Closed cmyr closed 3 months ago
Is the impact detectable (https://github.com/googlefonts/fontc?tab=readme-ov-file#comparing-branch-performance-with-hyperfine)? - happy to land either regardless, just curious.
no it's not... if there's any benefit I think you'd need to have a parsing-specific benchmark for it to show up. Mostly about ergonomics in any case. :)
Particularly for types that end up being SmolStr later on, like glyph names; it's inconvenient and inefficient to parse to String and then later convert to SmolStr, so this is a minor simplification.
Arguably the whole plist parsing process should be using smolstr, which should offer a reasonable speedup, but we can worry about that when we're thinking more generally about optimization.
...looking at this again, maybe it would make sense to use
GlyphName
whenever we're dealing with a glyph name? that's the type we're going to want later anyway, and it would save us from needing to convert things later...