Issue 5324: \abs-fontsize and set-global-staff-size in books

Issue 5324: \abs-fontsize and set-global-staff-size in books

The font buffering hashtable will provide an outdated LilyPond
scale factor after changing global staff size between books. Mainly
\abs-fontsize concerned, because \fontsize will usually not match
a Pango font-size used in a differently scaled book and fresh fonts
will have to be loaded anyway.

The straight-forward solution is to simply apply #ly:restet-all-fonts
from within #set-global-staff-size.

[dak, 2018-08-13 12:17:29]

Does not help with the other problems we had with fontsize mismatches (out of
multiple calls of set-global-staff-size), right?

What would a correct solution look like?

[Be-3, 2018-08-13 12:58:34]

On 2018/08/13 12:17:29, dak wrote:
> Does not help with the other problems we had with fontsize mismatches (out of
> multiple calls of set-global-staff-size), right?
> 
> What would a correct solution look like?

Hmmm, what were the other fontsize mismatch problems?

The *only* problem this solves is a "successful" try_retrieve that delivers the
wrong (outdated) scale_ factor after
#(set-global-staff-size …).

The other solution I proposed (checking output_scale each time a font is
retrieved from the buffer and eventually re-calculating the scale_ value) would
be safer in so far as it ensures the correct scale_ factor no matter how or why
output_scale has changed (not just by set-global-staff-size).
Would that be of any help? I could provide an alternative patch here.

It'd be good to check the "other fontsize mismatch problems", too, in any case.

Thanks,
Torsten

[lemzwerg, 2018-08-13 16:21:10]

LGTM, thanks!

https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
File input/regression/book-change-global-staffsize-abs-fonts.ly (right):

https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
input/regression/book-change-global-staffsize-abs-fonts.ly:6:
@code{\\abs-fontsize}.
A single-word sentence looks strange...

[Be-3, 2018-08-13 16:46:59]

On 2018/08/13 16:21:10, lemzwerg wrote:
> LGTM, thanks!
> 
>
https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
> File input/regression/book-change-global-staffsize-abs-fonts.ly (right):
> 
>
https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
> input/regression/book-change-global-staffsize-abs-fonts.ly:6:
> @code{\\abs-fontsize}.
> A single-word sentence looks strange...

Right!
(That's a single-word sentence that definitely makes more sense than
"\abs-fontsize."
I'll see to it, thanks.

Apart from that: I just wonder if there's another possibility to change
output_scale other than by #set-global-staff-size.  If so, as DAK implied, we'd
better go for a more comprehensive solution.

Thanks,
Torsten

[dak, 2018-08-13 16:56:51]

On 2018/08/13 16:46:59, Be-3 wrote:
> On 2018/08/13 16:21:10, lemzwerg wrote:
> > LGTM, thanks!
> > 
> >
>
https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
> > File input/regression/book-change-global-staffsize-abs-fonts.ly (right):
> > 
> >
>
https://codereview.appspot.com/341450043/diff/1/input/regression/book-change-...
> > input/regression/book-change-global-staffsize-abs-fonts.ly:6:
> > @code{\\abs-fontsize}.
> > A single-word sentence looks strange...
> 
> Right!
> (That's a single-word sentence that definitely makes more sense than
> "\abs-fontsize."
> I'll see to it, thanks.
> 
> Apart from that: I just wonder if there's another possibility to change
> output_scale other than by #set-global-staff-size.  If so, as DAK implied,
we'd
> better go for a more comprehensive solution.

AFAIR, output-scale is changed in \score markups.  That changes the behavior of
dimensions in those: I implemented an issue to fix problems here but I was
basically flying blind.  Let's see.  There is

commit 4343d45b858e21f30e22a9d899163b35e4579dd1
Author: David Kastrup <dak@gnu.org>
Date:   Tue Aug 12 20:29:22 2014 +0200

    Issue 4065: Don't rescale \layout in \score markup when already scaled
    
    This change modifies the fix for issue 677 to avoid scaling problems
    when employing $(ly:grob-layout grob) in the \layout of a score markup.

and

commit 84cf69f7c840700a83ca6a9ae9e6b6f2e9f9f1f9
Author: David Kastrup <dak@gnu.org>
Date:   Mon Dec 31 18:15:41 2012 +0100

    Issue 677: \score markup confuses paper settings
    
    I actually have no idea whether this is the right fix, but it seems to
    do something related to the TODO comments and the issue report.

There is also the issue that layout-set-staff-size is iffish and does not really
yield the same result as global-set-staff-size.

[Be-3, 2018-08-13 17:16:45]

On 2018/08/13 16:56:51, dak wrote:
> […]

Hi David,

Thanks for the hints, I'll look into issues 4065 and 677.
As far as layout staff sizes, \magnifyStaff, fontSize/staff-space and all the
others are concerned (I've tested many variations back and forth) the results
were (surprisingly) stable.
That's why I used #ly:reset-all-fonts for #set-global-staff-size only.

But I'll definitely check the issues you mentioned.

Cheerio,
Torsten

[Be-3, 2018-08-16 10:43:41]

Correction of single-word sentence in regtest.