Re: UserLock oddity with Limit - Mailing list pgsql-hackers

From Rod Taylor
Subject Re: UserLock oddity with Limit
Date
Msg-id 00d201c0d7d7$57beb3b0$2205010a@jester
Whole thread Raw
In response to UserLock oddity with Limit  ("Rod Taylor" <rbt@barchord.com>)
List pgsql-hackers
As a general rule I don't.  But I'm having a hard time trying to find
out if there is a lock on a given item without attempting to lock it.
Seems to work that way with all locks but most delay until it can
obtain it.  Userlocks don't wait.


--
Rod Taylor  BarChord Entertainment Inc.
----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "Rod Taylor" <rbt@barchord.com>
Cc: "Hackers List" <pgsql-hackers@postgresql.org>
Sent: Tuesday, May 08, 2001 11:35 AM
Subject: Re: [HACKERS] UserLock oddity with Limit


> "Rod Taylor" <rbt@barchord.com> writes:
> > Fiddling with userlock stuff for the purposes of setting up an
action
> > queue.  Having the lock in the where clause causes the lock code
to
> > actually lock 2 rows, not just the one that is being returned.
>
> A WHERE clause should *never* contain function calls with side
effects.
> I do not regard this behavior as a bug.  Put the function call in
the
> SELECT's output list if you want to know exactly which rows it is
> evaluated at.
>
> regards, tom lane
>



pgsql-hackers by date:

Previous
From: Alexander Klimov
Date:
Subject: Re: Where `gcc -MMD' puts .d files
Next
From: Doug McNaught
Date:
Subject: Re: Is `#!/bin/sh' configurable?