Attaching marks to ligatures, but not mark-to-ligature

[simoncozens]

In Noto, I keep beating my head against googlefonts/glyphsLib#756. Basically, the issue is this:

• Someone designs a conjunct glyph of a-somescript and b-somescript and calls it a_b-somescript.
• It has a mark attachment anchor bottom.
• glyphsLib, being glyphsLib, assumes that because the glyph name is a combination of two other glyphs names, it's a Ligature glyph.
• Because it's a Ligature it ends up in the GDEF "ligature" glyph class.
• The mark feature writer looks at this glyph, notices it's a ligature, looks for "bottom_1", doesn't find it, and doesn't write anything.
• The mark doesn't attach.

I can get around it by setting all these glyphs manually to being bases (or alternatively by renaming all my anchors), but if I have a bottom anchor and some marks have a _bottom anchor, that means I want them to attach together. I think it should be fine to match these anchors and issue mark-to-base rules for these glyphs if no _1 etc anchors are present (heck, even if they are), even if the base glyph is actually in the ligature glyph class definition.

Does this make sense to others?

[anthrotype]

hm but then what's the point of marking a glyph as a Ligature in GDEF but then effectively treating it as a Base for mark feature, by attaching a mark-to-base rule?

[schriftgestalt]

Again, this would be fixed with the code in the GlyphsInfo3 branch.

[simoncozens]

How? This is a ufo2ft issue.