Michael Pliskin a écrit :
Hi Nicolas,
Thanks for the pointer, I didn't know that.
As for tora.api.call, I could probably use it. Is it an expensive thing? I.e.
if I would be making 10000 requests to make my socket expire in an hour and
queue them all, would that be bad?
Check the implementation (Tora.speedDelayLoop for delayed events and
ModToraApi.tora_call for request).
Yes, handling a lot of requests should be expensive since the delayQueue
is not sorted, and also because each event will cause an actual request
to be processed by your service, although no network IO is done.
Nicolas
--
Neko : One VM to run them all
(http://nekovm.org)