OK, get rid of this hashing/remove duplicates feature please

Archived discussion about features (predating the use of Bugzilla as a bug and feature tracker)

Moderator: Moderators

Locked
Deathbliss
Posts: 9
Joined: 2003-07-22 03:53

OK, get rid of this hashing/remove duplicates feature please

Post by Deathbliss » 2004-06-13 22:26

I've had enough... I have to put up with hub owners throwing absoluetly ridiculous minimum share requirements at me (150 gig was the worst so far), DC++ forgetting all that I have shared so I have to do it over, and a clunky POS Favorite Hub system. The last straw is that I have noticed recently with this latest build of DC++ that you have some sort of duplication removing system in place. HUB OWNERS DON'T NEED YOUR HELP! They can go and make their own fucking scripts if they want something like this - it should not be hardcoded into the program! I loose 20 gigs of things to share, and THEY AREN'T EVEN DUPLICATES! I tested this - I have for example some music in folders listed by author and album. Each song starts with the album name, then moves on the the song number, and the song title. DC++ is seeing these as duplicates and removing them! They are valid files, that I won't and shouldn't have to rename, and I have no control over this stupid, idiotic feaure.

Please address this in the next release. I am aware that the program coder's laptop has been stolen, which has probable delayed things. I'f I had the money I'd throw it their way and tell them to get another. I don't, so I have to wait until eveything is up and running again. That's fine, all I ask is that you make remove the hashing and duplication feature, and never add any unecassary feature like this which benifits only hub owners again.

Thank you -
- Deathbliss

Gratch06
Posts: 141
Joined: 2003-05-25 01:48
Location: USA

Post by Gratch06 » 2004-06-13 23:08

Perhaps you could go into settings and turn off the "Remove Duplicate Files" checkbox and that would solve your primary problem?

Todi
Forum Moderator
Posts: 699
Joined: 2003-03-04 12:16
Contact:

Re: OK, get rid of this hashing/remove duplicates feature pl

Post by Todi » 2004-06-14 01:30

Deathbliss wrote:I've had enough...

So stop using DC++ then, it's not like we're forcing you to use it. Or better yet, make your own personal mod with only the features you want in it.. the joys of open source.

Deathbliss wrote:I have to put up with hub owners throwing absoluetly ridiculous minimum share requirements at me (150 gig was the worst so far)

I actually checked this on hublist.org once, and it turns out that about 0.01% of the available hubs have > 120GB requirements (i don't remember exactly, but it was around there). I'm sure you could get all the files you want in hubs with lower requirements..

Deathbliss wrote:The last straw is that I have noticed recently with this latest build of DC++ that you have some sort of duplication removing system in place.

No actually it's always been there. Or atleast since way, way, waaay back. However, in DC++ 0.401 it was changed so that instead of detecting duplicates through filesize and filename, it does so by hashing the file and comparing the hashes. There has not yet been a case of hash collision (two files having the same hash) using the Tiger Tree Hashes, so i'd say the system has been greatly improved. Go hashing!


Deathbliss wrote:...rant... They are valid files, that I won't and shouldn't have to rename, and I have no control over this stupid, idiotic feaure.

If you want to submit a bug report, this is not the proper way. However, unless there is something seriously wrong with the hashing and hash comparing, i'm afraid these files actually are duplicates, and renaming them won't do any good at all (see previous statement).

Deathbliss wrote:Please address this in the next release. ..... That's fine, all I ask is that you make remove the hashing and duplication feature, and never add any unecassary feature like this which benifits only hub owners again.

Actually, i like these features. And to be honest with you, most users do. So it's highly unlikely you will see these features removed any time soon (read: ever). Once DC++ 0.402 is out you'll notice that YOU as a user will benefit from hashing much much more than hubowners ever will, and perhaps you'll come around. Until then, there is always the option i mentioned at the beginning.

GargoyleMT
DC++ Contributor
Posts: 3212
Joined: 2003-01-07 21:46
Location: .pa.us

Post by GargoyleMT » 2004-06-14 18:14

Removing duplicate files from your file list is already an option. They will never count towards your share size, but that's fine, since they're byte-for-byte duplicates (that's the purpose of hash functions - take megabytes of data and make a number out of it that has an infinitesimal chance of being the same number that a file with different contents generates.)

It looks like your concerns are already addressed.

Deathbliss
Posts: 9
Joined: 2003-07-22 03:53

No...

Post by Deathbliss » 2004-06-15 01:52

You didn't listen... These are different songs from the same album which are not counting towards my share, and I do not appreciate this. I SHOULD HAVE COMPLETE AND UNEQUIVACAL CONTROL OVER MY SHARE! If hub owners don't like it that's just tough.

I am curious about something... How in the hell does hashing help me? Explain this please - I would dearly love to know how removing my control helps me in any plausible way.

Oh and yes, I would make my own program or mod this one if I could program. You guys seem to be forgetting that not all of us are blessed with however many years of programming knowledge we'd need to be able to make our own stuff.

I left Direct Connect for many reasons, most of them obvious to anyone else who ever used that peice of garbage. I will leave this one too if I have to, but I don't want to and I shouldn't be forced to just because my concerns and desires are somehow considered irrelevant. If hashing is so great and this duplicate removal thing then let us turn it of or on in the options. Then everyone is happy - that's the way it should be.

Oh and if I did want to share duplicates I'd think it perfectly fair if it was the only way into a good hub. I should be able to defeat a hub owner's idiotic rules if I choose to... That is any rules that exclude others for some silly reason or another (AKA you don't have 150 gigs of files). There are other rules that I can see the reason for, they are not unreasonable and I see no problem with abiding by them.

Just because someone has power doesn't mean that they can use it however they want, and anyone that feels differently should go and join the RIAA, or perhaps the CIA.
- Deathbliss

Xan1977
Forum Moderator
Posts: 627
Joined: 2003-06-05 20:15

Post by Xan1977 » 2004-06-15 02:21

Deathbliss wrote:You didn't listen... These are different songs from the same album which are not counting towards my share, and I do not appreciate this.

Todi wrote:There has not yet been a case of hash collision (two files having the same hash) using the Tiger Tree Hashes, so i'd say the system has been greatly improved. Go hashing!

Don't argue with the TTH algorithm, it will beat you every time. :wink:

Deathbliss wrote:How in the hell does hashing help me?

It prevents your from downloading corrupt/fake files. Dupe checking and search by TTH are the two current benefits of hashing. More cool stuff, such as www.bitzi.com are possible. I'm sorry, but hashing will not, and should not, be optional. You are free to plea for a change to that, but I completely disagree with you.

Deathbliss wrote:I should be able to defeat a hub owner's idiotic rules if I choose to...

No you shouldn't. If you don't like and/or comply with a hubs rules, move on to another hub that suits you better. Isn't that one of the great attributes of the DC community? If you think that a hub has stupid rules, don't go to that hub.

Todi
Forum Moderator
Posts: 699
Joined: 2003-03-04 12:16
Contact:

Re: No...

Post by Todi » 2004-06-15 04:25

Deathbliss wrote:You didn't listen...

Right back at ya.

Deathbliss wrote:Oh and yes, I would make my own program or mod this one if I could program. You guys seem to be forgetting that not all of us are blessed with however many years of programming knowledge we'd need to be able to make our own stuff.

The options are clear:
1. Spend some time and learn to program. It will have more benefits than just changing stuff in DC++.
2. Pay someone to do the changes.
3. Argue your case and convince someone who knows how to program to make the changes. In your case, i doubt this will work.

Deathbliss wrote:I left Direct Connect for many reasons, most of them obvious to anyone else who ever used that peice of garbage.

Correction: Direct Connect is the p2p network you are (still) using, perhaps you mean the Neo-Modus Direct Connect client?

Deathbliss wrote:If hashing is so great and this duplicate removal thing then let us turn it of or on in the options. Then everyone is happy - that's the way it should be.

It is actually so great and wonderful that there is no reason to make it optional, since it benefits everyone much more if it's mandatory. And then everyone will be happy, except perhaps you. But honestly, i can live with that.

Deathbliss wrote:Oh and if I did want to share duplicates I'd think it perfectly fair if it was the only way into a good hub. I should be able to defeat a hub owner's idiotic rules if I choose to... That is any rules that exclude others for some silly reason or another (AKA you don't have 150 gigs of files). There are other rules that I can see the reason for, they are not unreasonable and I see no problem with abiding by them.

Going from bad to worse fast. What would the purpose of rules/laws be if everyone just followed the ones they thought were reasonable? They would all be useless then. Your reasoning clearly shows you're not a dedicated sharer, just a leech. If you think anyone will help you after this statement, you are clearly deluded.

GargoyleMT
DC++ Contributor
Posts: 3212
Joined: 2003-01-07 21:46
Location: .pa.us

Re: No...

Post by GargoyleMT » 2004-06-15 17:52

Deathbliss wrote:Oh and if I did want to share duplicates I'd think it perfectly fair if it was the only way into a good hub. I should be able to defeat a hub owner's idiotic rules if I choose to...


Just because there are hub rules that you cannot follow does not mean that DC++ is supposed to help you circumvent them. Sorry - that's the way it is.

Deathbliss
Posts: 9
Joined: 2003-07-22 03:53

You guys are just worthless...

Post by Deathbliss » 2004-06-16 18:06

I shouldn't have vented here in the first place, I've been foruming long enough to know better. So I give up. I've been using Direct Connect since the earliest versions of Neo Modus, and yes you are correct that was what I referring to. I know what's good and bad, what works and what doesn't. If you want certain files you'll only be able to get them in the top 10 hubs. Every hub afterwards gets worse and worse. So if the ones in teh top ten that you can acces don't have what you want, then what do you do? Plus what if you suspect someone you've ben downloading from is in one of the hubs you can't access because you don't have a big enough share?

Enough. DC++ is no longer user-friendly - it's as simple as that. Anything that's forced on you in a program to your disadvantage makes a peice of software non user friendly. Somewhere in the development of this program the hub owners have been considered more important. Unfortunatley most are stupid little peices of shit that misuse any authority they finally manage to get their grubby hands on.

Whatever... I don't need the stress, and for the record I'm not a leech. I have never shared duplicates, and if for some strange reason I kept using DC++ - which is unlikely - I never will. But I also don't have 100 gigs of shit to share, and when my legitamite share is cut down from 100 gig to 80 gig there's definitely something wrong.

It's as simple as that -
- Deathbliss

cologic
Programmer
Posts: 337
Joined: 2003-01-06 13:32
Contact:

Post by cologic » 2004-06-16 19:46

The minshares paragraph contains nothing but uninteresting moaning and whining, so I'll ignore it.

If DC++ falsely marks 20GiB of data as duplicate, even though it's not, that's a bug; can you show it's doing this?

Deathbliss
Posts: 9
Joined: 2003-07-22 03:53

Yes.

Post by Deathbliss » 2004-06-17 03:54

I've watched the little text readout, and when it hits the music folder it says, "such and such will not be shared" and something about duplicates. I suppose I can take a screenshot or send a log or something - just tell me what to do.
- Deathbliss

Todi
Forum Moderator
Posts: 699
Joined: 2003-03-04 12:16
Contact:

Post by Todi » 2004-06-17 04:30

In your Log settings, turn on System Logging. Then you can copy/paste from the system.log file.

GargoyleMT
DC++ Contributor
Posts: 3212
Joined: 2003-01-07 21:46
Location: .pa.us

Post by GargoyleMT » 2004-06-17 13:09

Once you have the filename, you can search your own system by file size and find which files are duplicates. Rest assured, DC++ is right, they are duplicates. I recently added a download directory to my share, and had to go through weeding out duplicate files - DC++'s hashing and system log (why I put that option in) were immensely useful to that end.

Coder
Posts: 7
Joined: 2004-09-21 06:34

No you are NOT right GargoyleMT

Post by Coder » 2004-09-21 06:44

,.,
Last edited by Coder on 2004-09-21 09:05, edited 1 time in total.

ivulfusbar
Posts: 506
Joined: 2003-01-03 07:33

Post by ivulfusbar » 2004-09-21 06:59

Coder: You are utterly wrong. The application that you have used does not inspect the content of files. This makes your comment and post utterly useless.
Everyone is supposed to download from the hubs, - I don´t know why, but I never do anymore.

Todi
Forum Moderator
Posts: 699
Joined: 2003-03-04 12:16
Contact:

Re: No you are NOT right GargoyleMT

Post by Todi » 2004-09-21 07:15

Coder wrote:sorry but i AM right.

Always nice to see someone be so utterly convinced they are right, when they're actually totally wrong =)

Coder
Posts: 7
Joined: 2004-09-21 06:34

Post by Coder » 2004-09-21 08:42

sorry maybe it doesn't hash files but i know i have no doubles. i checked my files there ARE NO DUPLICATE files

hashing is obviously wrong with dc++ my share is much smaler and ppl don't find their files anymore so what more proof do we need?

Coder
Posts: 7
Joined: 2004-09-21 06:34

Post by Coder » 2004-09-21 08:47

i checked the system.log of dc++ some doubles it so called identified aren't even of the same file size so stop calling me utterly wrong

Coder
Posts: 7
Joined: 2004-09-21 06:34

Post by Coder » 2004-09-21 08:50

,.,.,.,.
Last edited by Coder on 2004-09-21 09:07, edited 1 time in total.

cologic
Programmer
Posts: 337
Joined: 2003-01-06 13:32
Contact:

Post by cologic » 2004-09-21 08:57

Silly post-spamming aside (4 consecutive posts in 8 minutes?), actually finding a Tiger [tree] collision would be very, very interesting. Could you provide more specific data?

Coder
Posts: 7
Joined: 2004-09-21 06:34

Post by Coder » 2004-09-21 09:00

i said it before

i do not share doubles

i never share eg one song twice so i can't understand what duplicates i should havge it cannot be right

Coder
Posts: 7
Joined: 2004-09-21 06:34

Post by Coder » 2004-09-21 09:01

lol how more specific should i be

tell me how

the only however very unlikely explanation could be that one shareware program is common to 2 or more collections eg one is an collection that is release by one shareware site the other one is a sampler from another shareware firm that released some compilation but the hash should not be the same because eg the archive format is different and unique and therefore the hash MUST be different
Last edited by Coder on 2004-09-21 09:12, edited 2 times in total.

cologic
Programmer
Posts: 337
Joined: 2003-01-06 13:32
Contact:

Post by cologic » 2004-09-21 09:04

(Wonders how do to this while treading lightly on copyright law.)

Huh. Can you provide an example of two files legal to share where you can show the system.log messages and put both files up for download? You provide no way, currently, to verify your claims, and the stubbornness due to your potentially misplaced self-confidence hinders investigation.

Edit: and another two posts, there. Please stop. 1 post at a time. This isn't an IM/IRC/DC chat.

TheParanoidOne
Forum Moderator
Posts: 1420
Joined: 2003-04-22 14:37

Post by TheParanoidOne » 2004-09-21 14:47

Also please stop going back and erasing what you have said.
The world is coming to an end. Please log off.

DC++ Guide | Words

GargoyleMT
DC++ Contributor
Posts: 3212
Joined: 2003-01-07 21:46
Location: .pa.us

Post by GargoyleMT » 2004-09-22 14:15

I'm waiting, I would like to see two non-identical files that have the same tiger tree hash.

As an aside: I use duplic8 for duplicate file checking. it doesn't claim to use hashes, but it does compare files with the same size.

Todi
Forum Moderator
Posts: 699
Joined: 2003-03-04 12:16
Contact:

Post by Todi » 2004-09-23 01:42

I'm waiting too.. and the suspense is killing me! =)

Carraya
Posts: 112
Joined: 2004-09-21 11:43

Post by Carraya » 2004-09-23 02:30

me too...
<random funny comment>

Locked