Re: Horizontal spacing change between 2.15.9 and 2.15.10 (regression?)

[address@hidden]

On Oct 24, 2011, at 8:44 PM, Xavier Scheuer wrote:

> Dear developers,
> 
> I compiled yesterday (with 2.15.14) a file that was first compiled with
> 2.15.5.  I was surprised to see that the default line breaks were quite
> different and some measures were really ugly (speaking of horizontal
> spacing) with the latest version!
> 
> I have a passage with many 32th notes (and subdivide beams).  With
> 2.15.5 (as well as with 2.14.2) LilyPond settles 2 measures per line,
> but with 2.15.14 it puts only one measure per line and this measure is
> really stretched horizontally (whereas it was only "a bit compressed"
> when there were 2 measures per line), leading to a poor output.
> 
> It is hard to give a minimal example of code.
> I tracked down this change and it happened between 2.15.9 and 2.15.10.
> 
> 1. Am I the only one to have seen such a change in horizontal spacing
> between scores compiled with version <2.15.9 and the same files
> compiled with a version >2.15.10 ?
> Does some users see this change as well?  Was it enlightened in some
> regression tests?
> 
> 2. Is this change wanted?  I mean, is it kind of a feature that should
> give a better horizontal spacing and that does it for many scores
> except the one I am currently working on?  Or is it unintentional (and
> maybe it could be interesting to find the commit responsible for this
> change and see if we could get back to the behaviour where some scores
> (mine) look better)?
> 
> Thanks in advance.
> 
> Cheers,
> Xavier

Hey Xavier,

If you use git bisect, you can track it down to an individual commit.  This'll 
help us figure out how to fix the problem.

Cheers,
MS