Behdad Esfahbod

[Skrifa] Minor speedup to iup mechanism

2

PoC: ```diff diff --git a/skrifa/src/outline/glyf/deltas.rs b/skrifa/src/outline/glyf/deltas.rs index c59021a7..426cd86b 100644 --- a/skrifa/src/outline/glyf/deltas.rs +++ b/skrifa/src/outline/glyf/deltas.rs @@ -265,6 +265,18 @@ where } else { D::zeroed() }; + if scale == D::zeroed() { +...

Working from here: https://github.com/harfbuzz/boring-expansion-spec/blob/main/iso_docs/WG03-beyond-64k-glyphs-2024-01e.pdf The changes are a lot more than what's in HB from the original proposal. So I've been slowly (and with help from ChatGPT5 codex) updating to...

Subsetting problems with markFilteringSets

6

This affects both hb-subset and fonttools subset. There are two problems, one is an inefficiency of the final subset, the other is a bug that changes shaping behavior. It was...

In a few places we have: ``` hb_array (font->coords, font->has_nonzero_coords ? font->num_coords : 0), ``` This should go into a function and be reused. Bonus points if we can make...

Same `benchmark-shape` tests, tested against each HB major release would be a good start.

For simpler clients. I'll write my thoughts here over time.

Direction, script, language, font, etc... would be nice to have well-integrated with the rest of the library.

[wasm] Remove font abstraction, do everything in font units?

1

The only downside is that then we cannot expose things like "get_advance", "get_extents", etc, without major reworking of how things work...

See: https://github.com/googlefonts/fontations/issues/1562 Blocks rolling fontations forward right now.