qBittorrent unable to handle very large torrents #8449
Comments
|
@arvidn |
|
This is due to default limits on number of tokens and tree-depth of bencoded structures when parsing it. It's to prevent an old exploit where a recursive-decent parser could cause a stack overflow by just sending lots of "l" (for list) or "d" (for dictionary). These default limits can be specified when calling If that's the case, I would suggest separating the |
|
See |
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is changed to 10000000.
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is changed to 10000000.
|
I've submitted PR #8455. |
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is changed to 10000000.
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is also changed to 10000000.
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is also changed to 10000000.
`torrent_info` constructor has default limits that can't be changed via parameters, so we handle the loading process manually and explicitly specifiy the limits to `bdecode()`. The token_limit is also changed to 10000000.
Fix loading very large torrents. Closes #8449
|
@MrEldritch |
qBittorrent version and Operating System
qBittorrent v4.0.4 (64 bit); Windows 10 Version 10.0.16299 Build 16299
What is the problem
When attempting to download a very large torrent (Danbooru2017, a extremely-large-scale tagged-image dataset intended for machine learning, found at https://www.gwern.net/docs/anime/danbooru2017-torrent.tar.xz), I encounter an error:
Failed to load the torrent: (path to torrent file). Error: bencoded item count limit exceeded.What is the expected behavior
The .torrent file itself is not defective; I have successfully added and downloaded the torrent in another client (Transmission). I would expect to be able to add and download this .torrent in qBittorrent, as I would normally do with any other .torrent file.
Steps to reproduce
-Download and extract the .torrent files from the linked archive
-In qBittorrent, File>Add Torrent File and select one of the included .torrents
-Error displays.
Extra info(if any)
Danbooru2017 is really unusually large, as far as datasets go - each of the .torrents contains ~300k images weighing in at around ~200GB. It is entirely possible that this might not be a "bug" at all, but that it's simply some deliberate size-check exceeding a bound set by someone who assumed nobody would ever intend to download that many items. However, since the .torrent is clearly well-formed and compatible with other clients (such as Transmission), and since clearly torrents large enough to exceed that limit do exist in practice, it seems like qBittorrent really ought to be able to handle it.
Presumably any other sufficiently large torrent would trigger this error; this is just the first time I have ever needed to download a collection this large.
(screenshot of the Torrent Properties window from Transmission, to give specific numbers on size)
The text was updated successfully, but these errors were encountered: