Skip to content

v7.13.0-RC2

Pre-release
Pre-release
Compare
Choose a tag to compare
@kozlovsky kozlovsky released this 01 Jun 19:55
· 847 commits to main since this release
v7.13.0-RC2
0371e2a
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

Overview

This release has important bug fixes that improve Tribler speed and stability.

What's Changed Since the Latest RC1 Release:

Stability fixes:

  • Properly stop the non-primary process by @kozlovsky in #7349
  • Detect slow coroutines and freezes in the asyncio loop by @kozlovsky in #7414
  • Fixes #7344: AssertionError "transport is not None" in aiohttp/web_request.py by @kozlovsky in #7385
  • Temporarily increase the core manager API port check timeout to be sure it is big enough by @kozlovsky in #7419
  • Profiling of slow coroutines by @kozlovsky in #7446
  • Fix the step interval of DiscoveryBooster by @kozlovsky in #7449
  • 100x speedup of EdgeWalk.take_step() by caching results of GigaChannelCommunity.guess_address() by @kozlovsky in #7450
  • Add extended logging to RESTManager.site.start() call by @kozlovsky in #7451
  • Send events to GUI only before shutdown and in the proper order by @kozlovsky in #7364
  • Skip on_tribler_exception if EventsEndpoint is shutting down. by @drew2a in #7357
  • Decrease the frequency of the downloads request by @drew2a in #7377
  • Check on the ancient version by @drew2a in #7383
  • Fix #7386 "The connection to the Tribler Core was lost" by @xoriole in #7387
  • Ignore old responses for the DownloadsPage by @drew2a in #7391
  • Stop sending requests after shutdown by @drew2a in #7405
  • Suppress ConnectionResetError by @drew2a in #7404
  • Make KnowledgeRulesProcessor less aggressive by @drew2a in #7418
  • Add queue for multithread processing by @drew2a in #7417

UI-related fixes:

Core-related fixes:

  • Fix invalid handle when there is no torrent metadata by @xoriole in #7415
  • Use a separate misc table for the KnowledgeDB by @drew2a in #7413
  • Add reprocess feature for KnowledgeRulesProcessor by @drew2a in #7379
  • Add CONTENT_ITEM relation by @drew2a in #7445

Test, build, and reports fixes:

  • Disable logging in the faker module for pytest. by @drew2a in #7348
  • Disable logging in tests for faker.factory and PyQT5.uic by @drew2a in #7384
  • Fix test check core API port timeout by @kozlovsky in #7420
  • Add tribler_stopped tag to the Sentry by @drew2a in #7365
  • Sentry: add the field additional_information by @drew2a in #7394
  • Add an option to cease the collection of breadcrumbs by @drew2a in #7426
  • Increase the size of log files from 100 kilobytes to 1 megabyte, and the number of log backups from 1 to 2 by @kozlovsky in #7455
  • Pin pip version for the macOS build by @drew2a in #7375
  • Remove events and requests from Sentry report by @drew2a in #7376

What's Changed Since the Tribler 7.12:

Overview

We're ecstatic to introduce the latest version of Tribler, packed with exciting features and significant enhancements! Here's a rundown of the most notable changes:

Instant search with superior result quality:

  • Experience lightning-fast local search queries, now 100x faster than in the previous release (because of broken feature) (#7166);
  • Enjoy real-time, remote results from other peers as they appear instantly (#7025);
  • Benefit from an advanced ranking algorithm that considers multiple factors, such as title structure, creation date, and the number of seeders/leechers (#7025);
  • Discover similar results for known content categories, now conveniently grouped for an improved UI experience.

We're confident that you'll love the revamped search experience. As more peers upgrade to the new version, the distributed search will deliver even higher-quality results with the enhanced ranking system!

With these updates, Tribler is gradually evolving into a decentralized search engine, rivaling the speed of YouTube's video search.

Introducing the Knowledge Component: A Leap Towards Decentralized Search Excellence

Google's power resides in its PageRank algorithm. Creating a decentralized alternative to Big Tech has proven challenging, but with this release, we take another small step on this journey. Our ultimate goal is to achieve self-organizing trust, the fundamental building block for a decentralized PageRank system. The concrete use-case we aim to address is collective decision-making in the peer-to-peer world, such as determining whether Justin Bieber is gay.

Is Justin Bieber gay?

Explore the details of this scientifically serious issue related to trust, spam, and crowdsourcing. For the record, we believe Justin Bieber's music is neither Black Metal nor gay. We consider this an exemplary use-case for Tribler's zero-trust architecture to solve. Our distributed trust ledger has been under incremental development since before Bitcoin. We have not yet deployed our decentralized reputation and trust framework, as our MeritRank algorithm is still not ready for production. However, we have upgraded the BitTorrent tag-based metadata enrichment from pure text labels into the KnowledgeComponent (#7070). As a test, we offer semantic clustering within Tribler when searching for "Ubuntu," providing cleaner search results and filtering duplicates.

No other peer-to-peer project, such as IPFS, Mojo Nation, or Freenet, has ever cracked semantic search. With the KnowledgeComponent, we take another challenging step towards a Google-quality search engine. It signifies a move towards semantics, a search engine that understands your preferences and helps you discover content without interference from lawyers or governments.

Creative Commons licensed promotional picture:

Enhanced Torrent Health: Accurate Sharing Between Peers

Tribler peers exchange information about torrent popularity, allowing torrents with a substantial number of seeders to appear on the "Popular" page and receive priority in search results. We identified several critical bugs (#7295) in the previous algorithm, causing invalid torrent health records to spread across the network and skew the results on the "Popular" page. These bugs have now been fixed, and the gossiping algorithm has been optimized for efficient sharing of popular torrents (#6950).

Improved Stability for a Smoother Experience

This release of Tribler is set to be the most stable in recent times. Our development team has successfully identified and resolved numerous obscure bugs that previously led to crashes and freezes. We're confident that this version of Tribler will impress you with its enhanced stability and reliability.

Get paid to work on Tribler and earn a phd?

See open position: PhD Decentralised Machine Learning

  • Challenge: Developing and building a fully decentralised video search engine.
  • Change: Eliminating the need for servers when developing machine learning.
  • Impact: Safe machine learning benefiting billions of users.

Features

Bugfixes

(click to view)

Code improvements

(click to view)

Build/CI improvements

(click to view)

Test improvements

(click to view)

Thank You for Your Support

We sincerely appreciate your continued support and feedback, which help us make Tribler better with every release. Your engagement drives our mission to create a powerful, decentralized platform that respects users' privacy and freedom. We encourage you to download and try the latest version, and as always, please share your thoughts and suggestions with us. Together, we can make Tribler the ultimate peer-to-peer experience. Happy sharing!

Contributors

vi, xoriole, and 6 other contributors
Assets 16