Обсуждение: [HACKERS] Warnings "unrecognized node type" for some DDLs with log_statement = 'ddl'

Hi all,

While reviewing another patch, I have bumped into a couple of failures
when running installcheck if log_statement = 'ddl'. This pops
regression failures for 4 tests: object_address, alter_generic,
alter_operator and stats_ext involving commands CREATE STATISTICS and
ALTER OPERATOR.

You can as well reproduce the failures using simply that:
=# create table aa (a int, b int);
CREATE TABLE
=# CREATE STATISTICS aa_stat ON a, b FROM aa;
WARNING:  01000: unrecognized node type: 332
LOCATION:  GetCommandLogLevel, utility.c:3357
ERROR:  42P17: extended statistics require at least 2 columns
LOCATION:  CreateStatistics, statscmds.c:220
=# ALTER OPERATOR = (boolean, boolean) SET (RESTRICT = NONE);
WARNING:  01000: unrecognized node type: 294
LOCATION:  GetCommandLogLevel, utility.c:3357
ALTER OPERATOR

Attached is a patch to fix all the failures I have spotted. As CREATE
STATISTICS is new in PG10, I am adding an open item as things come
from 7b504eb2. The problems of ALTER OPERATOR are introduced by 9.6.
Still I would suggest to fix everything at the same time. The problem
comes from GetCommandLogLevel() which forgot to add T_CreateStatsStmt
and T_AlterOperatorStmt. I have noticed as well that
T_AlterCollationStmt was missing.

Thanks,
-- 
Michael

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

On Thu, Sep 14, 2017 at 9:24 AM, Michael Paquier
<michael.paquier@gmail.com> wrote:
> Hi all,
>
> While reviewing another patch, I have bumped into a couple of failures
> when running installcheck if log_statement = 'ddl'. This pops
> regression failures for 4 tests: object_address, alter_generic,
> alter_operator and stats_ext involving commands CREATE STATISTICS and
> ALTER OPERATOR.
>
> You can as well reproduce the failures using simply that:
> =# create table aa (a int, b int);
> CREATE TABLE
> =# CREATE STATISTICS aa_stat ON a, b FROM aa;
> WARNING:  01000: unrecognized node type: 332
> LOCATION:  GetCommandLogLevel, utility.c:3357
> ERROR:  42P17: extended statistics require at least 2 columns
> LOCATION:  CreateStatistics, statscmds.c:220
> =# ALTER OPERATOR = (boolean, boolean) SET (RESTRICT = NONE);
> WARNING:  01000: unrecognized node type: 294
> LOCATION:  GetCommandLogLevel, utility.c:3357
> ALTER OPERATOR
>
> Attached is a patch to fix all the failures I have spotted. As CREATE
> STATISTICS is new in PG10, I am adding an open item as things come
> from 7b504eb2. The problems of ALTER OPERATOR are introduced by 9.6.
> Still I would suggest to fix everything at the same time. The problem
> comes from GetCommandLogLevel() which forgot to add T_CreateStatsStmt
> and T_AlterOperatorStmt. I have noticed as well that
> T_AlterCollationStmt was missing.
>
Hmm...I am also able to reproduce the failures reported by you. Your
patch does solve the problem. Just to confirm if we are still missing
TAGS for any other Statement nodes, I had a quick look into the list
given in 'nodes.h' file but couldn't find any missing nodes. So, yes,
your patch does solve the problem completely and looks good to me.
Thanks.

-- 
With Regards,
Ashutosh Sharma
EnterpriseDB:http://www.enterprisedb.com


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

On Thu, Sep 14, 2017 at 6:02 AM, Ashutosh Sharma <ashu.coek88@gmail.com> wrote:
> Hmm...I am also able to reproduce the failures reported by you. Your
> patch does solve the problem. Just to confirm if we are still missing
> TAGS for any other Statement nodes, I had a quick look into the list
> given in 'nodes.h' file but couldn't find any missing nodes. So, yes,
> your patch does solve the problem completely and looks good to me.
> Thanks.

Committed and back-patched.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

On Fri, Sep 15, 2017 at 6:19 AM, Robert Haas <robertmhaas@gmail.com> wrote:
> On Thu, Sep 14, 2017 at 6:02 AM, Ashutosh Sharma <ashu.coek88@gmail.com> wrote:
>> Hmm...I am also able to reproduce the failures reported by you. Your
>> patch does solve the problem. Just to confirm if we are still missing
>> TAGS for any other Statement nodes, I had a quick look into the list
>> given in 'nodes.h' file but couldn't find any missing nodes. So, yes,
>> your patch does solve the problem completely and looks good to me.
>> Thanks.
>
> Committed and back-patched.

Thanks Ashutosh and Robert.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers