headers
Stas Khirman | 5 Apr 2007 05:08
Picon

8-bytes transaction id

Quite often , I see some of the DHT nodes are responding with 8-bytes
transaction ID despite the fact that queries have only 2-bytes ID.
Apparently,  all responses in question a pretty similar format : two first
bytes are correct transaction ID, then 4 bytes of 0x000018FF and then two
more bytes (often similar between different peers). For example:
xxxx000018FFyyyy
xxxx000018FFyyyy

It seems to be a behavior of some "not-so-standard" implementation - does
anyone know what client have such behavior? What is a meaning of the last 6
bytes? Version?

Regards
Stas
Permalink | Reply | 0 comments |
headers
MooPolice | 5 Apr 2007 07:58
Picon
Favicon

Re: 8-bytes transaction id

> Quite often , I see some of the DHT nodes are responding with 8-bytes
> transaction ID despite the fact that queries have only 2-bytes ID.
> Apparently,  all responses in question a pretty similar format : two first
> bytes are correct transaction ID, then 4 bytes of 0x000018FF and then two
> more bytes (often similar between different peers). For example:
> xxxx000018FFyyyy
> xxxx000018FFyyyy
> 
> It seems to be a behavior of some "not-so-standard" implementation - does
> anyone know what client have such behavior? What is a meaning of the last 6
> bytes? Version?
> 

There are 1/2/4/8 byte transaction ids in the different dht implementations, unfortunally some 
of them cut off/expand the received ids. For example, my dht uses a 4byte id and some 
clients respond with 1 byte (the first from of my 4 bytes) while others respond with 8 bytes 
(my 4 bytes + 4 more). I don't think that there is any meaning to them. It's just that those 
clients transform the ids into their own format. :(
Btw, one is supposed to return the id as it is. So if i get 1 byte i return 1 byte, if i get 8 then i 
return those 8. Basically i don't care thats inside the id of other clients.

MassaRoddel

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |
headers
Alan McGovern | 5 Apr 2007 09:44
Picon
Gravatar

Re: 8-bytes transaction id

You should not reply to nodes which send invalid transaction ID's. Any node which does *not* send back the transaction ID you sent it exactly as you sent it should be dropped.

It's not "not-so-standard", it's just wrong.

There was a thead about this a while ago which tried to clarify exactly how the transaction ID's and tokens should be encoded. Check that out.

Alan.

On 4/5/07, Stas Khirman <staskh <at> comcast.net> wrote:
Quite often , I see some of the DHT nodes are responding with 8-bytes
transaction ID despite the fact that queries have only 2-bytes ID.
Apparently,  all responses in question a pretty similar format : two first
bytes are correct transaction ID, then 4 bytes of 0x000018FF and then two
more bytes (often similar between different peers). For example:
xxxx000018FFyyyy
xxxx000018FFyyyy

It seems to be a behavior of some "not-so-standard" implementation - does
anyone know what client have such behavior? What is a meaning of the last 6
bytes? Version?

Regards
Stas

_______________________________________________
BitTorrent mailing list
BitTorrent <at> lists.ibiblio.org
http://lists.ibiblio.org/mailman/listinfo/bittorrent

_______________________________________________
BitTorrent mailing list
BitTorrent <at> lists.ibiblio.org
http://lists.ibiblio.org/mailman/listinfo/bittorrent
Permalink | Reply | 0 comments |
headers
Stas Khirman | 5 Apr 2007 19:07
Picon

LSD == Cache Discovery Protocol ??

What is newly implemented LSD functionality ? Is it an open implementation
for "secret" CacheLogics Discovery Protocol or something else?

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 4 comments |
headers
Arvid Norberg | 5 Apr 2007 20:12
Picon
Picon
Favicon

Re: LSD == Cache Discovery Protocol ??

It is short for local service discovery, it is just a periodic multi  
cast message to announce a peers presence on a local network.
And the ability to receive such messages from other peers and connect  
to them. This has been available in Mainline for a while.

It's not very well tested though. I couldn't get the multicast to  
work on OS X.

On Apr 5, 2007, at 10:07 , Stas Khirman wrote:

> What is newly implemented LSD functionality ? Is it an open  
> implementation
> for "secret" CacheLogics Discovery Protocol or something else?

--
Arvid Norberg

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 3 comments |
headers
Stas Khirman | 5 Apr 2007 20:49
Picon

Re: LSD == Cache Discovery Protocol ??

Thank!

Is any information about "cache Discovery Protocol" available? Wikipedia
claims that it is a close protocol, but implemented by MainLine. Does
MainLine went close source ( I was searching net for sources, but find
nothing).

Regards
Stas

> -----Original Message-----
> From: libtorrent-discuss-bounces <at> lists.sourceforge.net [mailto:libtorrent-
> discuss-bounces <at> lists.sourceforge.net] On Behalf Of Arvid Norberg
> Sent: Thursday, April 05, 2007 11:13 AM
> To: General discussion about libtorrent
> Subject: Re: [libtorrent] LSD == Cache Discovery Protocol ??
> 
> It is short for local service discovery, it is just a periodic multi
> cast message to announce a peers presence on a local network.
> And the ability to receive such messages from other peers and connect
> to them. This has been available in Mainline for a while.
> 
> It's not very well tested though. I couldn't get the multicast to
> work on OS X.
> 
> On Apr 5, 2007, at 10:07 , Stas Khirman wrote:
> 
> > What is newly implemented LSD functionality ? Is it an open
> > implementation
> > for "secret" CacheLogics Discovery Protocol or something else?
> 
> --
> Arvid Norberg
> 
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share
> your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Libtorrent-discuss mailing list
> Libtorrent-discuss <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/libtorrent-discuss

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 2 comments |
headers
Gregor Riepl | 6 Apr 2007 01:45
Picon

Re: LSD == Cache Discovery Protocol ??

hi stas

> Is any information about "cache Discovery Protocol" available?  
> Wikipedia
> claims that it is a close protocol, but implemented by MainLine. Does
> MainLine went close source ( I was searching net for sources, but find
> nothing).

mainline is the "official" client, right?
there is no source release as such, because it's written in python  
and the program you download is in fact the source code. just get  
yourself the unix tgz and you're set.

> It's not very well tested though. I couldn't get the multicast to
> work on OS X.

*sigh* why didn't they just use mdns/bonjour/zeroconf? that's a  
standardized system and should be easy enough to implement on osx.
what specific problems did you have with multicast?

cu
gregor

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 1 comment |
headers
arvid | 6 Apr 2007 01:59
Picon
Picon
Favicon

Re: LSD == Cache Discovery Protocol ??

Quoting Gregor Riepl <seto-kun <at> freesurf.ch>:
> > It's not very well tested though. I couldn't get the multicast to
> > work on OS X.
> 
> *sigh* why didn't they just use mdns/bonjour/zeroconf? that's a  
> standardized system and should be easy enough to implement on osx.
> what specific problems did you have with multicast?

I don't receive the packets I send out on my multicast socket, even though I
have loopback enabled. As a matter of fact, I don't receive any messages
received on the multicast socket. They show up in ethereal/wireshark though.

Here is the source:

http://code.rasterbar.com/libtorrent/browser/trunk/src/lsd.cpp#L118

--

-- 
Arvid Norberg

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |
headers
MooPolice | 8 Apr 2007 08:53
Picon
Favicon

upnp (test)

Hydri,

the reason why the upnp test is does not end normally is the outstanding refresh timer.
If i cancel that timer in upnp::close() then the test ends normal.
I think the right place to cancel the timer would be unmap_port function.

Maybe that is what blocks my client too.

MassaRoddel

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 6 comments |
headers
arvid | 8 Apr 2007 20:24
Picon
Picon
Favicon

Re: upnp (test)

Quoting MooPolice <at> massaroddel.de:
> Hydri,
> 
> the reason why the upnp test is does not end normally is the outstanding
> refresh timer.
> If i cancel that timer in upnp::close() then the test ends normal.
> I think the right place to cancel the timer would be unmap_port function.
> 
> Maybe that is what blocks my client too.

Thanks for finding this. I haven't had time to fix this yet. I have another
modification that might fix the blocking destructor too.

--

-- 
Arvid Norberg

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 5 comments |

Gmane