Here’s a license dilemma for you:
Imagine company X hosting a tarball on their public web server. There’s no publicly available link to this tarball, but if you access the URL with your browser or download tool, you can download it with no restrictions from anywhere in the world.
The tarball contains GPL code. That is, the code in question has GPL license headers (in addition to Copyright (C) by Company X notices).
If you get your hands on said code, is it to be considered GPL and thus valid to be used by a GPL-compatible open source project?
Arguments against this include that the tarball, while being accessible, may not actually have been meant for distribution and thus the license may perhaps not be the one intended for the code in the end.
What if someone would publish the link on a totally unrelated site and say “get the code [here]” and link to the above mentioned code. Wouldn’t that cause at least some people to get the code in good faith and then would the GPL apply?
(Any resemblance to a real-life scenario is purely coincidental. Names have been changed to protect the innocent.)