On 2014-10-27 09:30:33 -0400, Tom Lane wrote:
> Andres Freund <andres@2ndQuadrant.com> writes:
> > On 2014-10-27 12:51:44 +0200, Heikki Linnakangas wrote:
> >> replication/slot.c and replication/logical/snapbuild.c use a CRC on the
> >> physical slot and snapshot files. It uses the same algorithm as used e.g.
> >> for the WAL. However, they are not doing the finalization step, FIN_CRC32()
> >> on the calculated checksums. Not that it matters much, but it's a bit weird
> >> and inconsistent, and was probably an oversight.
>
> > Hm. Good catch - that's stupid. I wonder what to do about it. I'm
> > tempted to just add a comment about it to 9.4 and fix it on master as
> > changing it is essentially a catversion bump. Any objections to that?
>
> Yeah, I think you should get it right the first time. It hardly seems
> likely that any 9.4 beta testers are depending on those files to be stable
> yet.
Since both state files have the version embedded it'd be trivial to just
do the FIN_CRC32() when loading a version 2 file. Does anybody object to
the relevant two lines of code + docs?
Greetings,
Andres Freund
-- Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services