Обсуждение: Broken CIDR: no fix in 7.4.6?
As been said, in 7.4.5 CIDR data type has been accepted values with non-zero bits to the right of >24-bit mask: http://archives.postgresql.org/pgsql-bugs/2004-10/msg00081.php In CVS head src/backend/utils/adt/network.c (1.54) has a fix by Bruce, but 7.4.6 still has 1.47.2.1 with no fix. I wonder if I can safely use diff -r1.53 -r1.54 .../network.c to patch 7.4.6 sources... -- Fduch M. Pravking
Alexander M. Pravking wrote: > > As been said, in 7.4.5 CIDR data type has been accepted values with > non-zero bits to the right of >24-bit mask: > > http://archives.postgresql.org/pgsql-bugs/2004-10/msg00081.php > > In CVS head src/backend/utils/adt/network.c (1.54) has a fix by Bruce, > but 7.4.6 still has 1.47.2.1 with no fix. > > I wonder if I can safely use diff -r1.53 -r1.54 .../network.c to patch > 7.4.6 sources... Yes, you can, but you should be sure all your existing data is clean or you will have problems with revalidating some of your data. We didn't backpatch this fix because it was something that might break existing bad data that got into your system. -- 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
On Fri, Nov 12, 2004 at 11:37:24AM -0500, Bruce Momjian wrote: > Alexander M. Pravking wrote: > > > > As been said, in 7.4.5 CIDR data type has been accepted values with > > non-zero bits to the right of >24-bit mask: > > > > http://archives.postgresql.org/pgsql-bugs/2004-10/msg00081.php > > > > In CVS head src/backend/utils/adt/network.c (1.54) has a fix by Bruce, > > but 7.4.6 still has 1.47.2.1 with no fix. > > > > I wonder if I can safely use diff -r1.53 -r1.54 .../network.c to patch > > 7.4.6 sources... > > Yes, you can, but you should be sure all your existing data is clean or > you will have problems with revalidating some of your data. We didn't > backpatch this fix because it was something that might break existing > bad data that got into your system. Good point. Of course, I'll check, thanks. -- Fduch M. Pravking