Re: Weird failure with latches in curculio on v15 - Mailing list pgsql-hackers

From Nathan Bossart
Subject Re: Weird failure with latches in curculio on v15
Date
Msg-id 20230202200957.GA3944544@nathanxps13
Whole thread Raw
In response to Re: Weird failure with latches in curculio on v15  (Michael Paquier <michael@paquier.xyz>)
Responses Re: Weird failure with latches in curculio on v15
List pgsql-hackers
On Thu, Feb 02, 2023 at 01:24:15PM +0900, Michael Paquier wrote:
> On Wed, Feb 01, 2023 at 09:34:44PM -0500, Tom Lane wrote:
>> I was vaguely wondering about removing both of those functions
>> in favor of an integrated function that does a system() call
>> with those things before and after it.
> 
> It seems to me that this is pretty much the same as storing
> in_restore_command in shell_restore.c, and that for recovery modules
> this comes down to the addition of an extra callback called in
> startup.c to check if the flag is up or not.  Now the patch is doing 
> things the opposite way: like on HEAD, store the flag in startup.c but
> switch it at will with the routines in startup.c.  I find the approach
> of the patch a bit more intuitive, TBH, as that makes the interface
> simpler for other recovery modules that may want to switch the flag
> back-and-forth, and I suspect that there may be cases in recovery
> modules where we'd still want to switch the flag, but not necessarily
> link it to system().

Hm.  I don't know if we want to encourage further use of
in_restore_command since it seems to be prone to misuse.  Here's a v2 that
demonstrateѕ Tom's idea (bikeshedding on names and comments is welcome).  I
personally like this approach a bit more.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

Attachment

pgsql-hackers by date:

Previous
From: Peter Geoghegan
Date:
Subject: Re: Amcheck verification of GiST and GIN
Next
From: Peter Geoghegan
Date:
Subject: Re: Amcheck verification of GiST and GIN