Bruce Momjian wrote:
> 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?
>
I tried two more runs of 10000, and the average is pretty steady at 0.0087.
However the total range is a fair bit wider than I originally reported.
I added a forth time() call to see what the effect would be. It increased the
average to 0.0089 (two runs of 10000 connects each), so I don't think the
time() call explains the entire difference.
Not sure this is worth worrying about or not. I'd guess anyone serious about
keeping connect time to a minimum uses some kind of connection pool or
persistent connection anyway.
Joe
