New Check: STAT table has Axis Values

[m4rc1e]

Rosalie recently made a pull request for a VF family which didn't contain any STAT Axis Values. By default fontmake will export a STAT table but won't include any meaningful Axis Values. We need to make a check that will fail these fonts.

cc @RosaWagner

[RosaWagner]

Few other possible related STAT checks:
(some may be already implemented)

<- weight axis ->

• Regular is correctly linked to bold (when there are both styles)
• Regular is elided
• Values are correct according to axis registry
• There is a value every 100

<- italic axis ->

• There is an ital axis in both upright and italic fonts (when there is an italic font)
• Upright : "Roman"
• "Roman"=0
• "Roman" is elided
• "Roman" is linked to 1 ("Italic")
• Italic: "Italic"
• "Italic"=1

[chrissimpkins]

By default fontmake will export a STAT table but won't include any meaningful Axis Values

Is that a better default than not adding a STAT table at all? I've wondered about this approach in the past too.

[davelab6]

This is now urgent @felipesanches :)

[miguelsousa]

@felipesanches Have you started working on this? If you haven't, please assign it to me.

[felipesanches]

Feel free to got for it! I have way too much on my plate now, so if you see issues in which I reasigned release milestone multiple times in succession, it simply means it is likely beyond my ability to address in time, so it is safe to request ownership of it for yourself if you want to work on them.

[felipesanches]

thanks, @miguelsousa!

[aaronbell]

I have a question about this check. In a project I'm working on, I have an fvar location at 5 that is being flagged as "missing Axis Value for 'opsz' value".

However, the STAT table has a rangeMinValue set at 0, and rangeMaxValue set at 16 (and a nominalValue at 11). Wouldn't that indicate that the 5 location is covered by the STAT table? As far as I can tell, the check that @miguelsousa implemented only covers the nominalValue (or value) entry and ignores the min/max range.

[miguelsousa]

@aaronbell I think you're right. It seems like format 2 is not being handled as well as it needs to.

[felipesanches]

reopening this one to track progress. Thanks, @aaronbell & @miguelsousa