Repository navigation
foundations: fix the small findings of #3363 - #3365
Conversation
check_dependencies.py now reads a closed crate list from dependency_rules.json: the five foundation crates take only heapless, libm and each other as production dependencies. The foundations design doc cites the check and records why the sound cue crosses the port. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011wmjvKzamiTf3Th8ePmUoi
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011wmjvKzamiTf3Th8ePmUoi
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011wmjvKzamiTf3Th8ePmUoi
…ments The module holds the Navigator checkpoint codec. The waypoint category comment now says what readers do: every OBCM PoiCategory id reads as that category. DateTime::to_unix states where its u32 wraps. Version-history wording is gone from obcr.rs and cache.rs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011wmjvKzamiTf3Th8ePmUoi
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011wmjvKzamiTf3Th8ePmUoi
|
Caution Review failedThe pull request is closed.
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. |
|
The Generated by Claude Code |
|
The Generated by Claude Code |
|
The re-run of Generated by Claude Code |
Mode: fix
Requirements: none
Public docs: yes (the architecture guide, in its own
docs:commit)This pull request fixes the findings with the action fix in #3363. It changes no behavior and no bytes.
What changes
check_closed_dependenciesinfirmware/tools/check_dependencies.pyrefuses a production dependency of the five foundation crates other thanheapless,libmand each other. Its list isclosed_dependenciesindependency_rules.json. A name in the list that is not in the workspace also fails. CI runs it in the existing dependency-direction step. The Rules table ofdocs/design/foundations.mdcites it.ContentRefcodec. The landmarks codec ispub(crate). The article and peak records use it in place of their own copies.Crc32::checksumis gone. Its 3 callers, all inobc-storage, usecrc32.obc_formats::assistantis nowobc_formats::navigator, because it holds the Navigator checkpoint codec. The code-authority path inspecs/ride-archive-metadata.mdfollows. No byte changes.obcr.rsandobc-route/src/reader.rsnow say what the readers do: everyPoiCategoryid reads as its category. Theto_unixdoc says where theu32wraps. Two comments no longer state version history.Cueis inobc-ports.Checks that ran
cargo testforobc-crc(5),obc-formats(51),obc-ports,obc-storage(193) andobc-app: all passed.cargo clippy --all-targets -- -D warningsfor the same crates;clippy --lib -D warningsforobc-routeandobc-host-core.cargo fmt --all, andcargo fmtinfirmware/obc-fw-nrf54l.cairn check: 56 violations, all in the baseline, 0 new.check_dependencies.pyand its unit tests;obc suites check;prose.py --check;build_docs.py --check-links.Checks that did not run
obc-routeandobc-host-core, and--all-targetsclippy for them: GEOS is not installed. A grep finds no remainingobc_formats::assistant.usepaths. CI'sembeddedjob builds the board.Notes for review
dependency_rules.jsonalready has afoundationgroup that is not the same set: it hasobc-elevationandobc-link, and notobc-file-source. That group sets the direction of dependencies between layers; the new list is the cairn system. I kept them apart in this pull request.specs/obcr-route.md:204says that a waypoint category outside1..=6renders as generic. The readers accept everyPoiCategoryid. The spec is unchanged here; the owner decides which side is right.Findings: #3363
Generated by Claude Code