when adding a duplicate torrent, harvest its info for the extant torrent to use

[mattia-b89]

I'd like to have transmission capable of adding new tracker from torrent that point to the same file.
e.g. I added a torrent via the magnet link; I want to add trackers contained in the .torrent version of the same file; actually, transmission only says that it can't add the file because already in download

I'm on Linux, transmission-2.92

[haywirephoenix]

I would like this feature too on windows client.
"Torrent already downloading, add the trackers from this list?"

[Aerozolic]

Any update to this improvement request?

[linuxgirl22]

+1 for this also (linux).

[lps-rocks]

I would make it so if the torrent is marked as private additional trackers can't be added. It's common for people to do this not understanding the ramifications.

[JustMyGithub]

I would make it so if the torrent is marked as private additional trackers can't be added. It's common for people to do this not understanding the ramifications.

Would you mind explaining this? If it is a private torrent, no peer exchange will happen, so what is the matter?

[lps-rocks]

I would make it so if the torrent is marked as private additional trackers can't be added. It's common for people to do this not understanding the ramifications.

Would you mind explaining this? If it is a private torrent, no peer exchange will happen, so what is the matter?

Private trackers often use the statistics from the 'announces' to track a user's ratio. Adding multiple trackers to the same private torrent makes the same statistics get announced to both trackers. However one tracker may have a more active pool than the other. This causes the user to get upload credit on a site they didn't contribute to the swarm on. This then causes them to be flagged as ratio cheaters.

Unless transmission can track the statistics on a per-tracker and per-peer basis and keep the statistics separate then this change should be prohibited on 'private' torrents. That fix however inherently breaks the whole point of multiple trackers / tiers of trackers in a torrent. It's for redundancy not cross-seeding.

[JustMyGithub]

Thanks. I did silently expect a per-tracker statistic.

[ckerr]

I like this idea for public torrents!

Unless transmission can track the statistics on a per-tracker and per-peer basis and keep the statistics separate then this change should be prohibited on 'private' torrents.

Completely agree that this feature should not be used on private torrents. That would probably get Transmission banned off private trackers.

In addition to that, it's common practice these days for private trackers to ensure their torrents have a unique info hash by injecting a "source" entry into the torrent's info dictionary. This helps people to cross-seed by giving otherwise-identical torrents a unique per-tracker info hash. So even if Transmission did allow this feature for private trackers (it won't) it still probably wouldn't be an issue in practice.

[ckerr]

#787 has another variation of this but it's essentially the same issue. I'm going to close it as a duplicate and modify the description for this ticket

[Tockman]

So, do I understand right that the feature won't ever be implemented?

[ckerr]

So, do I understand right that the feature won't ever be implemented?

If that were the case, this issue would have been closed with a declined label.

It's still open with a PR welcome label because if PRs for this feature are welcomed! This one probably wouldn't even be that difficult, if anyone wants to take a stab at it I'm happy to answer questions.

[drewoprea]

I think the proposed handling of torrents not marked by the private flag can be made to work.
I came across this issue when considering ways to enable cross seeding on private trackers. I imagine we can allow two torrents to read the same data while behaving separately for seeding / announce purposes. (I've had great results when different tracker's torrents expect the same data, but don't have a hash collision, so my desired behavior "just works") I'll have look harder to understand the implications of downloading from two or more swarms.