mldonkey-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Mldonkey-users] Re: pango-20021128b


From: Sven Hartge
Subject: [Mldonkey-users] Re: pango-20021128b
Date: Thu, 28 Nov 2002 20:42:41 +0100
User-agent: tin/1.5.14-20020926 ("Soil") (UNIX) (Linux/2.4.19-109 (i586))

Pierre Etchemaite <address@hidden> wrote:
> Le Thu, 28 Nov 2002 17:05:47 +0100, Steffen Solyga <address@hidden> a
> écrit :

>> Citing address@hidden (Thursday, 2002/11/28 16:34)...

>>> Does it also works if you only remove the "client_md4=..."
>>> in the downloads.ini file ?
 
>> This is it! Works fine when having deleted the client_md4 line
>> each time before running mldonkey...

> Very interesting debugging. Some code handling client_md4 was modified,
> because I tried implementing the same kind of "signature" that eMule uses,
> "magic" bytes in client_md4.
 
> I had to change the signature (again), because I'm afraid no signature based
> on magic file requests can be fully compatible with eMule.

I think the magic file request ID was a very bad idea from the
beginning.

> In a couple of HighID clients, any can initiate a connection to the
> other.  eMule requires that you don't ask for the same file hash more
> than once per 10 mins. But if mldonkey needs to identify itself at the
> beginning of each connection (even incoming ones), and since it
> doesn't control when incoming connections happen, it will certainly
> trigger a ban...

So removing the old-style identification is the Way To Go[tm].

But mldonkey should still listen to the old-way and respond like:

a) stop uploading to 1.x clients.
b) stop sharing sources with 1.x clients
c) stop sharing servers with 1.x clients

mldonkey 2.x should not be affected by this phase-out action.

S°

-- 
Letzte Worte eines Autofahrers: Wenn das Schwein nicht abblendet, ich
tue es auch nicht!




reply via email to

[Prev in Thread] Current Thread [Next in Thread]