en%
% Bar count 0
\def\Notes{\vnotes1.41\elemskip}%
\def\NOtes{\vnotes2.00\elemskip}%
\def\NOtesp{\vnotes2.45\elemskip}%
\def\NOTes{\vnotes2.83\elemskip}%
\NOtes\cu{'A}|\cl a\nextvoice\bigna c\cu c\en%
% Bar count 1
\xbar
\NOtesp\qup G|\roff{\ql b}\nextvoice\ibu3b1%
\makeatletter\def\C@Point#1#2{\PMXpt{1.0}{1.0}0}\makeatother\qbp3b\en%
\Notes|\nextvoice\tbbu3\qb3c\en%
\NOtes|\cl b\nextvoice\tbu3\qb3d\en%
\NOTes\qu{'C}|\ql c\nextvoice\isu8d{.8}\qup e\en%
\NOtes\cu{'B}|\cl d\nextvoice\en%
\NOTes\qu{'A}|\ql c\nextvoice\tslur8e\qu e\en%
\NOtes\ds|\lower4\internote\ds\nextvoice\raise2\internote\ds\en%
\endpiece
\vfill\eject\endmuflex
\bye
Now if I let MiKTeX process this file I get - besides the right- and
upward adjustment of the dot in the first upper righthand note of bar 1
- 32'th flags on the stem of that note. If I remove the dot adjustment
directives the flags disappear.
I don't get these flags at all, however, if I let linux TeTeX process
the musixtex code. But as I normally do my typesetting from win32 I find
the problem interesting enough to post it here.
You may look at the two different dvi outputs in the attached zip file.
Will some TeX guru mutex subscriber please explain what goes on here :-)
Regards
--
Christian Mondrup, Computer Programmer
Scandiatransplant, Skejby Hospital, University Hospital of Aarhus
Brendstrupgaardsvej, DK 8200 Aarhus N, Denmark
Phone: +45 89 49 53 01, Telefax: +45 89 49 60 07
dvi.zip
