[brettheap]

Lane: openreposhape-1 (openRepoShape-1)
Claim: #938 (comment)

A v1.2.0 draft on main 01a2b016, which holds #881. It is #881's follow-up, #938, from lane 5's re-check of #881 at 1d21a850. It changes test docstrings and comments only, in 8 files under tests/: ast.dump of every changed file, outside its docstrings, is unchanged. Line numbers below are at the head, fc4c06a3.

What

1. tests/test_adopt_split_branch.py:21-23 no longer says its gh "is never reached".

   • insteadOf redirects git, not gh.
   • Given no --local-remote-dir, once nothing before it refuses and the plan's names and org: pass, check asks gh repo view --json nameWithOwner of testorg/Northwind-spec and of testorg/Northwind-code (adopt-project.py execute: each leg repository is asked about only as it is made, so a code leg whose name is taken leaves the spec leg created, on GitHub as under --local-remote-dir, and a --local-remote-dir that is a file or cannot be written ends in a traceback after check said plan ok #798). Off Windows, the session's guard answers each as a gh with no login does.
   • The file's in-process execute runs replace run with no_gh, so none of execute's gh reaches the guard.

2. "Every machine" is said off Windows only.

   • The places: tests/gh_guard.py:28-30 (the machine row: "off Windows a logged-in operator gets the output CI gets"), tests/gh_guard.py:49 (the NOTE), and tests/test_template_text_placeholders.py:32 ("NO NETWORK AND NO GITHUB, off Windows").
   • On Windows a bare gh skips the guard's gh.cmd (tests/conftest.py). Where a logged-in gh.exe is installed, it answers instead.

3. The condition names what ends check first: "once nothing before it ends check and the plan's names and org: pass". That covers a mistyped plan, which ends check with exit 1 and findings, and a refusal, which ends it with exit 2. It is in seven places:

   • tests/gh_guard.py:45-46;
   • test_adopt_check_plan_name.py:38;
   • test_adopt_project_is_the_source.py:26;
   • test_adopt_tracking_branch.py:24;
   • test_adopt_origin_repository.py:33;
   • test_template_text_placeholders.py:36-37;
   • the new sentence in test_adopt_split_branch.py.

4. The guard's own summary names adopt-project.py execute: each leg repository is asked about only as it is made, so a code leg whose name is taken leaves the spec leg created, on GitHub as under --local-remote-dir, and a --local-remote-dir that is a file or cannot be written ends in a traceback after check said plan ok #798's answer: tests/gh_guard.py:2 ("three answers, a refusal, every call logged"), :17-22, and tests/conftest.py:13-17, which now name gh repo view <owner>/<name> --json nameWithOwner, answered as a gh with no login answers it.

5. tests/test_adopt_split_branch.py:732: check "reads no network for the branches".

Measured, with the guard's log

• test_adopt_split_branch.py: 86 calls in 31 of its 47 tests (20 with 2 calls, 10 with 4, and 1 with 6), as lane 5 counted. Every call is repo view testorg/Northwind-spec|-code --json nameWithOwner.
• The 8 test functions (11 cases) that use no_gh and run no check: no calls.
• test_check_of_a_plan_whose_branch_was_deleted_is_refused_by_name (test_adopt_tracking_branch.py:390): no calls, and it exits 2.
• The other files: their counts match tests: the docstrings of five adopt test files and tests/gh_guard.py say that check asks gh repo view once of each leg once the plan's names and org: pass, which the session's guard answers, and what execute's in-process runs replace (#880) #881's table: 62, 28, 18, 4 and 20.

Tests

• Review: r938 (sonnet), not the writer, found it READY at 0b292a53 with one SHOULD, the exact condition; round 2 at fc4c06a3 makes it and four NITs, and r938 found it READY there.
• The nine files tests: the docstrings of five adopt test files and tests/gh_guard.py say that check asks gh repo view once of each leg once the plan's names and org: pass, which the session's guard answers, and what execute's in-process runs replace (#880) #881's gate ran: 450 passed and 2 skipped. Both skips are test_no_auto_maintenance.py:222, which skips when the session sets GIT_CONFIG_GLOBAL, as the probe environment does.
• After the :28 round: test_gh_guard.py, test_template_text_placeholders.py and test_repo_hygiene.py, 231 passed.
• Style: every added line is ASCII and at most 79 columns, and git diff --check is clean.

Known limits

• Windows was not run. The Windows sentences rest on the comment in tests/conftest.py.
• tests/test_gh_guard.py:203 keeps "on every machine", because that test is marked WINDOWS_SKIP.

Closes #938

🤖 Generated with Claude Code

Summary by Sourcery

Clarify the test suite's GitHub CLI guard behavior and adoption-check documentation, including platform-specific limitations and the conditions that trigger repository lookups.

Enhancements:

• Clarify test documentation to accurately describe GitHub CLI calls made by adoption checks and the conditions under which they occur.
• Document the guard's handling of repository metadata queries and distinguish its behavior from real GitHub CLI execution on Windows.

[sourcery-ai]

Reviewer's Guide

Documentation-only updates align test docstrings and comments with observed gh execution: adoption check can invoke guarded repository lookups, while Windows may bypass the guard, and execute uses its own stubbed command path.

File-Level Changes

Change	Details	Files
Correct test documentation to describe the actual gh reachability and guard behavior for adoption checks.	Explain that Git URL rewriting does not redirect gh, and that check reaches gh repo view after earlier validations pass. Document the guard’s off-Windows scope and the possibility of a real logged-in gh.exe on Windows. Clarify that in-process execute calls use no_gh and therefore do not reach the session guard. Qualify repeated test-module descriptions with the prerequisite refusal and plan-name/org validation conditions.	tests/gh_guard.py tests/test_adopt_split_branch.py tests/test_adopt_check_plan_name.py tests/test_adopt_project_is_the_source.py tests/test_adopt_tracking_branch.py tests/test_adopt_origin_repository.py tests/test_template_text_placeholders.py

Assessment against linked issues

Issue	Objective	Addressed	Explanation
#938	Correct the split-branch test docstring to explain that git's insteadOf redirect does not prevent check from running gh, and clarify that the session guard answers those calls off Windows while the in-process execute calls are stubbed.	✅
#938	Qualify claims that the guard answers gh or that GitHub is not reached so they apply off Windows, and describe the Windows behavior where a real gh.exe may answer instead.	✅
#938	Update the relevant docstrings and guard documentation to state that check asks gh only when nothing earlier refuses and the plan's names and org: pass.	✅

Possibly linked issues

• tests: test_adopt_split_branch.py still says check's gh is never reached, gh_guard.py and test_template_text_placeholders.py still say "every machine" beside "off Windows", and the docstrings' condition leaves out a refusal that comes first (follow-up to #881) #938: The PR makes the requested corrections to gh reachability, Windows qualifiers, and check’s earlier-refusal condition.
• tests: the docstrings of five adopt test files and tests/gh_guard.py say that check asks gh repo view once of each leg once the plan's names and org: pass, which the session's guard answers, and what execute's in-process runs replace (#880) #881: The PR refines the same docstrings’ account of guarded gh calls, adding when calls occur and clarifying Windows behavior.

---

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-1, 2026-10-18T18:47:18Z) Independent review of #966 by lane 1's reviewer r938 (sonnet), who is not its writer: READY at fc4c06a3, with no MUST, SHOULD or NIT left.

• Round 1, at 0b292a53: READY with one SHOULD.

  • "Once nothing before it refuses" was not exact. A plan whose keys have the wrong type ends check with exit 1 and findings, before any name is read, and never reaches gh (measured: plan-bad-legs and plan-bad-paths, exit 1).
  • Every other sentence was true against cmd_check, the guard and tests/conftest.py.
  • ast.dump, outside docstrings, was unchanged in all 7 files.
  • 450 passed and 2 skipped, and the guard log matches tests: the docstrings of five adopt test files and tests/gh_guard.py say that check asks gh repo view once of each leg once the plan's names and org: pass, which the session's guard answers, and what execute's in-process runs replace (#880) #881's table.

• Round 2, at fc4c06a3:

  • "once nothing before it ends check", in all 7 places;
  • "off Windows" in the machine row;
  • check "reads no network for the branches";
  • the guard's own summary (gh_guard.py:2, :17-22, conftest.py:13-17) names adopt-project.py execute: each leg repository is asked about only as it is made, so a code leg whose name is taken leaves the spec leg created, on GitHub as under --local-remote-dir, and a --local-remote-dir that is a file or cannot be written ends in a traceback after check said plan ok #798's repo view answer.

  It changes docstrings only, in 8 files, and ast.dump is identical. test_gh_guard.py and test_repo_hygiene.py: 89 passed. The writer's nine files: 450 passed, 2 skipped.

• The body is updated to the head.

#966 takes its turn once its tests are green at fc4c06a3, gated over main as it then stands.

Lane: openreposhape-1 (openRepoShape-1)

[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 23 hours and 26 minutes by commenting @sourcery-ai review. Upgrade to get a review now.