DC++

sandos

With the GPL, if you were to say modify someone elses code, and all you did was added a check for a dll file or something, and if it doesn't exist run as normal, otherwise let the dll take control, would the dll now become a program of its own? Or would it also be covered by the licence of the orig...

sandos

I don't recall saying download in any of my previous posts about this ignorant method of stealing slots. Maybe you should learn how to comprehend before responding? Oh and since when was DC++ the haven for ignorance? Oh, that's right, if it's open source and it's catching on... "hey cops, don't bot...

sandos

I prefer to use a Tiger Tree hash as opposed to, say, an SHA1 tree hash because the TTH (calculated with, if not always storing all the leaves down to, a 1kB block size) will allow compatibility with other users of Bitzi's code/algorithms. Too bad that you cant yet lookup files in bitzi using only ...

sandos

Id say calling it localport is a bit confusing, as it depends on which side youre at. Its actually the source-port if you ask me, as opposed to destination port, which should be 2501 for a hublist server.

sandos

The solution to the problem with cheating clients might be to have (super-)trusted clients, either ops or a bot, which seeds the rating server with ratings that are correct: http://www.advogato.org/trust-metric.html The problem here is that their model of a attack is different from this attack as ne...

sandos

Hashes. Hashes. Hashes and MP3 info. It seems that an XML of sorts might be really nice to give information on a file - hash, hash of just the audio part of mp3s (so people who retag [like me] can still be alternate sources), possibly ID3 tag information, AVI codecs, bitrates, image resolution, etc...

sandos

Interesting, Im playing around with non-blocking stuff in linux, and this sounds pretty much like an idea I had, where I want to run a searchbot as a cgi.

sandos

Meanwhile, if someone (GargoyleMT?) could take a peek at BCDC and determine whether a queue could be implemented in a snap (especially if some nifty algorithm could be use to "bump" clients with good ratings to better positions) this'd help, because we probably need to do some distributed work to g...

sandos

One note: dont use dc-dreamland.dhs.org for registering, as it doesnt work. Either use dreamland.gotdns.org or dcreg.mine.nu. dreamland.gotdns.org MIGHT disappear, though Ive been told it wont. Safest bet is probably to stick with dcreg.mine.nu, I will update that to point wherever the hublist is at...

sandos

yilard wrote:btw. javaDC looks quite good, but I haven't tried it yet.

It was kind-of nice once, but its a bit buggy. I think most bugs are pretty easy to fix though.

sandos

This domain will not be valid much longer, so: dreamland.gotdns.org will work for some time still (1 month maybe?) and the new one will be: dc-dreamland.com I will also have to implement fake-checking soon, lots of fakehubs already. Ill also probably have to check hub from another hostname than the ...

sandos

0.232 has a pretty severe bug:

http://sourceforge.net/tracker/index.ph ... tid=427632

and it seems it should be easy for Arne to fix. Not for me though.

sandos

I think thats dctc, not sure.

sandos

This can be generalized to "Support of file metadata", which Im sure is already in a feature request. Thing is, this should be done right, and its not a 5 minute hack. (well it could be, but Im sure it wouldnt be too good) BTW: Does anyone know of any opensource library for extraction of metadata? I...

sandos

If DC++ would accept and support ED2K links just as eMule or overnet it would be the optimal filesharing tool! Perhaps. What is an ED2K link? How does one use it, both as a programmer and a user? Give links - don't expect people to either know what you are talking about or care enough to go digging...

sandos

One thread per user shouldnt be a real problem, if done correctly, but non-blocking (using one thread) will probably be more effective when the load is huge. (compare thttpd to apache)

sandos

Have you looked into Tiger Tree (or something like that) hashes? From what I understand they have built in segmentation kind of thing. Look it up. When we developed the hash idea fully last time, long long time ago, someone brought the idea of the Tiger hash and it sounded great. TTH are great, but...

sandos

Why is the DC++ hub list not updating? This one seems much larger. http://www.mds.mdh.se/~dtv01jbd/PublicHubList.config.bz2 Whose is this and how often is it updated? Are there any others we can use? Its mine. I dont know why the DC++ list isnt updating. There are other hublists to use, I dont know...

sandos

Who wrote:when calling CryptoManager::decodeBZ2, with a partially downloaded BZ2 file, it may enter a endless loop in:

Yeah, I can confirm that this has happened for me. I never reported it, since I thought it was just my list that happened to be corrupt at that moment, but I guess not then.

sandos

Look at the list in my signature, if its in there, it has been up on the nm list.

sandos

if hashes are used, it would probably be best if they were implemented for files larger than a certain size ( let's say.. 100 megs ) because... well, i don't care about the authenticity of some 4-meg mp3 or some doc or smth. however, the issue of the size of some shares ( 100gigs + ) would cause a ...

sandos

I think this has to do with default calling conventions/symbol decorations. (cdecl, fastcall, stdcall)

sandos

One list at a time.

sandos

I did a dirty grep for "config" in this weeksÃ‚Â´ http logs at the machine hosting www.mds.mdh.se, and I got 14114 downloads, and 7554 404s. This is due to this bug: http://sourceforge.net/tracker/index.ph ... tid=427632

sandos

Ah, ok. I can remove those ugly numbers, actually.

sandos

I havent been able to get a hublist for a few days now. The two list address's in my DCPP setup are http://dcplusplus.sourceforge.net/PublicHubList.config; http://dcpp.lichlord.org/PublicHubList.config.bz2 i know that second one isnt suposed to be online atm... what are some other hublists?? The on...

sandos

The hublist now shows shareamount in the description field. Too bad you wont be able to sort on it. :) The list actually has, for some time, sent the shareamount in the fifth column of the hublist, but dc++ doesnt display that one. For now, only hubs that registers themselves have any amount, all el...

sandos

Defiance wrote:IÃ‚Â´d love to be able to see how much is shared in a hub
from the hublist view.

/Defiance

'

I could fix that for the dreamland list. Ill keep you posted. Im not sure I can do this for pinged hubs.

sandos

aDe wrote:It is the total amount of shared bytes in the hub, afaik.

It is indeed. Recently a buggy Open DC hub made my hublist throw an exception (it sent 2^64 bytes shared, java only has signed ints) since my hublist actually parses the number. I have code that appends "(shared bytes)" to the description.

sandos

But I cant help myself, I always think about how the system can be abused There is a problem associated with hashing: cheaters would forge their hashes intentionaly to prevent them from being chosen as alternate source and thus they will reduce the outgoing traffic from them. It's probably just as ...

sandos

LinkSync wrote:I think u has to PAY whoever owns the reg servers to list yer hub or something...

Not the dreamland list, no.

sandos

a TTH (TigerTreeHash) implemenetation is available at http://sourceforge.net/projects/tigertree/, but the problem is it has a security issue, so we do not want to use it as it stands today. I was going to give a link to a explanation, but the server is down so Ill get back with that.

sandos

Hm, maybe chat doesn't add so much load on hubs. On many hubs, there is very little chat and they experience performance problems. I think search requests are what put load on them. The builtin chat is what gives some feel of community and sharing. (in contrast Napster was, and kazaa is just about ...

sandos

Ooops, seems my diffs were reversed. Ah well, just use -R. :)

sandos

The header aswell: --- 0.22/client/HttpConnection.h 2003-01-20 18:48:36.000000000 +0100 +++ 0.22-orig/client/HttpConnection.h 2002-07-24 15:56:50.000000000 +0200 @@ -48,7 +48,7 @@ { public: void downloadFile(const string& aUrl); - HttpConnection() : ok(false), port(80), size(-1), socket(NULL), moved...

sandos

My diff: --- 0.22/client/HttpConnection.cpp 2003-01-20 22:02:18.000000000 +0100 +++ 0.22-orig/client/HttpConnection.cpp 2002-12-28 02:31:50.000000000 +0100 @@ -32,8 +32,6 @@ void HttpConnection::downloadFile(const string& aUrl) { dcassert(Util::findSubString(aUrl, "http://") == 0); - ok = false; - i...

sandos

I did exactly what you did, and It only worked the first time around. It seems DC++ keeps the same object around till the next time it downloads, so the ok variable isnt reset. I just added ok = false; to the top of downloadFile(), and now it works every time. This might also be causing bugs in norm...

sandos

It seems to me you should probably fire the event up from the beginning, its probably cleaner. For example what if youve got multiple chained 302Ã‚Â´s, that wont work with your code, right?

sandos

I can think of some more minior protocol changes to but not so drastic as these ones. I think the DC community is great but its starting to become a chaos of difrent protocol hacks that is badly documented so i think some project that have a large influens in the community need to sit down and writ...

sandos

Just wanted to take this up, might be good to know if using tiger hashes for a tree hash:

http://www.shareaza.com/forum/viewthread.aspx?ID=4261

So care must be taken when implementing tigertree, and the correct one choosen aswell.

sandos

But what about a simple 32bit, 64bit, or 128bit CRC value. They are pretty much the same as hashes, differring mostly in number of bits, and cryptographical strength afaik. Hashes are generally slower though, but not that slow. Disk I/O will still be a bit of bottleneck, just as with CRCs, reading ...

sandos

ender wrote:Actually it should be Windows-1252, which difers a little from ISO-8859-1...

Thats weird, since all is fine using iso-8859-1 in java. I dont think there is any windows-1252 even. Hmm. :)

sandos

I'm not very interested in hashing, either.. I've been using dc for a little over a year now, and sure, sometimes you get a bad download, so what, get it again (and perhaps send the user a pm) currupted downloads is not a problem for me. maybe if I was on a slower connection, i'd care about it. cur...

sandos

ender: yes, the client needs to save hashes in some way. next problem, then the client needs to save a map between a hash and what file is called on DC. If you store this in memory i wouldn't be suprised if you dubbled the amount of memory consumed by my dc++ client (if you save it in a similiar wa...

sandos

Also, imho, we should just make a new search command including both hashes and metadata, much like whats in Luke-JrÃ‚Â´s proposal, whatever that is called, even though I think the command could be made simpler than his version. (I havent checked back on his drafts though, just doing this from my ver...

sandos

do-you-know-how-long-it-takes-to-get-a-hash-of-more-than-100000-files-ly'ers? eDonkey on linux, VIA Eden 533 MHz machine, 240 GB of data shared, it takes a little less than two hours to generate the hashes for the first time. When it's run the next time, the files are checked in around 20 minutes (...

sandos

so now i really dont know what shoul i fix... I written small DC+ hub - it shows everything what clients sends and i connected to it with my DC and DC++ and everything was the same. I thing that it must be same problem with sockets so i should find someone who knows Java... Ive done dc-programming ...

sandos

I really dont like the idea of hashing only the first 4k bytes though.

sandos

users should have the choice to enable this option for their files or not... I recall someone saying it took a while to /refresh his files... If the metadata must be extracted from them at the same time, it may be worse... Yup, the metadata stuff should not be entwined into the refreshing of the fi...

sandos

Another alternative also: if dc++ used hashes that can be found in the bitzi database, then all metadata can be found/serched/etc on bitzi, and only hashes need ever be sent using dc. Of course, depending on bitzi.com being alive isnt that great either.

DC++

Search found 177 matches

Re: GPL

Re: some ideas (not mine!) to discuss ...

Re: dc++ language

Address change

Warning, this post is on-topical!

Re: How to add feature request for ED2K Link support

Re: Hub List

Re: Bug in CryptoManager.cpp?

Re: hubs lists NOT working

Re: Total share in Hublist

Re: Publists Never WORKED

Re: What about a simple CRC