Consider not showing small header delimiter if columnTypes = false in DataFrame.renderToString

It was recently introduced along with switching default columnTypes = true because two-row header is easier to confuse with data, so some small distinctive indicator is needed. Maybe it's better to only render it when there's type?
`println(df.renderToString(columnTypes = false))`
Actual:
```
name age city
---- ---- ----
Alice 30 Berlin
Bob 25 Paris
Charlie 35 London
```

Expected:
```
name age city
Alice 30 Berlin
Bob 25 Paris
Charlie 35 London
```

IMO it's not worth adding an extra parameter dedicated just for it because it somewhat conflicts with border parameter already. If we don't agree - let's just keep as is.

Start at the DataFrame.renderToString entry point and reproduce the shown output with columnTypes = false. Update the rendering behavior so the small header delimiter is absent when columnTypes is false, then verify that rendering with columnTypes enabled still retains the delimiter and matches the intended output.

Written by the indexing model from the issue text.