Re: Bizarre coding in recovery target GUC management - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Bizarre coding in recovery target GUC management
Date
Msg-id 6c30a0e9-eecb-7b5a-9e3f-63108fa067f7@2ndquadrant.com
Whole thread Raw
In response to Bizarre coding in recovery target GUC management  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On 2020-10-12 18:00, Tom Lane wrote:
> Would someone explain to me why assign_recovery_target_lsn and related GUC
> assign hooks throw errors, rather than doing so in the associated check
> hooks?  An assign hook is not supposed to throw an error.  Full stop, no
> exceptions.  We wouldn't bother to separate those hooks otherwise.

That code is checking whether more than one recovery target GUC has been 
set.  I don't think the check hook sees the right state to be able to 
check that.

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services



pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: "unix_socket_directories" should be GUC_LIST_INPUT?
Next
From: Peter Eisentraut
Date:
Subject: Re: warn_unused_results