Add network timeouts and more robust error handling across scrapers/hosts, improve logging concurrency, and refine download handling:
- Hosts/sources: add request timeouts (15s) and exception handling to buzzheavier, gofile, steamrip, uztracker and rutracker; use urllib.parse.quote for search queries and better parse error checks.
- GoFile: use Session with try/finally, create guest account with timeout, validate API responses and close session.
- Context menu: import DirectDownloadHandle and handle stopping direct downloads separately from libtorrent; simplify file deletion with single try/catch; improve logging on exceptions.
- GUI/search flow: move UI state changes (clear table, disable searchbar, hide empty results) into MainWindow._start_search and reduce duplicated UI changes in run_search.
- Main: remove duplicate is_running assignment.
- Libtorrent: snapshot seeded magnets in update loop to avoid concurrent access; minor indentation change in add_seed.
- State/logging: add _log_lock to AppState; protect state.log_buffer with lock in consoleLog and flush_log_buffer and fix flush semantics.
- Updater: replace busy-wait with QEventLoop/QTimer polling for DirectDownloadHandle progress, stop timer and quit loop on completion/error, and remove unnecessary sleeps.
These changes aim to make network operations more reliable, prevent UI freezes, avoid race conditions when logging, and handle direct vs libtorrent downloads more cleanly.
Replace brittle index-based host resolution in SteamripScraper with substring checks on the chosen download URL (buzzheavier, gofile, vikingfile, megadb). Add a console log for the resolved link and adjust returned values for gofile. Remove the unnecessary links.index lookup and its exception handling.
Simplify SteamripScraper link/name extraction by using BeautifulSoup results directly (append absolute URLs and use link.get_text()) and avoid fragile regex; adjust scrape_steamrip_game_downloads to accept a full URL argument. Update context menu text: replace "Copy Magnet URI" with "Copy Download Link" in two places and rename "Download Item" to "Download" for clearer UI wording.
Improve thread-safety, logging, and various bug fixes across the app.
Key changes:
- Add and use state.downloads_lock around accesses to state.active_downloads in multiple modules (UI, libtorrent, misc, direct downloads) to prevent race conditions.
- Prevent concurrent dl_status_loop runs using a non-blocking _loop_lock and ensure proper iteration over a snapshot of downloads.
- Add metadata timeout when waiting for torrent metadata and fix free-space check logic to remove or reject downloads appropriately.
- Refactor download log persistence: add _load_downloads and _save_downloads helpers and centralize JSON read/write logic in utils/logging/logs.py.
- Optimize DirectDownloadStatus speed window by switching to deque and using popleft to trim older samples.
- Trim console log widget to 500 lines to avoid unbounded growth.
- Add small state fields for custom image positioning (x/y and enabled flag).
- Fix Steamrip scraper caching behavior and make Uztracker return an empty list on fetch failures.
- Add defensive checks in run_download to handle missing links.
These changes focus on stability and correctness for concurrent download handling and durable logging.
Rename scraper method get_magnet -> get_download_link in monkrus, rutracker and uztracker. Update downloader code to call .get_download_link and use the .is_magnet flag on tracker objects (replace dict access). Switch direct_download package imports to absolute paths and perform a small import reorder in direct_download/utils.py. These changes unify the download API and adapt call sites to the new tracker interface.
Vxrtrauter
KeksNino