yutto-dev/yutto - CodSpeed

yutto

Blog Docs Changelog

Performance History

Latest Results

✨ feat: decouple the download core and add an experimental server mode (#748) Co-authored-by: Codex <noreply@openai.com> Co-authored-by: MrXnneHang <xnnehang@gmail.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

main

13 hours ago

🐛 fix: harden FFmpeg merge defaults Co-authored-by: Codex <noreply@openai.com>

codex/core-server-architecture

13 hours ago

🐛 fix: make wire paths platform-stable and user-info cache semantics explicit - Serialize Path values with as_posix() (and dump models in python mode so Path fields reach the shared serializer): the RPC wire format no longer emits backslash paths when the server runs on Windows - Reject drive-relative/rooted output paths via path.anchor in ServerPolicy and ensure_output_path_is_scoped, matching the subpath-template check; on Windows `/x` now fails with `must be relative` instead of relying on the resolve() fallback - Check the request-level user-info cache before building an httpx client in validate_user_info, and reword the mid-batch revalidation comment in the download loop: with the cache introduced by this PR the check no longer re-fetches, so credential expiry within one context is not re-detected (documented tradeoff instead of a misleading comment) - Make test path assertions platform-neutral; the offline suites now pass on Windows (190 passed, previously 6 platform failures invisible to the ubuntu-only CI) - Document that --token-file auto-creates a 0600 token file on first start Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

codex/core-server-architecture

14 hours ago

♻️ refactor: split EpisodeData into listing-time info and lazily resolved data Hoist the identity and planned-path information (avid / cid / name / cover_url / path / display_group) out of the per-episode lazy coroutine closures into a new EpisodeInfo, available at listing time. Extractors now return ResolvableEpisode(info, data_coro) pairs; the data coroutine keeps the exact same fetch sequence (playurl, subtitles, danmaku, cover) and EpisodeData embeds info alongside the untouched resource fields. CLI behavior is unchanged: same request count, same fetch order, same terminal output. This is the internal half of the parse/download separation discussed in #744; no new verb or wire surface is added. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

codex/core-server-architecture

18 hours ago

Update src/yutto/utils/filter.py

codex/core-server-architecture

19 hours ago

🎨 style: format server documentation Co-authored-by: Codex <noreply@openai.com>

codex/core-server-architecture

20 hours ago

🐛 fix: harden task runtime and server transport Co-authored-by: Codex <noreply@openai.com>

codex/core-server-architecture

1 day ago

✅ test: avoid platform-specific datetime messages Co-authored-by: Codex <noreply@openai.com>

codex/core-server-architecture

1 day ago

Latest Branches

-15%

✨ feat: decouple the download core and add an experimental server mode#748

13 hours ago

8b9fed0

codex/core-server-architecture

-11%

⬆️ deps: Update dependencies (non-major)#747

1 day ago

df9d3ac

renovate/dependencies-(non-major)

-12%

⬆️ deps: Update Python dev dependencies#746

1 day ago

71a8d33

renovate/python-dev-dependencies

© 2026 CodSpeed Technology

Home Terms Privacy Docs