Some links when they were being checked returned the http code 404,
but the links were working correctly.

This was happening because before the request the link was concatenated
with the / character at the end, making it a different link from the
original. If the original link didn't have a path that is
accessed by / at the end, it would return a 404 error.

This behavior made it a false negative.

Changes:

- It is now possible to only check for duplicate
  links (without checking if the links are working)

- Check for duplicate links when a push occurs

- Update the scripts documentation accordingly

Related:
- #2409
- #2960