"Christopher Kings-Lynne" <chriskl@familyhealth.com.au> writes:
>> Yeah. This is fixed in current sources, and I back-patched it into
>> the REL7_2 branch, but current plans don't seem to include a 7.2.2
>> release --- we'll be going straight to 7.3 beta instead.
> Is it worth doing a 7.2.2 patch that will dump people's foreign keys as
> ALTER TABLE/ADD FOREIGN KEY instead of a bunch of CREATE CONSTRAINT
> TRIGGERs, so that they actually become constraints in 7.3?
I don't think it's any easier to do that as a backpatch than as new
functionality in 7.3 pg_dump. As far as the general issue of a 7.2.2
release goes, I'm personally in favor of one --- there are several
important bugfixes in the 7.2 branch right now --- but I don't do the
gruntwork for patch releases, so I can't complain too much.
> Also Tom - did you check if you can CREATE CONSTRAINT TRIGGER on a dropped
> column - I think I neglected to look at that in the patch I submitted
> originally.
I'm pretty sure that won't get past the ATTNAME cache patches, but try
it...
regards, tom lane