[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Figured bass bug in 2.23.82?

From: Jean Abou Samra
Subject: Re: Figured bass bug in 2.23.82?
Date: Mon, 12 Dec 2022 23:59:16 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.5.1

Le 12/12/2022 à 23:33, Lukas-Fabian Moser a écrit :
Yes, I kept a backup. I attach the .ly file and the ps file generated by 2.23.10.

Thanks. I think this finishes to explain what is happening.

In the big blob of data that embeds the emmentaler-11 font in the PS file, there is the string "2.23.7" somewhere, next to licensing info. Apparently, Fontconfig is preferring that font file from 2.23.7 under ~/.local for glyphs that exist in it (numbers), but it falls back to the good one for glyphs that 2.23.7 doesn't contain, the accidentals. This results in a mix of font files, as we have seen in the stencil expression dumps you posted.

Then, the two font files, one from 2.23.7 and the good one from 2.23.10, are conflicting under the name "Emmentaler 11" for being embedded in the PS file. Depending on how the input file is made exactly, either one could get to it first. In this case, the 2.23.7 one gets embedded. The 2.23.10 one doesn't because LilyPond apparently thinks it has already embedded that font. Actually, scm/framework-ps.scm line 256 tries to display a warning in this case, so we need to investigate why this warning isn't triggering.

Then the accidental glyphs aren't found by Ghostscript, leading to these squares.

So we need to figure out a solution for this font finding issue.

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]