Fixed VariationAxis.hidden being initialized with garbage data

[SamusAranX]

In fvar.rs, the hidden flag is currently read like this:

hidden: (flags >> 3) & 1 == 1,

According to the docs linked in that file, the bitshift here is wrong, and a simple flags & 1 == 1 would do.
If you have access to a Mac, you can verify this by checking the hidden GRAD axis in /System/Library/Fonts/SFNS.ttf.
(I'd point to another, more freely accessible font, but I'm afraid I'm not aware of any with hidden axes, sorry)

[asibahi]

Thanks for this. I fixed this in my Zig port, but I do wonder where the >> 3 comes from. It seems that it used to be auto-generated somehow.

I wonder if other font parser exhibit the same behaviour.

[SamusAranX]

Autogenerated how, though? The original commit is from October 2021, which is well before coding agents etc were mainstream.

[asibahi]

Python scripts. They’re all the rage in Unicode libraries

[SamusAranX]

Point taken. Wish I knew where the bitshift came from though.