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

[Xavier Scheuer]

On 24 October 2011 22:13, address@hidden <address@hidden> wrote:
>
> 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.

Hey,

Sorry for late follow-up.

I used git bisect between 2.15.9 and 2.15.10 (following instructions in
the CG).  It gave me  f7278543a3ca65ee239696c21b4a8d0155bb79db
Merge remote branch 'origin' into release/unstable
as the first bad commit.
But I must mention that in the lasts steps I got no bar lines any more,
so I considered them as good because I had 2 measures per line, but
maybe if the bar lines were present I would have had only 1 measure per
line and would have considered it as bad!

I also get, with the last one when I compiled the file, this error:
programming error: Improbable offset for stencil: inf staff space
Setting to zero.
continuing, cross fingers
programming error: Improbable offset for stencil: inf staff space
Setting to zero.
continuing, cross fingers


On 25 October 2011 02:00, Keith OHara <address@hidden> wrote:
>
> Xavier,
> Possibly I did this.  Sometime in 2007, the space after a key-signature
> or time-signature or clef became very compressible, because some spacing
> parameters were no longer respected (issue 1785).  I restored the
> enforcement of this space for version 2.15.9 with Janek's help to
> adjust parameters so that scores looked nice.
>
> A side effect of that fix was issue 1856: the space after time signatures
> could now stretch, which made some highly-stretched regression tests
> look different.  The fix to issue 1856 made the space at the beginning of
> each line perfectly rigid, and appeared in 2.15.10.
>
> Does the override below restore your scores?
>
> \override Score.KeySignature #'space-alist #'first-note
>  = #'(semi-fixed-space . 2.5)
> \override Score.Clef #'space-alist #'first-note
>  = #'(semi-fixed-space . 2.5)
> \override Score.TimeSignature #'space-alist #'first-note
>  = #'(semi-fixed-space . 2.0)

Hi Kieren,

I think you find the good spot.
It did restore my scores.  Actually  KeySignature #'space-alist
#'first-note = #'(semi-fixed-space . 2.5)  is sufficient (the other
ones do not show such a significant change).

But I do not understand why semi-fixed-space restores my scores,
because these scores look good with 2.14.2 (which has fixed-space,
isn't it?)!!  Maybe it is the conjunction of different changes in the
horizontal spacing that lead to a bigger change in my particular
scores, what do you think?

Thank you for your interest.  Please let me know if I can do something
about this.

Cheers,
Xavier

-- 
Xavier Scheuer <address@hidden>