Re: minimal-breaking not working as expected

[David Wright]

On Mon 01 May 2017 at 21:31:15 (+0200), Simon Albrecht wrote:
> Am 01.05.2017 um 20:37 schrieb Kieren MacMillan:
> >Hello all,
> >
> >In v2.19.59, I’m using
> >
> >   page-breaking = #ly:minimal-breaking
> >   page-count = 10
> >   systems-per-page = 2
> >
> >… and the output has 16 pages, 14 of which contain one system (2 pages have 
> >2 systems each).
> >
> >As I misunderstanding and/or misusing these features?
> 
> I’ve never used the minimal-breaking algorithm, so this is of
> limited value, but: The default algorithm handles systems-per-page
> with some flexibility, so e.g.
>   system-count = 15
>   systems-per-page = 4
> will likely result in one page having only 3 systems without any
> warning given.
> However, page-count has always been strictly obeyed in my scores, so
> at least that would qualify as a bug, I think.

I find that page-count can be disruptive; ie a score that looks
good on three pages will sometimes, given page-count=3, produce
two relatively slack pages followed by one very tight one.

> I take it that creating a minimal example will be very difficult?

Quite!

Cheers,
David.