Обсуждение: \dP and \dX use ::regclass without "pg_catalog."

I noticed that for \dP+ since 1c5d9270e, regclass is written without
"pg_catalog." (Alvaro and I failed to notice it in 421a2c483, too).

+       if (showNested || pattern)
+               appendPQExpBuffer(&buf,
+                                                 ",\n  c3.oid::regclass as \"%s\"",
+                                                 gettext_noop("Parent name"));
+
+       if (showIndexes)
+               appendPQExpBuffer(&buf,
+                                                 ",\n c2.oid::regclass as \"%s\"",
+                                                 gettext_noop("On table"));

\dX is new in v14, and introduced the same issue in ad600bba0 (and modifies it
but not fixed in a4d75c86).

I searched for issues like this, which finds all 4 errors with 1 false positive
in psql/describe.c

|time grep -wF "$(grep -oE 'pg_catalog\.[_[:alpha:]]+' src/bin/psql/describe.c |sed -r 's/^pg_catalog\.//;
/^(char|oid|text|trigger)$/d')" src/bin/psql/describe.c |grep -Ev 'pg_catalog\.|^    *[/ ]\*'
 
|#include "catalog/pg_am.h"
|                                                  ",\n  inh.inhparent::regclass as \"%s\"",
|                                                  ",\n c2.oid::regclass as \"%s\"",
|                                                  "  es.stxrelid::regclass) AS \"%s\"",
|                                                  "es.stxrelid::regclass) AS \"%s\"",

Tom informs me that this is not considered to be interesting as a security patch.

-- 
Justin

On 2021-Aug-27, Justin Pryzby wrote:

> I noticed that for \dP+ since 1c5d9270e, regclass is written without
> "pg_catalog." (Alvaro and I failed to notice it in 421a2c483, too).

Oops, will fix shortly.



-- 
Álvaro Herrera           39°49'30"S 73°17'W  —  https://www.EnterpriseDB.com/

=?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@alvh.no-ip.org> writes:
> My point is that it's good to have it, but it would be much better to
> have something bulletproof, which we could use in an automated check
> somewhere (next to stuff like perlcritic, perhaps).  I don't know what,
> though.

Meh ... this is mostly cosmetic these days, so I can't get excited
about putting a lot of work into it.  We disclaimed search path
bulletproofness for all these queries a long time ago.

I don't object to fixing it in the name of consistency, but that's
not a reason to invest large effort.

            regards, tom lane