I have entries for unencoded stacked diacritics in my custom GlyphData.xml file. Each entry has a sortName field.
I also have a set of standard encoded .narrow diacritics, and they too have a sortName field.
In the font window, what I am finding is that an unencoded character with a .narrow name and a sortName entry which is in sequence with the standard group of encoded .narrow diacritics will not sort with the standard group but always groups with all the other unencoded diacritics.
Since all these unencoded characters are pushed to the end of the Marks section, I’m wondering if G2 is treating them as ligatures because each has an underscore in the name, following Adobe’s naming. Is there a way to get around this default behavior?
What are you naming them exactly? Can you paste the glyph names here?
dieresiscomb_tildecomb.narrow is an example of the naming style, following the style of Adobe Latin 5 except they are using Unicode rather than Nice Names.
However, I did more testing and during one test using the default GlyphData.xml file, omitting mine from the equation, the stacked accents appear in quite a different order, which I expected – the unencoded stacked glyphs that I have in my character set are mixed in with their non-stacked counterparts, which is also what would be expected.
So my original question likely has nothing to do with how Glyphs is sorting them and more to do with how I have my custom data file set up, and now that I know that, it’s perfectly OK. The two .narrow glyphs that I thought were appearing in the wrong place are actually appearing in the group they belong to, which is the stacked accents. I don’t know exactly how this is happening, but it is OK.