Re: pgbench - allow backslash-continuations in custom scripts

On Thu, Mar 17, 2016 at 4:12 AM, Kyotaro HORIGUCHI
<horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp> wrote:
> Thank you for the comment, but could you please tell me what kind
> of criteria should I take to split this patch? The discussion
> about splitting criteria is in the following reply (in the
> sentence begins with "By the way").

Well, I'm trying to find a piece of this patch that is small enough
that I can understand it and in good enough shape that I can commit it
independently, but I am having some difficulty with that. I keep
hoping some other committer is going to come along and be able to grok
this well enough to apply it based on what you've already done, but so
far it seems to be the all-me show.

> These patches are made so as to keep the compilable and workable
> state of the source files. It might be a bit more readable if
> unshackled from such restriction.

Keeping it compilable and workable after each patch is essential, but
the first patch is still big and doing a lot of stuff. I'm wondering
if it can be further decomposed.

>> I see this went to psqlscan_int.h, but there's no
>> obvious reason for that particular name, and the comments don't explain it;
>
> I assumed that is a convention of naming by looking libpq-int.h
> (though it doesn't use underscore, but hyphen). But the file
> needs a comment like libpq-int.h. I'll rename it and add such
> comment to it.

OK.

>> - yyless(0);
>> + my_yyless(0);
>>
>> Why do we need to do this? Is "my_" really the best prefix? Is this
>> another change that could be its own patch?
>
> Oops! Sorry for the silly name. I was not able to think up a
> proper name for it. Does psqlscan_yyless seems good?

That does sound better.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company