--master-retry-count=0 can still stop

From https://github.com/MariaDB/server/pull/3764#discussion_r1917630832:

A value of 0 means the replica will not stop attempting to reconnect.
⸺ https://mariadb.com/kb/en/mariadbd-options/#-master-retry-count

But the code had no special case for 0 since 2002, which means it treats 0 as ULONG_MAX + 1.
It's not forever forever, only practically forever (2^32 tries × 1 millisecond/try ≅ 50 days).

We should either add this special case or correct (clarify?) the doc.