From: Greg Smith on
Robert Haas wrote:
> On Fri, Jun 25, 2010 at 1:56 PM, Josh Berkus <josh(a)agliodbs.com> wrote:
>
>>>> Shouldn't this be backpatched, or was this a new bug in 9.0?
>>>>
>> We've always output bytes. I'd have noticed the discrepancy myself if I'd
>> read the actual docs ;-)
>>
>> KB would be more useful. And I don't think people have enough scripts built
>> on this yet to make this break anything. We should backport to 8.4.
>>
>
> log_temp_files was introduced in 8.3, so we'll need to backpatch this
> to 8.3, not just 8.4. Greg Smith tells me Simon has been busy with
> other things, so I'm going to pick this up. Barring objections, I'm
> going to revert Simon's patch and change the behavior instead, per the
> attached patch. I will also back-patch the fix to 8.4 and 8.3.
>

I was expecting Josh's suggestion to get argued down...I'm curious why
anyone is considering a behavior change in 8.3 or 8.4 over this? It's
reasonable to change the behavior in 9.0, and/or to correct the docs for
8.3 or 8.4, but I would think a behavior change in the earlier versions
would be out of line with the project's version policies. I know I have
scripts that will mysteriously break if this goes into a stable
version. I'm used to writing things that have code for detecting major
version and acting accordingly, but I'd really prefer this project
doesn't ever wander down the path where I need to start writing in code
for "do this on <=8.3.11, this other thing for >=8.3.12".

--
Greg Smith 2ndQuadrant US Baltimore, MD
PostgreSQL Training, Services and Support
greg(a)2ndQuadrant.com www.2ndQuadrant.us


--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Greg Smith on
Robert Haas wrote:
> OK, so I talked to Bruce about this and I guess I've been persuaded
> that we should just apply the patch I sent upthread to HEAD and leave
> the back-branches broken, for fear of creating an incompatibility.
>

The only thing that might be appropriate to backport is the docs fix,
change "kilobytes" to "bytes" in config.sgml where the parameter is
described. The bug complaint here was mad at the documentation not
matching the behavior, rather than the behavior itself. I agree with
the change to HEAD you've suggested though, that's the right way to do
this for future releases.

This change is something worth mentioning in the release notes for 9.0 too.

--
Greg Smith 2ndQuadrant US Baltimore, MD
PostgreSQL Training, Services and Support
greg(a)2ndQuadrant.com www.2ndQuadrant.us


--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

From: Greg Smith on
Robert Haas wrote:
> Bruce felt it wasn't worth putting that amount of work into backbranch docs that nobody's likely to read
> anyway, but I suppose that view could be overruled if there's a strong consensus.
>

I was never arguing in favor of touching anything in the back branches;
if you recall I didn't even voice an opinion here until I got concerned
about too many changes happening in them. I think a proper fix in 9.0
combined with a release notes comment noting the old/new behavior, so
it's clear what was broken in the old versions, would be quite enough here.

Something like this maybe:

Change log_temp_files to be set in and output its value in kilobytes.
Older versions had the parameter labeled and documented as being in KB,
while it was actually set and outputting to the logs in bytes.

Thanks for following this through, I think it's a useful small bit to
get sorted out fully.

--
Greg Smith 2ndQuadrant US Baltimore, MD
PostgreSQL Training, Services and Support
greg(a)2ndQuadrant.com www.2ndQuadrant.us


--
Sent via pgsql-hackers mailing list (pgsql-hackers(a)postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers