add-font adds spaces to ABeeZee font name

[thundernixon]

I'm using add-fonts to generate new metadata for all families in google/fonts, in order to make sure we don't exclude latin-ext glyphs where they exist (see google/fonts#187). This has led me to find a number of issues with add-font. I'll document them separately, so that we can hopefully correct things that are real issues, and I can be better informed if some of them aren't real issues.

---

If add-font is run on ABeeZee, the METADATA.pb file ends up with

name: "A Bee Zee"

However, in the name table of ABeeZee, nameID 1 is

    <namerecord nameID="1" platformID="1" platEncID="0" langID="0x0" unicode="True">
      ABeeZee
    </namerecord>

(without spaces)

So, we may need to add an exception for ABeeZee, unless this is really the intended behavior.

[felipesanches]

fontbakery does have exceptions for this kind of thing. Maybe we can copy the exceptions into gftools add-font as well

[felipesanches]

https://github.com/googlefonts/fontbakery/blob/d2cb61bf4c1f97b03f437a0f2da3d454c8c9b6c9/Lib/fontbakery/specifications/googlefonts.py#L2838-L2851

[thundernixon]

That's pretty much what I had in mind! I thought it actually did exist in add-font, but that must have been in another tool.

[m4rc1e]

I don't like this camelcase splitting function. I also use it in GFR. Couldn't we get the family name in the following manner?

nametable = font['name']

family_name = None
typographic_name = nametable.getName(16, 3, 1, 1033)
if typographic_name:
    family_name = typographic_name.toUnicode()
else:
    family_name = nametable.getName(1, 3, 1, 1033).toUnicode()

[m4rc1e]

The above method does depend on the fonts having good name tables though? so maybe the filename may be a better indicator of the family name. Anyone willing to test this across our collection?

[thundernixon]

I think I will probably need to check name changes anyway, so I'd be happy to try this.

What is nameID 1033? I don't see it in the MS spec, and I don't recall coming across it before.

[m4rc1e]

it's the langID of the record.