A blkno argument passed to pg_checksum_page procedure should be not a
relative block number in a given segment but an absolute block number
in the entire relation. It's the same for the segment number 0, but for
segment number N the difference is RELSEG_SIZE * N.
Aleksander Alekseev
Add PostgreSQL Global Development Group to the copyright notices, and
remove a couple of no-longer-appropriate references to Red Hat.
Unfortunately I can't undo Red Hat's choice of GPL licensing, but it is
what it is.
Also reduce the pain of version-stamping by coalescing references to the
version into one pair of macros, and removing not-especially-useful
change logs. IMO the commit logs serve that purpose just as well.
