As of now, when a regex like #"(^|\n)#[^#](.*\n)*\n#[^#]"# is specified which matches whitespaces or some kind of prefix first, where the actual problematic part lies somewhere in the middle of the matched string, the violation is still reported at the beginning of the matched string. This should be configurable, for example by providing a specific capture group name like pointer reserved for the violation location.

Also, we could consider to support multiple violating locations that are related, this way lint checks could for example point out duplicate code showing both (or all) locations of the duplicated code.