Re: Re: Dynamic Searching

Hi Peter,

Thanks for the comment.

Yes you are right, we can generally do nothing about these searches where the probes return more than the needed results. Maybe these searchers are too popular and are not so likely to occur when one performs a regular search for something one needs. All defines I use are according to the dynamic search specs. I suppose that the values for the number of probes, ttl and the time to wait for responses are chosen so that in average the searches return the desired number of results. Of course there can still be something to be optimised there and if you really think so we can set up the question in the GDF and see what the others think :).
About your last point whether the PacketProcessor should know about the Searcher. You are right I also intended to implement it with signals and slots but decided to test it like this at first.

Regards Slavcho

Would you like to post a relpy?


This post is a reply to:
Re: Dynamic Searching
Hi Slavcho, Sorry for the late comments. As you know I was away for awhile... So, I did some tests using searches of different popularity and with different number of peer connections. (more...)

Follow-ups:
Re: Re: Re: Dynamic Searching
OK! I committed your patch as revision 9. We should do further testing after we add support for metadata searches. With metadata in use, we may observe different behavior.