Horton

Re: pg_upgrade fails to detect unsupported arrays and ranges

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Daniel Gustafsson <daniel@yesql.se>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Tomas Vondra <tomas.vondra@2ndquadrant.com>
Date: 2019-11-10T21:05:43Z
Lists: pgsql-hackers 
Daniel Gustafsson <daniel@yesql.se> writes:
> On 10 Nov 2019, at 20:07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Although this is a really straightforward patch and I've tested it
>> against appropriate old versions (9.1 and 9.2), I'm very hesitant
>> to shove it in so soon before a release wrap.  Should I do that, or
>> let it wait till after the wrap?

> Having read the patch I agree that it's trivial enough that I wouldn't be
> worried to let it slip through.  However, given that we've lacked the check for
> a few releases, is it worth rushing with the potential for a last-minute
> "oh-shit"?

Probably not, really --- the main argument for that is just that it'd fit
well with the fixes Tomas already made.

>> +		/* arrays over any type selected so far */
>> +						  "			SELECT t.oid FROM pg_catalog.pg_type t, x WHERE typelem = x.oid AND typtype = 'b' "

> No need to check typlen?

Yeah, that's intentional.  A fixed-length array type over a problematic
type would be just as much of a problem as a varlena array type.
The case shouldn't apply to any of the existing problematic types,
but I was striving for generality.

			regards, tom lane

Commits

  1. Improve wording of some pg_upgrade failure reports.

  2. Fix some more omissions in pg_upgrade's tests for non-upgradable types.

  3. Handle arrays and ranges in pg_upgrade's test for non-upgradable types.