Improving integration with keyring and pip authentication.

[mangin]

What's the problem this feature will solve?

In the current implementation of integration between keyring and pip authentication we could ask username of keyring user and use this method of authentication as 401 fallback.

I would like to improve the integration by:

1. removing all logic that asks people to promt password. In the CI scripts it couldn't possible to type in some environments
2. start to use keyring authentication as an optional method like netrc file (not as fallback)
3. changing a way how we pick up username for keyring:

• from artifact url
• from index url
• from pip.cong parameter (I added a new parameter)

These changes improve integration with keyring and help private companies to block anonymous access to their internal PyPi reposiotires by setting up keyring authentification on their dev machines and CI clusters.

Describe the solution you'd like

Instead of describing the solution I already prepared the fix:
#11823

I didn't create new test cases for my logic because first I would like to here your oppinion about this change and only after that add unittest/integration tests + update docs.

Alternative Solutions

As an alternative solution we can use:

1. Basic HTTP authentication in index urls and artifact urls
   But in this case people should store unencrypted passwords/tokens in their home directories.
   => user processes will be able to read passwords/tokens => some viruses could steal passwords/tokens
2. netrc
   There are the same problem

Keyring allows storage encryption on dev devices and CI machines.

Additional context

I don't have any additional context.

Code of Conduct

• I agree to follow the PSF Code of Conduct.

[jfly]

changing a way how we pick up username for keyring

• from artifact url
• from index url
• from pip.cong parameter (I added a new parameter)

This feels wrong to me, as keyring already has support for discovering a username. I've put together a POC with support for that here: #12748.