Re: pgbench - allow to create partitioned tables

On 2019-Sep-26, Fabien COELHO wrote:

>
> Hello Alvaro,
>
> > pgbench's main() is overly long already, and the new code being added
> > seems to pollute it even more. Can we split it out into a static
> > function that gets placed, say, just below disconnect_all() or maybe
> > after runInitSteps?
>
> I agree that refactoring is a good idea, but I do not think it belongs to
> this patch. The file is pretty long too, probably some functions could be
> moved to distinct files (eg expression evaluation, variable management,
> ...).

I'm not suggesting to refactor anything as part of this patch -- just
that instead of adding that new code to main(), you create a new
function for it.

> > (Also, we seem to be afraid of function prototypes. Why not move the
> > append_fillfactor() to *below* the functions that use it?)
>
> Because we avoid one more line for the function prototype? I try to put
> functions in def/use order if possible, especially for small functions like
> this one.

I can see that ... I used to do that too. But nowadays I think it's
less messy to put important stuff first, secondary uninteresting stuff
later. So I suggest to move that new function so that it appears below
the code that uses it. Not a big deal anyhow.

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