Re: BUG #1144: comments don't get propagated

From: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
To: comments on database <Oliver303(at)aon(dot)at>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #1144: comments don't get propagated
Date: 2004-05-20 01:40:05
Message-ID: 200405200140.i4K1e5C09312@candle.pha.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

PostgreSQL Bugs List wrote:
> I did a "create database testdb" from psql without a file ${HOME}/.pgsqlrc.
> The initial database was "template1". Then i tried a "comment on database
> testdb is '...';"
> ==> cannot create comment
>
> --> I'm using LC_ALL=de_AT.ISO8859-15
>
> translated: database comments can be applied to current database only
> so I did a "\c testdb"
> and then "comment on database testdb is '...';"
>
> success!
>
> but after "\c tempalte1"
> "\l+" doesn't give the comment on testdb.
>
> probably the fact is that pg_database is the same across a cluster, but
> pg_catalog.pg_description is not.
>
> After I did a "\c testdb" i saw the the comment.
>
> The docu says that pg_catalog.pg_database is shared between the cluster.
> Probably the same should be true for pg_description, since all oid are
> unique across the cluster. ?!
>
> In my opinion this fact could confuse some people.
>
> BTW:
> Could You give an example of an insert statement to pg_description to
> circumvent this inconvenience.

Interesting idea having pg_description span databases because oids are
unique. Yea, that would help the database description problem, but then
oids can wrap around and they wouldn't be unique anymore.

Basically, it would fix this problem, but it might cause more serious
problems other places.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Sün 2004-05-20 16:30:53 LATIN2 'bssz' and 'bszsz' fails on unique index
Previous Message Bruce Momjian 2004-05-20 00:16:44 Re: Renaming a table leaves orphaned implicit sequences which