Same FILES from two different sources - SOLVED manually, but would like it automatic

[joshace]

Quite often, I will find the same FILES on more than one site.  

One site is typically faster than the other, or neither is complete, or other reasons.

Even though I download them with the SAME NAME and to the SAME DIRECTORY, when one of them completes and gets moved to a finished directory, the incomplete one is not "aware" of the other one.   I can force it to see it by moving files around, but that is tedious. 

It would be nice if files were aware of each other's "twin" and complete together.    This would also help complete torrents that are incomplete on another site.  

IDEALLY Utorrent would automatically combine the trackers of the two sources... I can do it manually, but that is a lot of work.  

Sometimes the file NAMES are not exactly the same, it would be nice if utorrent would either assign both to the same name, or allow the user to pick one name or the other. 

===================

UPDATE:  The MANUAL fix that works is this:

Torrent A contains the file     Star Wars.mp4 with a length of 654,127   with a tracker   https://bobstorrents.org/qwe7

Torrent B contains the file     StarWars-NewHope.mp4 with a length of 654,127  with a tracker  http://plb.site/97715

Torrent A has already been started

When loading Torrent B, I must change the download NAME to Star Wars.mp4 then click on ADVANCED and add the tracker from Torrent A which is   https://bobstorrents.org/qwe7 (with a blank line between the two trackers as is required)

Then remove Torrent A (but not it's data)

What you have done is taken the existing filename and tracker from Torrent A, and added them to Torrent B.  

======

SHOULD work.. but isn't

When I do the above, I do get the 2 trackers on the same file  BUT on the added one, it says UNREGISTERED TORRENT instead of WORKING

They should BOTH be working

[mike20021969]

Are the torrent's hashes the same though?

If the torrent's hashes are different, then the files within the torrent will not be (classed) the same.

So 2 different torrents might look identical, but they're most likely not.

[joshace]

Thanks for correcting me.  I meant the files contained in two different torrents are identical.. not the torrent file itself.  

I think a working MANUAL fix is to add the tracker from one of the torrents to the other, and nuke the other torrent.  That is a little tedious.  

The torrent's hashes would be different because they are not the same torrents... but the files the torrents are linked to are 100% identical. 

This reminds me of a murder trial in which their computer "experts" were ignorant.   They were telling the jury that if you found a file on two different computers with the same hashes, then the file had to come from the same person.  Nonsense.  a "hash" is a checksum to indicate that a file is the same rather than comparing every byte of the file to check that they are the same.     For a hash of a file to be identical, the file length has to be identical.  However, just because the file length is identical, the data might not be the same.     Two different pictures can have exactly the same length.  The checksum is a value created by adding up every byte of the data.  For instance.. in modular arithmetic...   the data  502 mod 20 would be  2  (502 - 25 rollovers of 20).   522 mod 20 would also be 2.. however  512 mod 20 would be 12.   So, it IS possible to have the same checksum even though the data is not completely the same, but so unlikely for a long file that it is virtually impossible.  

Anyway, in the trial, they misled the jury into thinking that the falsely accused person sent a file to multiple destinations because the hashes were identical.  WRONG!   If the hashes were different, then it would be impossible for the accused to be the source for all instances of the file.. however,  just because the hashes were the SAME doesn't mean squat.   If I sent a file to you, and you sent that file to someone else without editing it or re-encoding it and it still had the same file length.. and whoever received that file sent it someplace else... that file is always going to have the same hashes.  In that trial, it was absurd, because they blamed the accused person of sending that file, when he was in custody at the time it was sent, and in a different country.   The person who posted it did so from an account on opposite side of the planet (Australia) and continues to post messages and files to this day.  One of the charges against the accused was based upon him sending this file (which he did not).  Since the idiot computer experts told the jury he sent it, the accused got convicted of an additional charge of a few years... but it is far worse than that.  By having that additional charge, the man who would be eligible for parole is NOT eligible for parole because of a "dangerous offender" designation.   One tiny technical lie to the jury resulting in a sentencing that will keep this man in prison for roughly 50 years at which time he will die of old age.  

Don't worry.. I will eventually take care of it when I have the time and money to take a trip.  I already know who the actual killer was - who spent years framing the accused guy.  The actual killer committed suicide months after the murder after confessing.  The actual killer's confession was never heard by the jury.    

[ShockDiamond]

I like this idea as I too have tried to merge the same data from multiple sites to complete the data faster, but the issue in general is not public sharing - it's giving people the ability to "leak" data outside the tracker.

If your client is counting data in and out and one of the trackers is a ratio-based site, you're going to get screwed if your client is giving data out to a public tracker, while the ratio-based tracker is in effect.  It won't be counting all the data you're uploading towards ratio.

Private sites will also not appreciate you downloading data from their members and giving it to members of other trackers, especially public ones.  They're based on you taking and giving bandwidth into their own swarms to better-serve their "customers".

[geter123]

 

On 15 sierpnia 2017 at 9:57 PM, joshace said:

   It would be nice if files were aware of each other's "twin" and complete together.

 

 

 

 

Vuze have this option "Swarm_Merging" for duplicate files. Please don't ban me! 
https://wiki.vuze.com/w/Swarm_Merging

On 22 września 2017 at 3:29 PM, ShockDiamond said:

I like this idea as I too have tried to merge the same data from multiple sites to complete the data faster, but the issue in general is not public sharing - it's giving people the ability to "leak" data outside the tracker.

No. Vuze generate statistic for 2 trackers, don't margin  statistic. 
 

[Brandy Down Under]

I've had this issue with two different very large collections that had almost exactly the same content.  Biggest problem was that neither had a 100% seed.  What I did was to stop the torrents, exchange the files from one to the other, force a recheck, and then resume.  What missing files the one had existed in the other.  I continued this painful operation for several days until I had all the files in both torrents.  That made me a 100% seed for those two torrents.  I continued to seed until a number of other peers made it to 100%.

It would be nice if there was a way this could be done within uTorrent, but I realize that it would be arbitrarily complex.  Each torrent has its own block size defined when the torrent was created.  Each file in the torrent starts within a different block, so merging would super messy and have to be done at the file level anyway.  I hear that Vuze can do this, but I don't know what mechanism they use nor what limitations there are.  It took me long enough to become conversant with uTorrent, and I'm not inclined to start over with another client.

[vincent163]

With different file names the infohash would be different, and according to what I know, BitTorrent protocol currently doesn't seem to allow peers to share the same file contained in different torrents with different infohash. Under the current BitTorrent protocol it doesn't seem practical.

However, BitComet has a long-term seed feature which allows files to be hashed one-by-one and shared directly using BitComet's server. It also computes ED2K hash for all files and shares the files using eMule network, allowing similar function. Hopefully uTorrent would consider adding such kind of features.

[joshace]

I frequently rename files so that they match another site and they are compatible using Utorrent.  

 

For instance.. Torrent Site A has a file called  "Pharkas.avi" which is 901 megs in size... while Torrent Site B has a file called "Pharcis.avi" which is the same other than the name.  I rename the local file to be "Pharkas.avi" and somehow Utorrent is able to make them compatible.   I suspect it uses the filename Pharcis.avi" for the torrent, but translates it to Pharkas.avi just for my computer and for my Utorrent database.