[DGD] call_out delay

[Felix A. Croes]

Jay Shaffstall <jshaffst at netwalk.com> wrote:

> I'm using call_out with a delay to trigger events in my mud that need to 
> happen at specific intervals.  The call_out docs say, "If the delay is an 
> integer, the function will be called after approximately the specified 
> number of seconds".

DGD handles timing for integer-delay callouts using the POSIX time.
That means that a one-second callout started at time T will be executed
(the rest of the MUD allowing) when the time turns into T+1, measured
in POSIX seconds since 1970.  So the actual delay can be between 0 and 1
seconds, and that's ignoring delays due to other callouts trying to
run at that same moment.

For callouts with a floating point delay, more care is taken that they
run precisely at the intended moment.  A 1.0 delay callout started at
T + 0.23 will be executed at T + 1.23.


> How approximate is approximate?  On my Windows PC, things seem to work as 
> I'd expect.  One of my builders is using a Linux laptop, and has seen the 
> delays decrease until events are being fired continually (the actual value 
> being passed in for the delay hasn't changed).  However on his Linux PC, it 
> seems to work fine.

This may legitimately (appear to) happen if each one second callout takes
.95 seconds to process.  Perhaps the laptop is fairly old, and a lot
slower than the desktop machine?  Otherwise, this may be a bug in DGD
which I would like to learn more about.

Regards,
Dworkin
_________________________________________________________________
List config page:  http://list.imaginary.com/mailman/listinfo/dgd