Open rsheeter opened 1 year ago
We have a few options: one simple one would be to have a GlyphId
trait, which would be implemented for GlyphId
, GlyphId24
, and GlyphId32
. Alternatively we could convert all glyph ids to a single type (u32) when we read them.
For reading being able to just always have u32 seems fine. Presumably the compiler needs direct control over width. If we u32 on read would that impede the compiler?
nope, codegen would know the underlying types and so would generate the correct representation for compilation.
Just want to be sure we aren't painting ourselves into a corner.