about cdm...

[adrian_007]

i don't know if you get the point. i said smart coder can pass it easily - it's not a problem.
also, we get to know if user have wrong connection settings (timeout on testsur).

nobody says it's 100% working detection. it can only give you a clue if user is clean or not.

and to get all info (lock, pk, supports) you don't need to download filelists (testsur have 0B size).

[Pietry]

indeed this could be useful for helping users... but still wont detect bad client and i still dont see a good reason to discriminate clients based on this test.
i think there should be better proof for faking then this one so that action is justified.

[adrian_007]

honestly, if someone want to be a faker, he will no matter what we will use to test it. it's a leak of dc protocol (nmdc and adc). clients are open source, there is no authorization for clients so anybody can fake for ex. dc++ without problem.

btw get any client with TestSUR check and open one of client profiles, you'll see what info is checked by this test.

[Toast]

There is another ways of detections

Like Alums client and Xaydies Clients

not to go on standard testsurs to send some random commands and see how the client responds most faker clients are greedy and respond after that structure..

[Pietry]

this client testing i think it should go for helping people rather than a way of determining which client is a "hacked" one. since everything is open src indeed there is no auth way, so maybe this "detection" should be called "identifying" what clients have and do not have. in ADC this isnt so useful since the INF already has a SU where clients can advertise what they can or cant do

[MikeJJ]

hmm so "no file" has higher priority then " no slots" . a stealth client can easily fake it.

Yes, but old stealth clients didn't. They have improved slightly over the years. Was effective back in the 0.301 - 0.4034 days. the 0.674 generation of stealth clients didn't fail this check.

oh, and about cdm's. . . . apparently they cause cancer. LoL