Re: Spacing problems with 2.9.21 (as opposed to 2.8)

[Joe Neeman]

On Sat, 2006-10-14 at 01:46 -0400, Marcus Macauley wrote:
> Han-Wen Nienhuys wrote:
> 
> > Marcus Macauley schreef:
> >
> >> So, Joe, whatever it is that you fixed, there still seems to be a  
> >> problem [with vertical spacing] in 2.9.22.
> >>  I can try to send an example if necessary.
> >
> > I think this went in for 2.9.23
> 
> Actually, now in 2.9.23-1, it seems to be even worse than 2.9.22, though  
> in the other direction.
> 
> At least judging from one score I'm working on, 2.9.17 was fine, 2.9.22  
> was way too spread out (vertically and horizontally), and 2.9.23-1 is way  
> too squished together (vertically and horizontally).
If there aren't any copyright issues, I'd be interested in seeing this
score.

> My score now has one system the right side of which goes beyond the margin  
> and literally off the page;

This one is puzzling. I can't think of an explanation off the top of my
head.

> another system, its bottom staff has been  
> pushed down entirely off the bottom of one page.

This is probably because Lily is underestimating the height of the
staves, so it thinks it can fit more systems on a page than it really
can. A workaround is to increase between-system-padding.

> 
> Setting ragged-last, ragged-bottom, and ragged-last-bottom to ##f in the  
> \paper block seems to trade these problems for others, slightly less  
> severe in this case, but still involving a staff pushed off the page.
> 
> And setting #(define page-breaking optimal-page-breaks) in the \paper  
> block, to disable the new page-breaking, seems to fix the vertical spacing  
> problems, but it still leaves the horizontal spacing somewhat more spread  
> out than it was in 2.9.17.

This would be because of the compression penalties (discussed earlier on
lilypond-devel). I suppose I should make it configurable, so it can at
least be overridden to the old behaviour.

Once again, thanks for the report,
Joe