From: Tom Lane <tgl@sss.pgh.pa.us>
To: "Daniel Verite" <daniel@manitou-mail.org>
cc: "PostgreSQL Hackers" <pgsql-hackers@postgresql.org>,
        Robert Haas <robertmhaas@gmail.com>,
        Jakub Wartak <jakub.wartak@enterprisedb.com>
Subject: Re: psql's FETCH_COUNT (cursor) is not being respected for CTEs
In-reply-to: <d5083352-abfc-4d1e-934c-875a44793f91@manitou-mail.org>
References: <d5083352-abfc-4d1e-934c-875a44793f91@manitou-mail.org>
Comments: In-reply-to "Daniel Verite" <daniel@manitou-mail.org>
	message dated "Wed, 01 Mar 2023 11:41:13 +0100"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2402442.1679688778.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 24 Mar 2023 16:12:58 -0400
Message-ID: <2402443.1679688778@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/2402443.1679688778%40sss.pgh.pa.us>
Precedence: bulk

"Daniel Verite" <daniel@manitou-mail.org> writes:
> PFA an updated patch.

This gives me several "-Wincompatible-pointer-types" warnings
(as are also reported by the cfbot):

common.c: In function 'ExecQueryAndProcessResults':
common.c:1686:24: warning: passing argument 1 of 'PrintQueryTuples' from i=
ncompatible pointer type [-Wincompatible-pointer-types]
       PrintQueryTuples(result_array, ntuples, &my_popt, tuples_fout);
                        ^~~~~~~~~~~~
common.c:679:35: note: expected 'const PGresult **' {aka 'const struct pg_=
result **'} but argument is of type 'PGresult **' {aka 'struct pg_result *=
*'}
 PrintQueryTuples(const PGresult **result, int nresults, const printQueryO=
pt *opt,
                  ~~~~~~~~~~~~~~~~~^~~~~~
common.c:1720:24: warning: passing argument 1 of 'PrintQueryTuples' from i=
ncompatible pointer type [-Wincompatible-pointer-types]
       PrintQueryTuples(result_array, ntuples, &my_popt, tuples_fout);
                        ^~~~~~~~~~~~
common.c:679:35: note: expected 'const PGresult **' {aka 'const struct pg_=
result **'} but argument is of type 'PGresult **' {aka 'struct pg_result *=
*'}
 PrintQueryTuples(const PGresult **result, int nresults, const printQueryO=
pt *opt,
                  ~~~~~~~~~~~~~~~~~^~~~~~

I think the cause is the inconsistency about whether PGresult pointers
are pointer-to-const or not.  Even without compiler warnings, I find
code like this very ugly:

-				success =3D PrintQueryTuples(result, opt, printQueryFout);
+				success =3D PrintQueryTuples((const PGresult**)&result, 1, opt, print=
QueryFout);

I think what you probably ought to do to avoid all that is to change
the arguments of PrintQueryResult and nearby routines to be "const
PGresult *result" not just "PGresult *result".

I find it sad that we can't get rid of ExecQueryUsingCursor().
Maybe a little effort towards reducing overhead in the single-row
mode would help?

			regards, tom lane