Re: LilyPond 2.23.7 released

[Jonas Hahnfeld]

On Fri, 2022-04-01 at 17:35 +0200, Jean Abou Samra wrote:
> Le 26/03/2022 à 22:36, Jonas Hahnfeld via LilyPond user discussion a écrit :
> > We are happy to announce the release of LilyPond 2.23.7. This is termed
> > a development release, but these are usually reliable. If you want to
> > use the current stable version of LilyPond, we recommend using the
> > 2.22.2 version.
> > 
> > Starting with this release, LilyPond requires Guile 2.2 and the
> > official binaries were created with the new infrastructure developed
> > over the past months. Issues reported for the previous release have
> > been addressed, in particular regarding performance (by integrating
> > compiled bytecode) and on Windows, where it is now possible to extract
> > the provided zip archive with the Windows Explorer and use special
> > characters in filenames. Please test this release and let us know about
> > problems that you encounter.
> 
> On the French list, Martial Rameaux reported that this snippet no
> longer worked:
> 
> https://lsr.di.unimi.it/LSR/Item?id=658
> 
> 
> \version "2.23.6"
> 
> %%postscript tuning name
> EADGBE =
> \markup {
>    \with-dimensions #'(0 . 0.8) #'(0 . 1.0)
>    \postscript #"/Arial-Bold findfont
>                  1.3 scalefont
>                  setfont 0 3.6 moveto
>                  (E) show 0 2.0 moveto
>                  (B) show 0 0.6 moveto
>                  (G) show 0 -0.8 moveto
>                  (D) show 0 -2.2 moveto
>                  (A) show 0 -3.6 moveto
>                  (E) show
>                  stroke"
> }
> %%------------------------------------------------
> \new TabStaff {
>    \set TabStaff.instrumentName = \markup { " " \EADGBE }
>    \set TabStaff.shortInstrumentName = \markup \EADGBE
>    e,8 c e gis b e'
> }
> 
> 
> I can reproduce on my systems with all binaries created with the new
> systems (including those of 2.23.6 and the experimental ones of 2.23.5),
> but it works with those from GUB. On the other hand, it also
> works with my local build.

Yes, I can reproduce this. The reason is that your system GS does the
following:

> Can't find (or can't open) font file Arial-Bold.
> Didn't find this font on the system!
> Substituting font Helvetica-Bold for Arial-Bold.
> Loading NimbusSans-Bold font from 
> /usr/share/ghostscript/9.54.0/Resource/Font/NimbusSans-Bold...

ie Ghostscript is loading its own font, whereas

> Can't find (or can't open) font file Arial-Bold.
> Didn't find this font on the system!
> Substituting font Helvetica-Bold for Arial-Bold.
> Can't find (or can't open) font file %rom%Resource/Font/%rom%Resource/F.
> Can't find (or can't open) font file NimbusSans-Bold.
> Can't find (or can't open) font file %rom%Resource/Font/%rom%Resource/F.
> Can't find (or can't open) font file NimbusSans-Bold.
> Didn't find this font on the system!

we don't ship the fonts twice (GUB did). The right fix is not to
reference fonts via \postscript and then letting Ghostscript figure out
where to find them.

> The GS version is not the problem, as it is the same between these
> two cases.

Not quite, our Ghostscript is stripped down to not include unnecessary
stuff and that is not prepared for users doing stuff with \postscript.

> Relatedly, another user reported that system fonts didn't load anymore
> (but didn't provide details yet).

If system fonts don't work anymore via LilyPond, that would be worth of
a bug report.

Jonas

signature.asc
Description: This is a digitally signed message part