Subtraction carry bug in xlog.c in 7.3 and 7.4

"J.R. Nield" <jrnield@usol.com> writes:

The attached patches against 7.3 and 7.4 fix a subtraction carry bug in
xlog.c.

This is simply a waste of cycles, because xrecoff can never be zero
(if it were, it'd be pointing at a page header, which is not a valid
record location). If you look around in the code, you'll note that
xrecoff==0 is actually used as an invalid-value flag in a couple of
contexts.

Can you point to anyplace where the behavior would change? If so
I think there's a deeper bug to fix.

regards, tom lane

Should we add an Assert() to make it clear the current code is OK?

---------------------------------------------------------------------------

Tom Lane wrote:

"J.R. Nield" <jrnield@usol.com> writes:

The attached patches against 7.3 and 7.4 fix a subtraction carry bug in
xlog.c.

This is simply a waste of cycles, because xrecoff can never be zero
(if it were, it'd be pointing at a page header, which is not a valid
record location). If you look around in the code, you'll note that
xrecoff==0 is actually used as an invalid-value flag in a couple of
contexts.

Can you point to anyplace where the behavior would change? If so
I think there's a deeper bug to fix.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Should we add an Assert() to make it clear the current code is OK?

A comment maybe, but not an assert.

regards, tom lane