Well, no. What I'm looking for is information on how the transactions behave in an error case, and why there is the requirement to have a savepoint in place to be able to continue a transaction after a failed statement.
As far as I'm aware neither PostgreSQL nor OS version do matter for this, I'm interested in the general behavior of the database. But as I said, I do find a lot of documentation on transactions in general, but not about their behavior in an error case.
Your first link is "kinda" what I'm looking for, because it closes with
> Moreover, ROLLBACK TO is the only way to regain control of a transaction block that was put in aborted state by the system due to an error, short of rolling it back completely and starting again.