pg.ddx.io pgsql-bugs@postgresql.org mailing list archive
help / color / mirror / Atom feedPostgres do not allow to create many tables with more than 63-symbols prefix
2+ messages / 2 participants
[nested] [flat]
* Postgres do not allow to create many tables with more than 63-symbols prefix
@ 2022-06-03 19:02 Andrey Lepikhov <a.lepikhov@postgrespro.ru>
2022-06-23 02:03 ` Re: Postgres do not allow to create many tables with more than 63-symbols prefix Masahiko Sawada <sawada.mshk@gmail.com>
0 siblings, 1 reply; 2+ messages in thread
From: Andrey Lepikhov @ 2022-06-03 19:02 UTC (permalink / raw)
To: PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>; +Cc: d.koval@postgrespro.ru
According to subj you can try to create many tables (induced by the case
of partitioned table) with long prefix - see 6727v.sql for reproduction.
But now it's impossible because of logic of the makeUniqueTypeName()
routine.
You get the error:
ERROR: could not form array type name for type ...
It is very corner case, of course. But solution is easy and short. So,
why not to fix? - See the patch in attachment.
--
regards,
Andrey Lepikhov
Postgres Professional
Attachments:
6727v.sql (application/sql, 397B) download
0001-Allow-postgresql-to-generate-more-relations-with-the.patch (text/plain, 3.6K) download
From e28e0b88ce7bfe79ebd3fa82a3124bf924a9fb53 Mon Sep 17 00:00:00 2001
From: "Andrey V. Lepikhov" <a.lepikhov@postgrespro.ru>
Date: Fri, 3 Jun 2022 21:40:01 +0300
Subject: [PATCH] Allow postgresql to generate more relations with the same 63
symbols long prefix. Rewrite the makeUniqueTypeName routine - generator of
unique name is based on the same idea as the ChooseRelationName() routine.
Authors: Dmitry Koval <d.koval@postgrespro.ru>, Andrey Lepikhov <a.lepikhov@postgrespro.ru>
---
src/backend/catalog/pg_type.c | 41 +++++++++++++----------
src/test/regress/expected/alter_table.out | 6 ++--
2 files changed, 27 insertions(+), 20 deletions(-)
diff --git a/src/backend/catalog/pg_type.c b/src/backend/catalog/pg_type.c
index 03788cb975..6977377f40 100644
--- a/src/backend/catalog/pg_type.c
+++ b/src/backend/catalog/pg_type.c
@@ -26,6 +26,7 @@
#include "catalog/pg_namespace.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_type.h"
+#include "commands/defrem.h"
#include "commands/typecmds.h"
#include "mb/pg_wchar.h"
#include "miscadmin.h"
@@ -936,17 +937,17 @@ makeMultirangeTypeName(const char *rangeTypeName, Oid typeNamespace)
* makeUniqueTypeName
* Generate a unique name for a prospective new type
*
- * Given a typeName, return a new palloc'ed name by prepending underscores
- * until a non-conflicting name results.
+ * Given a typeName, return a new palloc'ed name by prepending underscore
+ * and (if needed) adding a suffix to the end of the type name.
*
* If tryOriginal, first try with zero underscores.
*/
static char *
makeUniqueTypeName(const char *typeName, Oid typeNamespace, bool tryOriginal)
{
- int i;
- int namelen;
- char dest[NAMEDATALEN];
+ int pass = 0;
+ char suffix[NAMEDATALEN];
+ char *type_name;
Assert(strlen(typeName) <= NAMEDATALEN - 1);
@@ -956,23 +957,29 @@ makeUniqueTypeName(const char *typeName, Oid typeNamespace, bool tryOriginal)
ObjectIdGetDatum(typeNamespace)))
return pstrdup(typeName);
+ /* Prepare initial object name. Just for compatibility. */
+ type_name = makeObjectName("", typeName, NULL);
+
/*
- * The idea is to prepend underscores as needed until we make a name that
- * doesn't collide with anything ...
+ * The idea of unique name generation is similar to ChooseRelationName()
+ * implementation logic.
*/
- namelen = strlen(typeName);
- for (i = 1; i < NAMEDATALEN - 1; i++)
+ do
{
- dest[i - 1] = '_';
- strlcpy(dest + i, typeName, NAMEDATALEN - i);
- if (namelen + i >= NAMEDATALEN)
- truncate_identifier(dest, NAMEDATALEN, false);
-
if (!SearchSysCacheExists2(TYPENAMENSP,
- CStringGetDatum(dest),
+ CStringGetDatum(type_name),
ObjectIdGetDatum(typeNamespace)))
- return pstrdup(dest);
- }
+ return type_name;
+ /* Previous attempt was failed. Prepare a new one. */
+ pfree(type_name);
+ snprintf(suffix, sizeof(suffix), "%d", pass++);
+ type_name = makeObjectName("", typeName, suffix);
+ } while (pass >= 0);
+
+ /*
+ * Return NULL only after integer overflow. It seems impossible because
+ * of OID limits.
+ */
return NULL;
}
diff --git a/src/test/regress/expected/alter_table.out b/src/test/regress/expected/alter_table.out
index 5ede56d9b5..41a019eab7 100644
--- a/src/test/regress/expected/alter_table.out
+++ b/src/test/regress/expected/alter_table.out
@@ -197,9 +197,9 @@ SELECT typname FROM pg_type WHERE oid = 'attmp_array[]'::regtype;
(1 row)
SELECT typname FROM pg_type WHERE oid = '_attmp_array[]'::regtype;
- typname
-----------------
- ___attmp_array
+ typname
+-----------------
+ __attmp_array_0
(1 row)
DROP TABLE _attmp_array;
--
2.36.1
* Re: Postgres do not allow to create many tables with more than 63-symbols prefix
2022-06-03 19:02 Postgres do not allow to create many tables with more than 63-symbols prefix Andrey Lepikhov <a.lepikhov@postgrespro.ru>
@ 2022-06-23 02:03 ` Masahiko Sawada <sawada.mshk@gmail.com>
0 siblings, 0 replies; 2+ messages in thread
From: Masahiko Sawada @ 2022-06-23 02:03 UTC (permalink / raw)
To: Andrey Lepikhov <a.lepikhov@postgrespro.ru>; +Cc: PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>; d.koval@postgrespro.ru
Hi,
On Sat, Jun 4, 2022 at 4:03 AM Andrey Lepikhov
<a.lepikhov@postgrespro.ru> wrote:
>
> According to subj you can try to create many tables (induced by the case
> of partitioned table) with long prefix - see 6727v.sql for reproduction.
> But now it's impossible because of logic of the makeUniqueTypeName()
> routine.
> You get the error:
> ERROR: could not form array type name for type ...
>
> It is very corner case, of course. But solution is easy and short. So,
> why not to fix? - See the patch in attachment.
While this seems to be a good improvement, I think it's not a bug.
Probably we cannot backpatch it as it will end up having type names
defined by different naming rules. I'd suggest discussing it on
-hackers.
Regarding the patch, I think we can merge makeUniqueTypeName() to
makeArrayTypeName() as there is no caller of makeUniqueTypeName() who
pass tryOriginal = true. Also looking at other ChooseXXXName()
functions, we don't care about integer overflow. Is it better to make
it consistent with them?
Regards,
--
Masahiko Sawada
EDB: https://www.enterprisedb.com/
^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2022-06-23 02:03 UTC | newest]
Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2022-06-03 19:02 Postgres do not allow to create many tables with more than 63-symbols prefix Andrey Lepikhov <a.lepikhov@postgrespro.ru>
2022-06-23 02:03 ` Masahiko Sawada <sawada.mshk@gmail.com>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox