Horton

Re: ExecTypeSetColNames is fundamentally broken

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-12-06T20:30:28Z
Lists: pgsql-hackers 
On Sun, Dec 5, 2021 at 1:46 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> So 0001 attached fixes this by revoking the decision to apply
> ExecTypeSetColNames in cases where a Var or RowExpr is declared
> to return a named composite type.  This causes a couple of regression
> test results to change, but they are ones that were specifically
> added to exercise this behavior that we now see to be invalid.

I don't understand the code so I can't comment on the code, but I find
the regression test changes pretty suspect. Attaching any alias list
to the RTE ought to rename the output columns for all purposes, not
just the ones we as implementers find convenient. I understand that we
have to do *something* here and that the present behavior is buggy and
unacceptable ... but I'm not sure I accept that the only possible fix
is the one you propose here.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

  1. Don't bother to attach column name lists to RowExprs of named types.

  2. Revert applying column aliases to the output of whole-row Vars.