For reference, the original Adobe Glyph List is maintained at https://github.com/adobe-type-tools/agl-aglfn/blob/master/glyphlist.txt. It looks like psaltnames just picks the first code point from AGL, no matter what.

Looks good to me.

@dscorbett, did you use some sed-like script for this or just updated everything manually? There are some other names in the list which look dubious, e.g. "Ifraktur" is called "Ifractur".

I probably used some kind of automation but I don’t remember the exact commands.

Ifractur appears under the comment “From Richard Kinch's TeX list of glyph aliases” and is attested in various glyph name lists. I think it’s fine to keep old names, as long as they aren’t incorrect like the ones I removed here.

The issue of Ifractur vs Ifraktur bothers me because Ifraktur does appear in AGL, but is somehow missing altogether from psaltnames. If you can recall your automation, it would be great to add it as a comment in the code. I can re-develop it if you are short on time at the moment.

Overall, your change is definitely in place, but there are a lot of interdependent lists here, and their logic is [for me] very elusive.

Ifraktur appears in the agl_sans_p0_b21 array, which is part of the NameList-based system of glyph names, where each glyph name standard (e.g. AGL) gets an array of arrays of glyph names, with one string per code point (or NULL if there is no name for that code point). psaltnames, on the other hand, is a miscellaneous collection of alternative names like Ifractur that don’t fit elsewhere.

grep -v ';....$' glyphlist.txt | grep -v ^#
produces 81 lines, corresponding to the number of removed lines in this PR.

I checked the other names in AGL, and they are all present in namelist.c in some place, so I think we are good to go.

@dscorbett, could you please merge with the latest to trigger the CI?