Lloyd Bryant wrote:
> Checking through my errors file, I'm seeing quite a few of the following 
> type messages:
> - - - - - - - - - -
> 07-09-02 20:33:01 (MESSAGE): Tree depth (9) is below the "good" depth (10)
> 07-09-02 20:33:01 (MESSAGE): Discarding tigertree data from 
> <xxx.xxx.xxx.xxx:6346 'Shareaza 2.2.5.0'>: Bad THEX data
> 
> 07-09-03 19:52:53 (MESSAGE): Tree depth (9) is below the "good" depth (10)
> 07-09-03 19:52:53 (MESSAGE): Discarding tigertree data from 
> <xxx.xxx.xxx.xxx:6348 'BearShare Lite 5.2.5.1'>: Bad THEX data
> 
> 07-09-05 17:19:03 (MESSAGE): Hashtree has too few nodes (nodes=283, depth=9)
> 07-09-05 17:19:03 (MESSAGE): Discarding tigertree data from 
> <xxx.xxx.xxx.xxx:11312 'Shareaza 2.2.5.0'>: Bad THEX data
> 
> 07-09-06 11:21:45 (MESSAGE): Tree depth (9) is below the "good" depth (10)
> 07-09-06 11:21:45 (MESSAGE): Discarding tigertree data from 
> <xxx.xxx.xxx.xxx:6346 'BearShare 5.2.5.6 (Polska)'>: Bad THEX data
> - - - - - - - - - -
 
> I'm still seeing problems with the THEX downloads.  For example, I've had 
> several cases where the main download stopped with "Giving priority to 
> THEX", and then had the the THEX download timeout.  Which means that the 
> main download is effectively frozen until I either manually intervene, or 
> the THEX download actually completes.

There's of course no guarantee that a latter connection attempt succeeds if
it worked currently. Especially if a PUSH is required a lot of things can
fail. I did not do anything about the "hang-up and re-connect" issue because
it's a bit more difficult than I had hoped.
 
> To add insult to injury, as often as not when I *do* get the THEX download 
> to complete,  GtkG discards the results, for the reasons in the above 
> examples.

It's not completely wasted. At least during the same session, there will be
no other attempt to fetch the THEX data through this source again.

> So here's the question: Why are we even trying to download THEX data from 
> sources that seem to have an incompatible implementation?  Can GtkG be 
> restricted to downloading THEX data from only those sources that are known 
> to have a compatible implementation?

We cannot know what they are going to sent before we've received it. We have
no other information than the URL.

> Or is there some real hope that Bearshare, Shareaza, et. al , will be 
> updated to conform to whatever spec GtkG is following?

BearShare is discontinued forever. Shareaza might in theory be "fixed" in the
future but I doubt it. You could suggest it in their forum if you're brave.
Shareaza uses a fixed depth of 9. gtk-gnutella and others use an adaptive depth
up to 11. Others accept tigertree data which has one level fewer than the
desired depth. For simplicity, gtk-gnutella currently does not.

-- 
Christian

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
gtk-gnutella-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gtk-gnutella-devel

Reply via email to