luukvbaal

> Not sure if been discussed already but `conceal_lines` makes more sense to me than `conceal_line`. Hasn't been discussed. Why? Because it will hide multiple lines if the range spans...

`virt_lines` itself can take an array of lines so there the plural form makes sense. Setting a `conceal_lines` property for a singular `start_row` seems wrong to me. Setting `end_row` implies...

We could instead have the on_win callback (that now already iterates from topline > botline), store captures/metadata per line, and have the on_line callback iterate that. Would that be a...

Reduces some code duplication (~30 LOC less) and double calls to `iter()`, `get_range()`, and `unpack4()`. My assumption was that it would be more performant, but if those functions are cheap,...

So yeah it's about twice as fast, even though multiline marks need to be stored in each line it spans. Otherwise multiline highlights are only applied to the start line....

Thanks for noticing that, I'll look into it.

> If there's a slowdown from running the highlighting query twice, we can try moving everything that needs to run for the entire window into a separate query file to...

> There's currently an issue with lines not being concealed if they weren't visible before other lines were concealed. Resolved this by just extending `erow` whenever we place a `conceal_lines`...

In addition to the new `_on_decor_line` decoration provider callback, I painstakingly went through all instances of `has(Any)Folding()` to make sure the logic handles concealed lines properly. Unless deemed necessary I...

I've just been rebasing this until someone finds the time/courage to review this. I think this works well enough, though I believe scrolling doesn't always work as expected (with regards...