Re: cataloguing NOT NULL constraints

On 2023-Jul-03, Peter Eisentraut wrote:

> On 30.06.23 13:44, Alvaro Herrera wrote:
> > OK, so here's a new attempt to get this working correctly.
> 
> Attached is a small fixup patch for the documentation.
> 
> Furthermore, there are a few outdated comments that are probably left over
> from previous versions of this patch set.

Thanks!  I've incorporated your doc fixes and applied fixes for almost
all the other issues you listed; and fixed a couple of additional
issues, such as

* adding a test to regress for an error message that wasn't covered (and
  removed the XXX comment about that)
* remove a pointless variable addition to pg_dump (leftover from a
  previous implementation of constraint capture)
* adapt the sepgsql tests again (we don't recurse to children when
  there's nothing to do, so an object hook invocation doesn't happen
  anymore -- I think)
* made ATExecSetAttNotNull return the constraint address
* more outdated comments adjustment in MergeAttributes

Most importantly, I fixed table creation for LIKE inheritance, as I
described upthread already.

The one thing I have not touched is add ¬_valid to processCASbits()
in gram.y; rather I added a comment that NOT VALID is not yet suported.
I think adding support for that is a reasonably easy on top of this
patch, but since it also requires more pg_dump support and stuff, I'd
rather not mix it in at this point.  The pg_upgrade support is already
quite a house of cards and it drove me crazy.

So, attached is v10.

-- 
Álvaro Herrera        Breisgau, Deutschland  —  https://www.EnterpriseDB.com/
"Industry suffers from the managerial dogma that for the sake of stability
and continuity, the company should be independent of the competence of
individual employees."                                      (E. Dijkstra)