On Sun, 2011-07-17 at 15:10 -0700, Don Simons wrote: > I've just uploaded a bugfix beta version 2.614 of PMX to > > http://www.icking-music-archive.org/software/pmx/pmx2614.zip
Thank you. The results look good. > > This fixes a bug discovered by Terry Enger. It was quite a remarkable catch, > and reminded me of the PMX bugs that Werner was so adept at finding. It > depended on having all the following conditions: A (1) rest that was (2) > lowered and (3) dotted, followed by (4) a single note, followed by (5) two > beamed notes, with (6) all of the above occurring in the same notes group, > i.e., without any change in horizontal spacing increment \noteskip! Admittedly it seems a mite strange to put an \hbox in an \hbox just so that you can lower it, something that you might expect from the department of extra redundancy department. Still, is it so very strange that it gives the tex macros (musixtex.tex?, pmx.tex?, whatever!) an excuse to supply flags+beam to the 'cello G? I was expecting the fix to be somewhere in the tex macro definitions. Thank you for waiting; I am back <grin />. Experimentation with tex source code shows that neither \lower nor a nested \hbox is necessary to show the problem. Rather, I think that the \hbox is occupying space that \tbl does not account for when it lays down the start of the beams. An \hbox out at the left moves the start of the beam to the left of where it belongs; an \box after the first beamed note moves the start of the beam to the right of where it belongs. More precisely, I think that those \hbox's move the notes in the other direction from what I wrote. The effect is easier to see with an \hbox of a different width. The width \Noteskip creates a strong suggestion that the start of the beam is associated with the stem it abuts; that is just coincidence. Well, that is this newbie's guess, for what it is worth. > I still > don't understand precisely what was happening, but it was caused by PMX > putting the rest and dot together in an hbox, when the rest itself was > already automatically in an hbox. The basic error would be seen in the first > bar of the following file, which also shows how the revised algorithm tweaks > dot heights depending on whether the dot would hit a staff line. Neat! Thank you. > (snip) ------------------------------- [email protected] mailing list If you want to unsubscribe or look at the archives, go to http://tug.org/mailman/listinfo/tex-music
