The only purpose, according to the README and existing files, is to
parse an integer from those lines, so (\d+) is sufficient for that. This
avoids potential exponential backtracking as flagged by CodeQL. I have
compared the output of the script with and without these changes and the
resulting files are the same.