[css-pseudo] `::after-letter(an+b)` or `::before-char()`

[Crissov]

There have been several proposals for text or character matching pseudo-elements, e.g. #3208, which is computationally expensive, in the past, but as far as I can see there is no proposal yet for a simpler pseudo-element that is inserted before or after the nth ”letter“ or rather character.

One (weak) use case for this are indicators for boundaries, e.g. where you would need to split a text to fit it into, e.g., 140-, 160- or 280-char messages.

[Loirooriol]

What if the element contains a mix of text and children elements? Would you keep counting letters recursively into the children and insert there? What style would it use then, the one from the pseudo-element, the one from the nested element, or a broken mix like ::first-line?

Also, I suspect authors will commonly want to skip trimmed whitespace rather than just counting unicode code points. And possibly count combined characters as a single one?

[Crissov]

Those are all valid points, of course, since this is not a fully thought-through proposal but more of a shot from the hip so the idea gets tracked at all.

I guess this should be somewhat aligned with #2208 ::text.

I think the distinction between letter and char or character would mostly be just that: ignore whitespace or don't.