Re: pgbench more operators & functions

Bonjour Michaël, Hello Robert,

>> Let's mark this Returned with Feedback and move on. We've only got a
>> week left in the CommitFest anyhow and there are lots of other things
>> that still need work (and which actually have been revised to match
>> previous feedback).
>
> Done as such, let's move on.

Hmmm.

I think that there is a misunderstanding, most of which being my fault.

I have really tried to do everything that was required from committers,
including revising the patch to match all previous feedback.

Version 6 sent on Oct 4 did include all fixes required at the time (no if,
no unusual and operators, TAP tests)... However I forgot to remove some
documentation about the removed stuff, which made Robert think that I had
not done it. I apologise for this mistake and the subsequent
misunderstanding:-(

The current v8 sent on Jan 25 should only implement existing server-side
stuff, including with the same precedence as pointed out by Tom.

So for the implementation side I really think that I have done exactly all
that was required of me by committers, although sometimes with bugs or
errors, my apology, again...

As for the motivation, which is another argument, I cannot do more than
point to actual published official benchmark specifications that do
require these functions. I'm not inventing anything or providing some
useless catalog of math functions.

If pgbench is about being seated on a bench and running postgres on your
laptop to get some heat, my mistake... I thought it was about
benchmarking, which does imply a few extra capabities.

If the overall feedback is to be undestood as "the postgres community does
not think that pgbench should be able to be used to implement benchmarks
such as TPC-B", then obviously I will stop efforts to improve it for that
purpose.

To conclude:

IMHO the relevant current status of the patch should be "Needs review" and
possibly "Move to next CF".

If the feedback is "we do not want pgbench to implement benchmarks such as
TPC-B", then indeed the proposed features are not needed and the status
should be "Rejected".

In any case, "Returned with feedback" does not really apply.

A+

--
Fabien.