Enivex

> But the main point of this function is that it can be used on lengths. Without this, it becomes almost useless as it can be implemented trivially in Typst....

> I've changed this to only supports floats (integers are automatically coerced into floats) and it always returns a float. > > The behaviour now is (as before): > -...

To be clear, the character variant font features should be exposed as an option, just like stylistic sets. Then the user can enable cv1 for new computer modern math if...

> I'm not sure about this. After all, it seems to me that typst want to provide nice "bindings" to symbols, rather than inventing new symbol names. > > So...

> Oh, I thought you are talking about enabling that feature *only for default math font* that has that issue. Sorry for misunderstanding. > > So this is a bad...

I don't agree with that. The reasonable default is the default for the font. Regardless, right now the opentype font features are tied to text, and not the actual font....

I think that rules around how shorthands can be used should be made stricter first. At a minimum they should be required to be surrounded by whitespace (or be alone)....

Another thing to consider is ligatures in popular coding fonts. Ideally they shouldn't conflict. Though I'm not sure that's an achievable goal.

> but won't change anything inherently with shorthands It will, because it'll prevent users from accidentally triggering shorthands (which can be incredibly annoying). Moreover, it'll allow more shorthands to be...

I might add that I'd really like to see an ergonomic way to disable or enable individual shorthands.