Re: lilypond-book turns longas below the staff into breves

[Patrick McCarty]

On Fri, Aug 07, 2009 at 12:26:42AM +0200, Francisco Vila wrote:
> >
> 2009/8/6 Laura Conrad <address@hidden>:
> > It's scary if you play the wrong note even when you took all that care
> > to enter the right note.
> 
> A clue for our brave developers: The attached image made from
> 
>   \stemDown c4 c\longa
> 
> shows that the skyline mechanism does not consider the longa' stem
> worth to be skylined.

Thanks, Francisco, I've started looking into this.

The root problem lies in various bounding-box calculations in the Metafont
code (search for `set_char_box').

The Parmesan font glyphs (the \longa glyph is one example) often use
symmetrical lengths for their Y-extent -- for example, #(-0.5 . 0.5).

This sometimes seems appropriate, but it often seems grossly
inaccurate, as you found out by debugging the skylines.

Once I digest more of the Metafont code, I should be able to fix the
bounding boxes.  Unless, of course, the bounding boxes *need* to be
inaccurate for some reason.  :-)


Thanks,
Patrick