Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the the_gdf group. File : /Proposals/Gnet...
the_gdf@yahoogroups.com
Jan 1, 2003 10:34 am
12724
There were no written specifications for Vendor-Message in the files section of the GDF, so I wrote the following, based on messages exchanged on the GDF...
Raphael Manfredi
Raphael_Manfredi@...
Jan 1, 2003 3:33 pm
12725
I welcome added support for the VendorMessages, in particular Hops- Flow and Connectback. I'm experimenting with Hops-Flow in Swapper, where I send this...
... Servents SHOULD NOT drop connections if they receive unknown descriptors - we should make sure that the latest Gnutella protocol document (?) is clear on...
... Yes, they should not. ... Yes, I know BS is not sending it. However, I think it is a good practice to include this header if you support vendor-specific...
Raphael_Manfredi@...
Jan 1, 2003 6:42 pm
12728
... Can you document Connectback the way I documented Hops-Flow, in a separate document, and drop it into the Vendor-Messages folder I created in the ...
Raphael_Manfredi@...
Jan 1, 2003 7:06 pm
12729
I am just concerned by the fact that the Vendor message does not assign a common role for the GUID, TTL, Hops fields, so that they can be routed blindly (using...
... If you don't know a message, how can you tell how to route it? That's why Vendor messages are not a-priori routable. If you define a vendor-specific...
Raphael_Manfredi@...
Jan 1, 2003 10:52 pm
12731
... From: <Raphael_Manfredi@...> To: <the_gdf@yahoogroups.com> Sent: Wednesday, January 01, 2003 11:52 PM Subject: [the_gdf] Re: Vendor-Message specs ......
... Cool, it's the new year after all! It's just that Vendor messages are primarily meant to be single hop messages. As to the fact that unknown messages are...
Raphael_Manfredi@...
Jan 1, 2003 11:25 pm
12733
... From: <Raphael_Manfredi@...> To: <the_gdf@yahoogroups.com> Sent: Thursday, January 02, 2003 12:24 AM Subject: [the_gdf] Re: Vendor-Message specs ... ...
Philippe, ... I disagree with you. Just because it is possible to do what you say does not mean it must be done that way. I don't like overloading semantics ...
Raphael_Manfredi@...
Jan 2, 2003 12:05 am
12735
... GGEP ... augmented ... implement ... You did not write what I said: I have clearly said that this practice was already used (I say "abused"). And your...
Now that both of these features are starting to come up together, could I ask that the developers consider sending the hash once for each transmission that...
I like Raphael's compression proposal. The only thing that worries me is the amount of compression context memory needed on the Ultrapeer with 100+ leaves. ...
... messages. I think that if you advertise the header, you should send the supported vendor messages descriptor in an unsolicited fashion. BearShare doesn't...
How would you join SHA-1 hashes and bitprint hashes ? Is there somewhere a requirement that the URI resolver must be located at /uri-res/N2R with the URN after...
... will be ... risking to ... withTTL=1,Hops=0 and ... to the ... Gnutella/0.6 ... against ... TTL=1: any ... either Although I don't agree with Phillippe...
... bulk ... Negative. The savings do not outweigh the disadvantages. Its only a few bytes, stop worrying about it. I'm not going to lose sleep over it. ...
... me ... I like it too, even though it fails spectacularly in a UDP environment. Regardless, the original thread (Compression and [download] Mesh/HEAD data)...
The compression context is tunable: the lookup dictionary could be as small as 4KB only and still provide a useful compression (even if it won't produce the...
This may sound unnecessarily futuristic but Alt-locs as strings in HTTP headers are just the beginning, worthy of experimentation, IMHO. Eventually, if Oso's...
Not necessarily, back at OpenCola there was a lot of work done with using XPath for peer addressing and queries and quite a bit of it was independant from XML....
Justin Chapweske
justin@...
Jan 2, 2003 4:09 am
12749
Could you please post some links to that work? Sounds interesting. Due to the way I learned about it, for me XPath has a meaning only in the context of XML...
I'm half-joking, of course, with some 'heretic and futuristic' thoughts. Quotes because I am aware that this has been discussed many times in the past. But, in...
... Which HTML header fields are you talking about? The one in <HEAD>? How does this relate to Gnutella? If you're talking HTTP headers, then HTTP is a...
Raphael_Manfredi@...
Jan 2, 2003 8:09 am
12752
Yes, I meant HTTP headers. In particular, I had alternate locations in mind. It looks complicated and clumsy (even if not impossible) to impose any structure...
