Re: droped out precise time calculations in src/interfaces/libpq/fe-connect.c

Joe Conway wrote:
> Seems to work well. But one slight concern:
>
> with previous 2 line patch
> --------------------------
> good connect info, using hostaddr, timeout = 1 || 2 second(s)
> =============================================================
> unsuccessful 0 times: avg n/a
> successful 20000 times: avg 0.0074
>
> These tests were on the same, otherwise unloaded development box. Not sure if
> it is an artifact or not, but the average connection time has gone from 0.0074
> to 0.0087, an increase of about 17%. Also worth noting is that there was very
> little deviation from connect-to-connect on both of the tests (eye-balled the
> total range at about 0.0003). I did not bother calculating standard deviation
> of the connect times, but I'm certain it would not be enough to account for
> the difference. Could anything in Bruce's patch account for this, or do you
> think it is normal variation due to something on my dev box?

Yes, the new code has _three_ time() calls, rather than the old code
that I think only had two. I was going to mention it but I figured
time() was a pretty light system call, sort of like getpid().

I needed the additional time() calls so the computation of remaining
time was more accurate, i.e. we are not resetting the timer on a
select() EINTR anymore.

Should I try to rework it?

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073