Best practices for homepage_url in rules

Currently the `homepage_url` attribute in scancode rules can point to an arbitrary URL, but typically is pointing to some Git URL. This Git URL often points to the `main` or `master` branch, which is a moving target, as it can change. I would suggest to instead require that `homepage_url` points to a specific commit to prevent things no longer being in sync.

An example:

https://scancode-licensedb.aboutcode.org/bear-blog-2.0.html

`homepage_url` points to https://github.com/HermanMartinus/bearblog/blob/master/LICENSE.md but could also have pointed to: https://github.com/HermanMartinus/bearblog/blob/998e87263248d21c8c5b967de0047b0cdf4dc11c/LICENSE.md

This URL, unless the commit is deleted, would never change, whereas the current one (pointing to `master`) actually *can* change. To remedy URLs being deleted I would recommend making a copy in the Internet Archive, or to use some sort of identifier from Software Heritage (and these are not mutually exclusive).