it is a bit of an overhead to call the whole parser to just parse the path of the include

[rexim]

introduced in #225

It would be better to extract specifically string parsing into a separate function and call it directly.