In GlyphData.xml
I notice that the Arabic characters U+FBB2…U+FBC1 are given a subCategory="Nonspacing"
The Unicode Standard in section 9.2 says specifically:
The range of characters U+FBB2…U+FBC1 provides a set of symbols for this purpose. These are ordinary, spacing symbols with right-to-left directionality. They are not combining marks, …
I understand the need for zero-width versions of nukta glyphs for the purpose of building up composite characters, but I question that such should be encoded as characters that Unicode says are “ordinary, spacing symbols”.
It seems to me that FBB2…FBC1 should be subCategory="Spacing"
similar to characters like U+0060 GRAVE ACCENT"
The Unicode Standard goes on to say:
and are not intended for the construction of new Arabic letters by use in combining character sequences.
so it seems to me that whatever is encoded at FBB2…FBC1 should not have anchors such as _top
or _bottom
since having those invites (perhaps enables) users of the font to try to make new Arabic letters using them.
Bob