Repository navigation
docs(architecture): delete Known gaps that v2.0.3 closed - #2014
Conversation
The v2.0.3 release built and packaged llama.cpp, sd.cpp and audio.cpp on all three runners, and legacy-update-bridge ran on the tag, which settles four packaging lines and one updates line. desktop-tests.yml already ran the surfsense_local tests on pull requests. The fit.md line is rewritten, not deleted: aaa461f fixed the Windows reading, but Device.reports_live_memory is still read nowhere. Co-authored-by: Cursor <cursoragent@cursor.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Advanced Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
What
Deletes five Known gaps lines that are no longer true, rewrites a sixth that was only half true, and updates the roadmap bullet that asserted the same stale facts.
docs/architecture/packaging.md— four lines deleted:build-sdcpp.ymlhaving never runsurfsense_localunit or integration tests on pull requestsdocs/architecture/updates.md— one line deleted: the--repofix tolegacy-update-bridgenever having run on a tag.docs/architecture/local-models/fit.md— one line rewritten rather than deleted. The Windows half is fixed; theDevice.reports_live_memoryhalf is not, so the line now says only what is still true, plus the missing test.docs/ROADMAP.md— "Release and CI health" said the next release would be the first to carry the llama.cpp runtime. It now describes what v2.0.3 shipped and what CI work is genuinely left.Why
The v2.0.3 release on 26 Sep 2026 closed most of these, and nothing deleted the lines. Run 36250362768 on tag
v2.0.3was green on all eight jobs —sdcpp / Linux x64,sdcpp / macOS arm64,audiocpp / Linux x64,audiocpp / Windows x64, the threebuildjobs andlegacy-update-bridge— which settles the four packaging lines and the updates line between them. The macOS audio.cpp staging is covered becauserelease-local.ymlrunsscripts/audiocpp/stage.mjs --stricton every runner and skips only the artifact download on macOS, where upstream's archive is fetched instead.The CI line was already false before that:
desktop-tests.ymltriggers onpull_requestintomainanddevand runsuv run pytest -m unit,uv run pytest -m integrationandpnpm test.The
fit.mdline was fixed byaaa461fc9, "fix(local): read free memory on Windows" (24 Sep), which gavesystem_memory.available_bytes()awin32branch callingGlobalMemoryStatusEx.Device.reports_live_memoryis still read nowhere —devices.py:40is its only occurrence — so that half stays, andavailable_bytes()having no test is worth recording while the line is open.This matters beyond tidiness:
CONTRIBUTING.mdand all ten READMEs point contributors at the Known gaps lists as a place to find work, so a stale line sends someone to fix something that already works.Fixes #
How to test
96 doc files, 0 problems. Docs only, no code touched.
To confirm the claims rather than take them on trust:
High-level PR Summary
This PR removes outdated documentation about known gaps and issues that were resolved in release v2.0.3. Specifically, it cleans up statements about CI workflows that now run successfully, runtime builds that are now working, and features that have been fixed. The changes update the roadmap to reflect what v2.0.3 actually shipped and rewrites one partially-resolved item about Windows memory reporting to only mention the part that remains unfixed.
⏱️ Estimated Review Time: 5-15 minutes
💡 Review Order Suggestion
docs/ROADMAP.mddocs/architecture/packaging.mddocs/architecture/updates.mddocs/architecture/local-models/fit.md