Обсуждение: Remove repeated calls to PQserverVersion

Hi,

I found a few functions making unnecessary repeated calls to
PQserverVersion(conn); instead of just calling once and assigning to a
local variable.

PSA a little patch which culls those extra calls.

------
Kind Regards,
Peter Smith.
Fujitsu Australia

On Tue, Jul 13, 2021 at 07:02:27PM +1000, Peter Smith wrote:
> I found a few functions making unnecessary repeated calls to
> PQserverVersion(conn); instead of just calling once and assigning to a
> local variable.

Does it really matter?  PQserverVersion() does a simple lookup at the
internals of PGconn.
--
Michael

Michael Paquier <michael@paquier.xyz> writes:
> On Tue, Jul 13, 2021 at 07:02:27PM +1000, Peter Smith wrote:
>> I found a few functions making unnecessary repeated calls to
>> PQserverVersion(conn); instead of just calling once and assigning to a
>> local variable.

> Does it really matter?  PQserverVersion() does a simple lookup at the
> internals of PGconn.

Yeah, it'd have to be mighty hot code to be worth caring about that,
and none of these spots look like it could be worth it.

            regards, tom lane

On Wed, Jul 14, 2021 at 12:15 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Michael Paquier <michael@paquier.xyz> writes:
> > On Tue, Jul 13, 2021 at 07:02:27PM +1000, Peter Smith wrote:
> >> I found a few functions making unnecessary repeated calls to
> >> PQserverVersion(conn); instead of just calling once and assigning to a
> >> local variable.
>
> > Does it really matter?  PQserverVersion() does a simple lookup at the
> > internals of PGconn.
>
> Yeah, it'd have to be mighty hot code to be worth caring about that,
> and none of these spots look like it could be worth it.

I agree there would be no observable performance improvements.

But I never made any claims about performance; my motivation for this
trivial patch was more like just "code tidy" or "refactor", so
applying performance as the only worthiness criteria for a "code tidy"
patch seemed like a misrepresentation here.

Of course you can judge the patch is still not worthwhile for other
reasons. So be it.

------
Kind Regards,
Peter Smith.
Fujitsu Australia

On 2021-Jul-14, Peter Smith wrote:

> But I never made any claims about performance; my motivation for this
> trivial patch was more like just "code tidy" or "refactor", so
> applying performance as the only worthiness criteria for a "code tidy"
> patch seemed like a misrepresentation here.
> 
> Of course you can judge the patch is still not worthwhile for other
> reasons. So be it.

Personally, I like the simplicity of the function call in those places,
because when reading just that line one immediately knows where the
value is coming from.  If you assign it to a variable, the line is not
standalone and I have to find out where the assignment is, and verify
that there hasn't been any other assignment in between.

-- 
Álvaro Herrera           39°49'30"S 73°17'W  —  https://www.EnterpriseDB.com/
"Pido que me den el Nobel por razones humanitarias" (Nicanor Parra)

> On 14 Jul 2021, at 02:19, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:

> Personally, I like the simplicity of the function call in those places,

+1

--
Daniel Gustafsson        https://vmware.com/