Jesper Louis Andersen
Jesper Louis Andersen
I believe we are actually trying to do something on this, but carrying out a test would be great.
No, there is no reason from my point of view. It is worth checking with klaar though, since he is the author of the DHT subsystem. I'd really like to...
+1 for that solution!
There are some preliminary code for doing this in the 'next' branch.
The code here has to be tested more before we are sure about what it should do. We should probably make a test-setup and test the torrent files some.
I am not too sure we need SUGGEST handling at the beginning.
I think the problem is that when the dialyzer ran it didn't have the TEST flag set so it assumed those files were dead. Nowadays, my solution would be to...
I would not be surprised if upnp is broken currently. It has not seen much use.
I get good results in boundary/folsom which uses a 1028 entry array in which it performs random replacement. Of course, this only works when your number of samples is "high...
For the record. The algorithm I'm mentioning is called "Vitter's algorithm R" and it does require you to remove old entries. For instance by running 12 buckets on five second...