[brettheap]

Lane: openreposhape-3 (openRepoShape-3)
Claim: #755 (comment)

Closes #755.

Lands after the v1.1.0 tag. (#751, the issue's "Lands after", is in main as e3137010.)

Tests only. No tool changes.

Re-measured on main e3137010

The issue's claims still hold. refusals_ahead has moved since b7312258: it now starts at shape-doctor.py:2813, its prepare refusal returns at :2843 and its conflict list is at :2845. Both of RV-f's mutants, applied on this main, pass every test the file held before:

• M1, :2843 return None → return {}.
• M2, :2845 if row.is_conflict → if row.state == us.COPY_IS_LINK.

What changed

All in tests/test_update_shape_merged_away.py:

• test_the_pins_row_keeps_the_gate_where_check_refuses_the_standard (:509). Both copies are merged away, and the doctor runs from a standard of another history (test_shape_doctor.standard_of_another_history), which lacks the pinned commit.

  • check refuses it, update-revision-unresolvable, as the test asserts first.
  • So refusals_ahead cannot ask its question, and the pins row must keep its gate: no merged_away key, and a NEXT that is not update-shape.py.
  • This is the issue's first option. A promisor = false-alone clone was the other; a standard short of the pinned commit needs no partial clone.

• test_the_pins_row_keeps_the_gate_beside_each_conflict_check_names[both|upstream-removed|unmapped] (:592). shape/Makefile is merged away, and beside it one of these:

  • both (both_beside, :550): the project's kept edit of AGENTS-shape.md, pinned as it stands, which is what apply --accept-local leaves, beside the standard's own edit of that copy one commit on.
  • upstream-removed (upstream_removed_beside, :562): the standard one commit on, without the AGENTS-shape.md template.
  • unmapped (unmapped_beside, :574): a pin row for a file of the project's own, as the issue suggests.

  Each case asserts four things in turn:

  1. check itself calls the path that state.
  2. The pins row keeps its gate, with refusal_ahead == {path: state} and the REFUSAL AHEAD clause naming it (the existing assert_the_pins_row_keeps_its_gate).
  3. The apply that check prints, pasted, exits 2.
  4. That exit is REFUSED update-conflict, as RV-f measured.

• Shared helpers:

  • pinned_as_it_stands (:531) rewrites, or adds, one pin row from the bytes on disk, matching shape_materialize.py's row form quoted or not, and commits.
  • assert_the_pins_row_is_the_gate (:502) is now used by the new test and by …_where_check_would_fetch (:488), whose two asserts it replaces.

• The builders are one table, CONFLICTS_BESIDE (:584), so the three cases share one test body.

Each named mutant, killed

Each mutant was applied to 837920c2 in a scratch clone, and the whole file run against it:

• M1 (return {}): 1 failed, 17 passed. The one failure is test_the_pins_row_keeps_the_gate_where_check_refuses_the_standard.
• M2 (copy-is-link alone): 3 failed, 15 passed. The failures are …_beside_each_conflict_check_names[both], [upstream-removed] and [unmapped].

Tested

With a stub gh first on PATH, GH_TOKEN=bogus and an empty GH_CONFIG_DIR:

/usr/bin/python3 -m pytest -q -p no:cacheprovider --basetemp=<scratch>/bt-w755-T -o tmp_path_retention_policy=failed tests/test_update_shape_merged_away.py tests/test_doctor_partial_clone.py tests/test_repo_hygiene.py tests/test_cli_reference.py

Result: 138 passed in 84.08s (0:01:24). The file needs git filter-repo, which this machine has, so nothing in it was skipped.

Pinned copies

None.

Docs

None. The tests pin behaviour the docs do not state.

🤖 Generated with Claude Code

Summary by Sourcery

Strengthen merged-away update tests so refusal-ahead classification preserves the pins gate across unresolved and conflicting standard states.

Enhancements:

• Expand merged-away shape update coverage to ensure the pins row remains gated when the standard is unresolvable and when it is adjacent to both, upstream-removed, or unmapped conflicts.

Tests:

• Add regression tests covering refusal-ahead handling and verifying that generated conflict updates remain refused.

[sourcery-ai]

Reviewer's Guide

Tests add focused regression coverage for issue #755: the pins row must remain a gated doctor step when standard resolution is unavailable and when each relevant conflict state is adjacent to a merged-away row, preventing incorrect refusals_ahead implementations from passing.

File-Level Changes

Change	Details	Files
Add regression coverage ensuring the pins row retains its gate when standard resolution is refused or when specific conflict states appear beside a merged-away row.	Cover a standard from another history, verifying check refuses before classification while the doctor still preserves the gate. Parameterize both, upstream-removed, and unmapped neighboring conflict scenarios and verify their refusal-ahead state, generated apply command, and update-conflict refusal. Add helpers to rewrite or add pin entries from current file contents and centralize gate assertions and scenario builders. Import hashing, regex, and standard-history helpers needed to construct and validate the fixtures.	tests/test_update_shape_merged_away.py

Assessment against linked issues

Issue	Objective	Addressed	Explanation
#755	Test that when check refuses the standard before classifying rows, the pins row keeps its gate and has no merged_away detail.	✅
#755	Test both, upstream-removed, and unmapped rows beside a merged-away row, asserting that the pins row keeps its gate and names the conflict state, and that the printed apply command exits 2 with update-conflict.	✅

Possibly linked issues

• tests: the pins row's merged-away check (#676, PR #751) has no test for a prepare that refuses, or for both, upstream-removed and unmapped beside a merged-away row; two mutants survive #755: The PR tests a refusing check and merged-away rows beside both, upstream-removed, and unmapped states, killing both named mutants.

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[brettheap]

(lane openRepoShape-3, 2026-10-17T01:09:28Z) Independent review of #770 by RV-f (opus), on head 837920c2: READY. No MUST and no SHOULD; three optional NITs. It lands only after the v1.1.0 tag, once its Linux tests is green.

The two mutants #751 left alive are now killed (applied to shape-doctor.py at e3137010, whole file run):

mutant	main's 14 tests	the head's 18
M1: :2843 return None → return {} (a prepare refusal reads as nothing ahead)	14 passed	1 failed: test_the_pins_row_keeps_the_gate_where_check_refuses_the_standard
M2: :2845 row.is_conflict → state == COPY_IS_LINK	14 passed	3 failed: [both], [upstream-removed], [unmapped]

RV-f ran four more mutants:

• dropping both, upstream-removed or unmapped from the conflict states each fails only its own case;
• return ahead or None fails four tests already on main.

Each case builds the state it names. RV-f copied each root out of tmp_path and ran update-shape.py check on it:

• prepare refuses → exit 2, update-revision-unresolvable;
• both → exit 1, REFUSAL AHEAD on AGENTS-shape.md;
• upstream-removed → exit 1, REFUSAL AHEAD on AGENTS-shape.md;
• unmapped → exit 1, REFUSAL AHEAD on notes-of-our-own.md;
• the control, merged-away alone → no REFUSAL AHEAD.

In every root, the root's own validate-pins.py reports only the merged-away finding.

Hygiene:

• no composite asserts;
• SonarCloud measures 0.0% new duplication;
• bytes only, so it holds on Windows;
• the file's time went from 36.30s to 43.99s.

NITs (optional):

• N1 :606-609: assert argv[1:3] == ["update-shape.py", "apply"] before running the pasted NEXT, as the first test in the file does.
• N2 :441 (assert_the_pins_row_keeps_its_gate): M2's kill is a KeyError. Use pins_row["detail"].get("refusal_ahead") == ahead, pins_row so a failure shows the row.
• N3 :531: pinned_as_it_stands overlaps tests/test_update_shape.py:944 set_pin_row_sha256. That helper uses write_text, which writes CRLF on Windows. Optionally have it call the new, byte-level helper.

Tests:

• the writer's set: 138 passed;
• the file on the head: 18 passed in 43.99s;
• the file on main: 14 passed.

Merges. All clean, each run with the file passing (18 passed, or 42 with test_doctor_partial_clone.py on the #754 merge):

• e3137010;
• validate-pins.py: the <repository>: N workflow @<sha> reference(s) agree with the gitlink note counts only the references at the gitlink, says N of M where some are not, and is not printed where none is, so no reference is both a pin-workflow-ref-mismatch and one that agrees (#739) #769 787d7e1e;
• shape-doctor.py: in a root that lost scripts/repo_shape.py, the legs and members lockstep NEXT is the pins and family rows' git -C <root> checkout HEAD -- scripts/repo_shape.py (none where HEAD no longer holds it), where it named the root's own gate, which dies on the import (#750) #771 c90a137c;
• siblings.py: a file, or a folder make siblings cannot list or enter, at ../<Project> is NOT A CLONE, reported and left as it is, so the run no longer ends in a traceback at iterdir before the members after it are placed, and make park no longer ends in one over a folder it cannot enter (#735) #772 1d149933;
• shape-doctor.py: the partial-clone row's NEXT names the first promisor remote, so it loops where the pinned commit sits only on another remote or another branch, and no test covers a partial clone that lacks the pinned commit itself #754's eb0568ac (shape-doctor.py: the partial-clone row's NEXT fetches each commit it lacks objects at, by id, from the promisor remote whose branches hold it, so it no longer loops where the pinned commit sits only on another remote or another branch, and its why says where the clone does not hold the commit at all (#754) #773).

Lane: openreposhape-3 (openRepoShape-3)

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[sourcery-ai]

Sorry @brettheap, you've used your own review budget of 250,000 diff characters for the last 7 days.

You can request another review in 1 day and 13 hours by commenting @sourcery-ai review. Upgrade to get a review now.