Laurenz

Thanks for the PR! Unfortunately, I must say that this seems a bit over-engineered overall (also indicated by the relatively large additive diff). For instance, the `VersionError` seems unnecessary as...

I made some changes: - I removed `typst::version()` in favor of directly using `typst_utils::version` (or `typst::utils::version` via re-export). It's important that the build script of `typst-cli` does not have a...

I think the fact that there are various different interpretations of what this API should do exactly does pose the question of whether we should add this in the first...

I don't want to rule out the possiblity of some sort of map function in the future (there was e.g. the proposal of `map-values` and `map-pairs`), but a bare `map`...

Maybe ` $ text(font: #(), emptyset) $` instead? That should have a similar effect. But it might warn as well in the future.

Looks like all crashes are related to math. We need to add a check in the ad-hoc math styling code, too.

Prior discussion: https://github.com/typst/typst/issues/492

I agree that the current shorthand list is not most consistent. However, I think an effort to improve the selection could then perhaps not try to fix this by exclusively...

> Though I think we would need to implement a way for users to define their own shorthands before any removals happen to prevent breaking existing documents. I don't think...

There is no current work. Supporting variable fonts is a major effort. It needs large changes to the font handling and to the PDF export.