The error/warning parser of the command execution
already sees the underscore as an alphanumeric
character, because a lot of filenames contain it. In my
projects, there are also a lot of files that contain a
dash. So i have added them to the list of
"alphanumerical" characters. Don't know if it's useful
but i thought i'd just submit it.
Also, the last part of the diff causes the parser to
only parse the filenames that are at the beginning of a
line (i got a lot of false positives in lines
containing ':' here).