--On Montag, Januar 26, 2009 20:03:41 -0800 Josh Berkus <josh(at)agliodbs(dot)com>
> Jaime, Bernd,
>> having said that, i don't think that inventing new syntax is the way
>> to go... a reloption seems better (thinking a little more, it could be
>> a problem if the user changes the reloptions of an already created
> There's also the issue with backup/restore: we need some kind of syntax
> for restoring a read-only view which doesn't depend on command ordering.
> So we need a ALTER VIEW SET READ ONLY or similar.
Hence my proposal with
CREATE [OR REPLACE] [READ ONLY|UPDATABLE] VIEW
This can easily be extended to ALTER VIEW SET [READ ONLY|UPDATABLE].
Besides other issues already mentioned, this looks more logical to me,
since this is going to change the behavior of a view completely.
In response to
pgsql-hackers by date
|Next:||From: Rick Gigger||Date: 2009-01-27 10:33:08|
|Subject: Re: 8.4 release planning|
|Previous:||From: KaiGai Kohei||Date: 2009-01-27 10:12:06|
|Subject: Re: SE-PostgreSQL Updated Revision (r1460)|
pgsql-committers by date
|Next:||From: Peter Eisentraut||Date: 2009-01-27 11:59:34|
|Subject: Re: [COMMITTERS] pgsql: Automatic view update rules Bernd Helmle|
|Previous:||From: Josh Berkus||Date: 2009-01-27 04:03:41|
|Subject: Re: [COMMITTERS] pgsql: Automatic view update rules Bernd