Horton

Re: using index or check in ALTER TABLE SET NOT NULL

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Stephen Frost <sfrost@snowman.net>
Cc: Robert Haas <robertmhaas@gmail.com>, Sergei Kornilov <sk@zsrv.org>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2017-11-29T15:47:21Z
Lists: pgsql-hackers 
Stephen Frost <sfrost@snowman.net> writes:
> Isn't the first concern addressed by using SPI..?

I did not look at the patch yet, but TBH if it uses SPI for sub-operations
of ALTER TABLE I think that is sufficient reason to reject it out of hand.
Doing things that way would create way too much of a vulnerability surface
for code touching a partially-updated table.  At minimum, we'd have to
blow holes in existing protections like CheckTableNotInUse, and I think
we'd be forever finding other stuff that failed to work quite right in
that context.  I do not want ALTER TABLE going anywhere near the planner
or executor; I'm not even happy that it uses the parser (for index
definition reconstruction).

			regards, tom lane

Commits

  1. Avoid using INFO elevel for what are fundamentally debug messages.

  2. Revert setting client_min_messages to 'debug1' in new tests.

  3. Allow ALTER TABLE .. SET NOT NULL to skip provably unnecessary scans.

  4. Improve predtest.c's internal docs, and enhance its functionality a bit.