Horton

Thread

Commits

  1. Log the location field before any backtrace

  1. Log the location field before any backtrace

    Peter Eisentraut <peter.eisentraut@2ndquadrant.com> — 2020-07-09T09:17:32Z

    In PG13, we added the ability to add backtraces to the log output. 
After some practical experience with it, I think the order in which the 
BACKTRACE and the LOCATION fields are printed is wrong.  I propose we 
put the LOCATION field before the BACKTRACE field, not after.  This 
makes more sense because the location is effectively at the lowest level 
of the backtrace.

Patch attached.

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

  2. Re: Log the location field before any backtrace

    Daniel Gustafsson <daniel@yesql.se> — 2020-07-09T11:25:43Z

    > On 9 Jul 2020, at 11:17, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
> 
> In PG13, we added the ability to add backtraces to the log output. After some practical experience with it, I think the order in which the BACKTRACE and the LOCATION fields are printed is wrong.  I propose we put the LOCATION field before the BACKTRACE field, not after.  This makes more sense because the location is effectively at the lowest level of the backtrace.

Makes sense, +1

cheers ./daniel

  3. Re: Log the location field before any backtrace

    Alvaro Herrera <alvherre@2ndquadrant.com> — 2020-07-09T16:31:38Z

    On 2020-Jul-09, Daniel Gustafsson wrote:

> > On 9 Jul 2020, at 11:17, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
> > 
> > In PG13, we added the ability to add backtraces to the log output. After some practical experience with it, I think the order in which the BACKTRACE and the LOCATION fields are printed is wrong.  I propose we put the LOCATION field before the BACKTRACE field, not after.  This makes more sense because the location is effectively at the lowest level of the backtrace.
> 
> Makes sense, +1

Likewise

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

  4. Re: Log the location field before any backtrace

    Michael Paquier <michael@paquier.xyz> — 2020-07-10T02:04:28Z

    On Thu, Jul 09, 2020 at 12:31:38PM -0400, Alvaro Herrera wrote:
> On 2020-Jul-09, Daniel Gustafsson wrote:
>> On 9 Jul 2020, at 11:17, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
>>> 
>>> In PG13, we added the ability to add backtraces to the log
>>> output. After some practical experience with it, I think the
>>> order in which the BACKTRACE and the LOCATION fields are printed
>>> is wrong.  I propose we put the LOCATION field before the
>>> BACKTRACE field, not after.  This makes more sense because the
>>> location is effectively at the lowest level of the backtrace. 
>> 
>> Makes sense, +1
> 
> Likewise

+1.
--
Michael

  5. Re: Log the location field before any backtrace

    Peter Eisentraut <peter.eisentraut@2ndquadrant.com> — 2020-07-10T06:36:07Z

    On 2020-07-10 04:04, Michael Paquier wrote:
> On Thu, Jul 09, 2020 at 12:31:38PM -0400, Alvaro Herrera wrote:
>> On 2020-Jul-09, Daniel Gustafsson wrote:
>>> On 9 Jul 2020, at 11:17, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
>>>>
>>>> In PG13, we added the ability to add backtraces to the log
>>>> output. After some practical experience with it, I think the
>>>> order in which the BACKTRACE and the LOCATION fields are printed
>>>> is wrong.  I propose we put the LOCATION field before the
>>>> BACKTRACE field, not after.  This makes more sense because the
>>>> location is effectively at the lowest level of the backtrace.
>>>
>>> Makes sense, +1
>>
>> Likewise
> 
> +1.

committed

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