Age | Commit message (Collapse) | Author |
|
fix: drop multi-stream
|
|
fix: use streamx instead of stream
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
* fix: web seed request URLs
When adding a multi torrent file (not magnet link) which has only web seeds, the torrent fails to download because webtorrent seems to try to call the wrong URL.
Suppose that we are using the default download path of `/tmp/webtorrent`, and we have a web seed at `https://some-s3-bucket.s3.amazonaws.com/my-torrent` where `my-torrent` contains some nested directory structure like this:
```
my-torrent/
file1.txt
file2/
file2.txt
```
Webtorrent will attempt to fetch the following URLs:
```
https://some-s3-bucket.s3.amazonaws.com/my-torrent//tmp/webtorrent/my-torrent/file1.txt
https://some-s3-bucket.s3.amazonaws.com/my-torrent//tmp/webtorrent/my-torrent/file2/file2.txt
```
This is obviously incorrect, but perhaps I'm doing something wrong to cause this. I dug through the code, and it seems that `fs-chunk-store` is [modifying `file.path`](https://github.com/webtorrent/fs-chunk-store/blob/master/index.js#L48) when the [chunk store is created](https://github.com/webtorrent/webtorrent/blob/master/lib/torrent.js#L485). Stripping off the temp path fixes this issue and the download progresses as expected.
* Update lib/webconn.js
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
.substr() is deprecated so we replace it with .slice() which works similarily but isn't deprecated
Signed-off-by: Tobias Speicher <rootcommander@gmail.com>
|
|
|
|
* chore: bump bittorrent-protocol to ^3.5.0
* feat: implement BEP6 Fast Extension
|
|
* prep for esm
* update min req node vers
* revert node prefix
|
|
* fix: add storeOpts, specify store path, align with docs
* make changes non-breaking by default
* debatable documentation update
* add torrent name to store name
* reduce hash name
* Update lib/torrent.js
* fix filegetmodtimes
* bump fs chunk store
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
|
|
* feat: add service worker server as an alternative to renderMedia
* code QL
* thanks gh auto merge
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
* reimplement #1384
* add option for secure - performance issue
* fixes after code review
* fix error
* use const in hex
* use stored hash
* Update lib/peer.js
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
* fix const
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
* Add speed limit to client
* Fix standard
* Update docs/api.md
* Add changes from PR feedback
Co-authored-by: Kadu Diógenes <kadu@fnix.com.br>
Co-authored-by: Ivan Gorbanev <ivang@van.work>
Co-authored-by: ultimate-tester <jordimueters@hotmail.com>
Co-authored-by: Julen Garcia Leunda <hicom150@gmail.com>
Co-authored-by: Niklas Johansson <niklas.y.johansson@se.abb.com>
Co-authored-by: ThaUnknown <kapi.skowronek@gmail.com>
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
Previously, early exits in FileStream._notify would
throw since cb was not defined in _destroy
|
|
* store.put is async and might fail
* Remove retry logic for now
Co-authored-by: John Hiesey <john@hiesey.com>
|
|
* fix: modernize code
* standard fix
|
|
* Support adding paused torrents.
* chore: use arrows, because semantic
Co-authored-by: Nas Kavian <>
Co-authored-by: Diego Rodríguez Baquero <github@diegorbaquero.com>
|
|
|
|
The fileStream._destroy() function would never run because when fileStream.destroy() is called then it sets a 'destroyed' getter to true, which shadows our own fileStream.destroyed property.
The code outside running eos() was reaching into fileStream internals, so I just moved it in there.
|
|
|
|
|
|
* Add optional utp-native
* Add error log when uTP cannot be loaded
* Fix review suggestions
* Remove an additional require call
Co-authored-by: Alex <alxmorais8@msn.com>
|
|
|
|
Made 3 arrays into a set
|
|
This is a non-standard interface and it's confusing. It caused a bug in https://wormhole.app
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Makes the EventEmitter warning show up less often, and avoids running potentially 100s or 1000s of chunkStore.get() calls concurrently in _verifyPieces
|
|
|
|
|
|
Code reviewed by @jhiesey
|
|
|
|
|
|
|
|
|
|
|
|
|