target_role definition clarification

2020-09-13 Thread PG Doc comments form
The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/12/sql-alterdefaultprivileges.html
Description:

The documentation on ALTER DEFAULT PRIVILEGES currently defines the
parameter target_role like this:

> The name of an existing role of which the current role is a member. If FOR
ROLE is omitted, the current role is assumed.

To be honest, this didn't help me to understand the purpose of the FOR ROLE
clause. And the rest of the section doesn't really explain it, either.

Also, the first part doesn't seem to be true. I mean, as user postgres,
something like this succeeds:

```
> ALTER DEFAULT PRIVILEGES FOR ROLE juser IN SCHEMA public GRANT select ON
TABLES TO backenduser;
ALTER DEFAULT PRIVILEGES
```

Seems like a contradiction to the first sentence because user postgres
(although it's a superuser) isn't a member of role juser.

What I expected from that section is an explanation of the FOR ROLE
clause.

My current understanding is this: default privileges definitions are bound
to one or multiple roles. That means that only when the FOR ROLE role(s)
create(s) a specified object then the specified GRANT is automatically
applied to it. Other roles aren't affected by that default privileges
definition then.


Fix the example in the document of file-fdw

2020-09-13 Thread btkatsuragiyu

Although csvlog has the leader_pid field, the document of file-fdw
that uses csvlog as an example does not reflect this fact.
Concretely, the DDL for table definition does not have the leader_pid 
field.

This patch fixes the DDL in the document of file-fdw.

Regards,
Yuta Katsuragidiff --git a/doc/src/sgml/file-fdw.sgml b/doc/src/sgml/file-fdw.sgml
index eefc6e7e5b..8831f5911f 100644
--- a/doc/src/sgml/file-fdw.sgml
+++ b/doc/src/sgml/file-fdw.sgml
@@ -265,7 +265,8 @@ CREATE FOREIGN TABLE pglog (
   query_pos integer,
   location text,
   application_name text,
-  backend_type text
+  backend_type text,
+  leader_pid integer
 ) SERVER pglog
 OPTIONS ( filename 'log/pglog.csv', format 'csv' );
 


Re: Fix the example in the document of file-fdw

2020-09-13 Thread Fujii Masao




On 2020/09/14 11:19, btkatsuragiyu wrote:

Although csvlog has the leader_pid field, the document of file-fdw
that uses csvlog as an example does not reflect this fact.
Concretely, the DDL for table definition does not have the leader_pid field.


This seems an oversight of commit b8fdee7d0c.


This patch fixes the DDL in the document of file-fdw.


Thanks for the patch! LGTM. I will commit it.

Regards,

--
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION




Re: Fix the example in the document of file-fdw

2020-09-13 Thread Michael Paquier
On Mon, Sep 14, 2020 at 11:33:23AM +0900, Fujii Masao wrote:
> On 2020/09/14 11:19, btkatsuragiyu wrote:
>> Although csvlog has the leader_pid field, the document of file-fdw
>> that uses csvlog as an example does not reflect this fact.
>> Concretely, the DDL for table definition does not have the leader_pid field.
> 
> This seems an oversight of commit b8fdee7d0c.

Dammit, I thought that I had all the places for that stuff done.  The
patch looks fine to me.

>> This patch fixes the DDL in the document of file-fdw.
> 
> Thanks for the patch! LGTM. I will commit it.

Hmm.  Instead of keeping a copy-paste of the same table in both
config.sgml and file-fdw.sgml, why not having only one?  It seems to 
me that we could just have the file_fdw doc refer to the main
configuration, and it is not difficult to guess how the CREATE TABLE
query needs to be changed to map with foreign table.
--
Michael


signature.asc
Description: PGP signature