Colin Rofls

The last one is a good reminder, there are a number of encodings we do not yet support for name records. That should have an issue. The others are mildly...

what format are your glyphs in initially? Assuming you have outlines in some raw format (i.e. not from an existing font) this is not a trivial task, and amounts to...

ah okay good catch!

That's actually an interesting complication for codegen, this is the first time where we will have a lower version of a table (2.0) that has more fields than a higher...

I just remembered one of the concrete motivations for this idea: in several places we need to choose between two different formats for a given subtable (specifically for contextual lookups)...

correct, this method would produce an 'upper bound' on size, since it could not perform deduplication. However for many of the cases that we care about I think duplicate objects...

The only planned serialization support would be to write tables out in the binary format provided by the spec, using write-fonts. What sort of API would you need, for things...

support for fields conditional on a flag is added in #937

One note is that the docs for `base_anchor_offsets` (referenced here) do actually include the info we want: https://docs.rs/read-fonts/latest/read_fonts/tables/gpos/struct.BaseRecord.html#method.base_anchor_offsets the methods like `base_anchors` were added later, to try and simplify things,...

An annoying thing here is that we now have two different special-cased records that are special in different ways. - I'm fine with deriving `Copy` for all records that can...