On Wed, Oct 20, 2010 at 10:35 AM, Heikki Linnakangas
> On 20.10.2010 17:32, Tatsuo Ishii wrote:
>>>> pg_is_in_recovery() returns a bool, are you proposing to change that?
>>> No. I just thought about adding more condition when it returns true.
>> Here is the patch. Comments are welcome!
>> ! /* use volatile pointer to prevent code rearrangement */
>> ! volatile WalRcvData *walrcv = WalRcv;
>> ! PG_RETURN_BOOL(RecoveryInProgress()&& walrcv->walRcvState ==
> This returns 'false' if you're in hot standby mode running against an
> archive. That seems wrong, I don't think the walreceiver state should play
> any role in this.
I think what we need is a second function, not a change to the
definition of this one.
The Enterprise PostgreSQL Company
In response to
pgsql-hackers by date
|Next:||From: Alvaro Herrera||Date: 2010-10-20 14:53:58|
|Subject: Re: max_wal_senders must die|
|Previous:||From: Robert Haas||Date: 2010-10-20 14:49:37|
|Subject: Re: leaky views, yet again|