PUA duplicate glyphs & OpenType features

[davelab6]

Filing this based on google/fonts#1743 :)

The private use area character (U+E000) is in the version on fonts.google, so it was just there to avoid a missing glyph on gfregression. It is now set to not export in the source.

I think these tools are still relatively early in development and need interpretation; looking at U+E000 it seems to be empty of contour data, so indeed the error "Glyph name: uniE000 Contours detected: 0 Expected: 1" is helpful because that glyph exists but has no contour data, neither in the current GF version or your source file, so it can just be deleted.

However, the font has a number of PUA encoded characters, which in the TTF on GF today have 'designer' glyph names that have somehow been duplicated along the way to your current source file:

Finally while looking at the Glyphs file, I see that if I click 'update' for Features, I get a 'no effort' additional OpenType feature :)

[davelab6]

I wonder that perhaps some nice custom liga features could be included to use these alts when it makes sense; http://opentypecookbook.com has a good guide to look-ahead rules