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

Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Tom Lane wrote:
> >> That is odd; seems like you should get between 1 and 2 seconds.  How
> >> were you measuring the delay, exactly?
> 
> > Remember, that if you add 1, the select() is going to get tv_sec = 2, so
> > yes, it will be two seconds.
> 
> Yeah, but only if the value isn't recalculated shortly later.  Consider
> 
>     caller computes finish_time = time() + timeout;
> 
>     ...
> 
>     inside select-wait loop, compute max_delay = finish_time - time();
> 
> If the time() value has incremented by 1 second between these two lines
> of code, you have a problem with a 1-second timeout...

Yep.  If you track finish time, you get that 1 second rounding problem,
and if you track just duration/timeout, you get into the problem of not
knowing when the timeout has ended.  I don't think these can be fixed
except by overestimating (+1) or by tracking subseconds along with
seconds so you really know when one second has elapsed.

Perhaps we need to modify a timeout of 1 to be 2 and leave other values
alone.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
359-1001+  If your life is a hard drive,     |  13 Roberts Road +  Christ can be your backup.        |  Newtown Square,
Pennsylvania19073