Skip to content

Release version 2.3.9 with schema updates and documentation fixes - #1054

Merged
tirth8205 merged 2 commits into
testingfrom
staging
Sep 18, 2026
Merged

tirth8205 merged 2 commits into
testingfrom
staging

Conversation

@tirth8205

Copy link
Copy Markdown
Owner

Pull Request

Linked issue

Closes #

What & why

How it was tested

uv run pytest tests/ --tb=short -q
uv run ruff check code_review_graph/
uv run mypy code_review_graph/ --ignore-missing-imports --no-strict-optional

Checklist

  • Tests added for new functionality
  • All tests pass: uv run pytest tests/ --tb=short -q
  • Linting passes: uv run ruff check code_review_graph/
  • Type checking passes: uv run mypy code_review_graph/ --ignore-missing-imports --no-strict-optional
  • Lines are at most 100 characters
  • Docs updated where behavior changed (README, docs/, docstrings)

tirth8205 and others added 2 commits September 18, 2026 19:24
Version strings, the CHANGELOG section, the Action pin in the five READMEs
and the shipped docs.

Review fixes folded in:

- docs/schema.md, docs/TROUBLESHOOTING.md and CLAUDE.md still said the
  current schema version was 10 while this release ships 13. docs/ is in the
  sdist, so the stale number would have gone to PyPI. CI's schema-sync job
  only compares migrations.py against the VS Code extension, so a new test
  pins the places it does not reach.
- The CHANGELOG said "This is schema version 10" — true of one migration,
  false of the release, which moves a database 9 -> 13. Migrations v11, v12
  and v13 were undocumented, including that v13 rewrites every node row and
  rebuilds the FTS5 index on first open.
- A new "Upgrade notes" section: the schema move, the VS Code extension
  needing a separate republish, the rebuild the parse-time fixes need
  (update skips unchanged files and nothing stamps the parser version), the
  reinstall the hook needs, and the Action cache key moving to schema13.
- The import benchmark's baseline was rewritten to "Measured against 2.3.8"
  without re-measuring. The figures were taken against a parent that already
  carried the batched writes this release also ships, so they isolate the
  import change and are not a 2.3.8 comparison.
- CRG_DISCOVERY_TIMEOUT was documented as never exceeding CRG_GIT_TIMEOUT.
  An explicit value is used as given; only the 5-second default is capped.
- The #821 entry claimed the generated instruction blocks changed. They are
  byte-identical to v2.3.8; hints and the prompt templates are what changed.
- Three user-visible changes had no entry at all: the FTS search rewrite,
  the seeded visualize neighbourhood, and the test-gap reclassification.
  The two CRG_CALLER_TEST_ROUTE_* variables are now in all five READMEs.
- uninstall could not remove the Qoder skills install now writes. #909 made
  install write the bundled workflows and stop copying the target project's
  skills/ directory, but uninstall still derived the list of .qoder skills
  from that directory, which a normal repository does not have, so it
  removed none of the seven. Both sides now read skills.bundled_skill_names;
  the pre-#909 shape is still removed as well.
- The VS Code extension carried SUPPORTED_SCHEMA_VERSION 13 while still
  calling itself 0.2.2, the same version as the build that accepts 9. Bumped
  to 0.3.0 with a changelog entry.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown

code-review-graph review

Overall risk: 0.65 (MEDIUM) — 8 changed function(s)/class(es), 2 affected flow(s), 4 test gap(s)

Risk-scored changes

Risk Level Symbol Location Tested
0.65 medium code_review_graph/skills.py::bundled_skills_dir code_review_graph/skills.py:2776 no
0.55 medium code_review_graph/skills.py::bundled_skill_names code_review_graph/skills.py:2803 no
0.40 medium code_review_graph/skills.py::install_qoder_skills code_review_graph/skills.py:2825 yes
0.35 low code_review_graph/uninstall.py::_qoder_skill_names code_review_graph/uninstall.py:951 no
0.35 low code_review_graph/uninstall.py::_process_repo code_review_graph/uninstall.py:1006 no
0.30 low tests/test_documentation.py::test_schema_version_is_current_everywhere_it_is_written_down tests/test_documentation.py:80 (test)
0.05 low tests/test_uninstall.py::test_shared_skill_directories_keep_user_files_and_unrelated_skills tests/test_uninstall.py:375 (test)
0.05 low tests/test_uninstall.py::test_qoder_skills_are_removed_without_a_repository_skills_directory tests/test_uninstall.py:419 (test)

Affected execution flows

  • bundled_skill_names — criticality 0.36, 2 node(s) across 1 file(s)
  • install_hermes_skills — criticality 0.26, 3 node(s) across 1 file(s)

Test gaps

  • code_review_graph/skills.py::bundled_skills_dir (code_review_graph/skills.py:2776)
  • code_review_graph/skills.py::bundled_skill_names (code_review_graph/skills.py:2803)
  • code_review_graph/uninstall.py::_qoder_skill_names (code_review_graph/uninstall.py:951)
  • code_review_graph/uninstall.py::_process_repo (code_review_graph/uninstall.py:1006)

Token savings: this graph-backed report used ~440,923 fewer tokens (~99%) than reading every changed file in full (estimated, chars/4 approximation).


Powered by code-review-graph — local-first analysis; no code leaves the CI runner.

@tirth8205
tirth8205 merged commit 57b5999 into testing Sep 18, 2026
40 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant