I’m implementing the TeX math layout algorithm based on the new OpenType Math Tables (instead of using TeX font metrics). Handling the complexities of the OpenType tables has been the hardest part. I’m still dependent on a Python utility that converts the binary font files into a very helpful XML file, which is much easier to inspect.
Ideally I’d write an OpenType parser direct in Elixir which would be much faster, but (1) I’d need to evaluate maxima and minima for Bézier curves to get vertical metrics (and the python utility does that for me) and (2) parsing the font file can be done at “build-time” by a mix task and the results can be stored for later use.
than 