@Leander Tentrup I was wondering if we could go over https://github.com/rust-analyzer/rust-analyzer/pull/4940#issuecomment-646690389. I didn't fully understand your comment, and what I attempted didn't really produce any noticeable changes (besides one producing something clearly wrong). For clarification, what do you mean by "change" when you say "the type of the splitted parent tokens prev[idx] and cloned have to be changed". Also, when you say "and
delete has to be changed to the overwriting semantic type", do you mean I need to keep what was previously deleted and add the overwriting semantic type (in other words, "injected" in this case?). When I attempted to keep the deleted portion for the doc tests, the doc comment ended up duplicated. If I removed the
HighlightTag::Operator then the original highlighting came back, but the elements marked
None had no highlighting at all. I couldn't figure out how to reconcile the two.