Check for USAGE (instead of MEMBER) privilege in all pg_has_role occurrences

[RealGreenDragon]

To work correctly Barman database user should be included in some roles ad this is checked by pg_has_role function.
In all pg_has_role functions MEMBER privilege is checked but in Barman code no SET ROLE is present so it better to check in pg_has_role functions USAGE privilege instead (that includes INHERIT check).

pg_has_role function documentation: https://www.postgresql.org/docs/16/functions-info.html#id-1.5.8.32.4.4.2.2.14.1.1.1

PR already discussed with @martinmarques here: EnterpriseDB/repmgr#807 (comment)

[martinmarques]

The code itself looks good. I would like to see a better description in the commit message, something closer to what you put in the description of the PR. Usually, if the commit message is written in a way that has a line subject and description in subsequent paragraphs, GitHub creates the PR with the user not needing to intervene in any way.

[martinmarques]

Oh, and you will likely need to rebase (don't think you'll find any conflicts as we have been only changing the GitHub actions)

[RealGreenDragon]

@martinmarques Thanks for the advice, I will do it at next PRs. I also rebased the branch as requested.

[martinmarques]

@RealGreenDragon you can amend the commit messages using git rebase --interactive

[RealGreenDragon]

@martinmarques: commits squashed, but I see a linter fail (python so I thinks is not for my changes...)

[martinmarques]

That linker error seems like something on our side.

[martinmarques]

Seems we are testing python black and that'¡s failing. But it has to do with code that is not yours.