Laurenz

This is intentional, but you can get rid of it by putting `#h(0pt, weak: true)` into the template, after the dash.

> Hmmm, then I guess the template should be updated. Fair enough.

So, you propose to not strip leading whitespace (e.g. `[ hello]`), but if there is a leading linebreak, strip that + following indent)?

> I personally think all leading whitespace should be stripped. That would have problematic consequences: For instance, `#([Name1], [Name2]).join[ and ]` and `[Hello] + [ ] + [World!]` wouldn't work...

> This would improve the UX for this specific case, but I'm not sure it's worth it. Or at least, having different behavior based on whitespace feels wrong. My thoughts,...

> Instead of [Hello] + [ ] + [World!] you can use "Hello"+" "+"World!" and instead of ([Name1], [Name2]).join[ and ] you could use ("Name1", "Name2").join(" and "). Also I...

Because only actual Unicode symbols and syntax shorthands participate in the auto-matching and auto-scaling. Symbol notation and escape sequence don't, without explicit `lr`.

Automatic fence matching was removed in https://github.com/typst/typst/commit/fe402759c03eb93cefc2879e751b8e732891bd3e because of false positives like this.

Spreadsheet functionality is out of the scope for the compiler, but can be implemented in user code. Could be a package.

> column footnotes If the whole page consists of columns, the footnotes are automatically in the columns. I doesn't work right now if there's a one-column header row, but I...