I just took a close look at the COMP_CRC64 macro in xlog.c.
This isn't a 64-bit CRC. It's two independent 32-bit CRCs, one done
on just the odd-numbered bytes and one on just the even-numbered bytes
of the datastream. That's hardly any stronger than a single 32-bit CRC;
it's certainly not what I thought we had agreed to implement.
We can't change this algorithm without forcing an initdb, which would be
a rather unpleasant thing to do at this late stage of the release cycle.
But I'm not happy with it. Comments?
regards, tom lane