[brettheap]

Stacked on #1106: its base is #1106's branch. Once #1106 lands, lane 3 brings main into this branch and moves its base to main.

What changed

A follow-up outside v1.2.0 (Brett Heap, 2026-10-21T10:32:04Z). It lands only after the v1.2.0 tag, on a later merge word.

The bug (#1112). Take a one-byte locale such as Latin-1, a plan written in UTF-8 whose record is a clone plan made, and a work directory outside ASCII (Thing plus U+00E9).

• When the clone is gone, neither spelling of the record names anything, so check clones it again at <--work-dir>/source-readonly/Thing. Those are the UTF-8 bytes (0xC3 0xA9) that --work-dir names, so the new clone is at the plan's record.
• _recorded_source then decided whether that clone is the record with at_the_record. That compared the clone's target with the spelling as written alone, the one byte 0xE9.
• So the clone was put on the list of clones to remove and was removed when check ended. Every run cloned again.
• This is adopt-project.py: the clone plan or check makes and removes is claimed and checked by path, so a shared, writable --work-dir lets another local user turn the removal onto a directory of theirs, two runs on one --work-dir remove each other's clone, a skipped removal says nothing, and a source-readonly that is a file ends in a traceback #803's bug under a one-byte locale. adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106 already fixes the same thing in _keep_the_clone_at_the_record (its mutant M4).

The fix (adopt-project.py), following the issue's Direction:

• A new helper, _recorded_paths(recorded). It lists the paths _recorded_path asks, in its order, with ~ expanded: first the spelling _recorded_spelling gives, then the plan's UTF-8 bytes where this locale spells the text otherwise.
• _recorded_path now asks its paths through _recorded_paths. Its behaviour is unchanged: the spelling as written is asked first, and where neither path is there, the path returned is the spelling as written.
• A new helper beside them, _at_the_record(target, path, there, recorded), decides whether a clone _recorded_source makes again is at the record. It compares the target's real path with every path _recorded_path asked:
  • where neither spelling names anything, each path in _recorded_paths;
  • where a spelling is there, that one and each asked before it, which names nothing (paths[:paths.index(path) + 1]).
• _recorded_source calls the helper, and its docstring has a paragraph for adopt-project.py: under a one-byte locale, check compares a clone it made again for a plan's clone record only with the spelling as written, so it removes that clone instead of keeping it and every run clones again #1112.

A decision the Direction leaves open: never a spelling asked after the one there. The Direction says to compare with both spellings. Where a spelling is there, I compare with that one and with any asked before it, but not with one asked after it.

• Why: where the spelling as written IS there but is no longer whole, it is the record _recorded_path reads first.
• Comparing the UTF-8 bytes there too would keep a clone made at them. The next check would then read the broken spelling as written first and be refused adopt-work-dir-occupied. That refusal calls the good clone "no longer whole" (mutant M2 below, measured).
• So that case behaves exactly as it did on adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106's head: each run clones again, and nothing is refused.
• A spelling asked before the one there names nothing, and _recorded_path asks it first next time, so a clone made there is the record (RV1153's N1, below).

Tests (tests/test_adopt_one_byte_locale.py). The three new tests run check twice in a child under the Latin-1 locale the file already builds. A new gh stand-in clones a repository on disk.

• test_a_clone_made_again_at_the_record_is_kept_under_a_one_byte_locale: the first run says the clone is cloned again and keeps it. The second gives plan ok, with no NOTE. gh is asked once in all, and the clone is still there.
• test_a_clone_made_again_beside_the_spelling_as_written_is_not_kept: a file sits at the spelling as written (0xE9). Both runs give plan ok with the "no longer whole" NOTE, gh is asked twice, and nothing is left at the UTF-8 bytes. That is adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106's head's behaviour.
• test_a_clone_made_again_at_the_spelling_asked_first_is_kept: the reverse. A file sits at the UTF-8 bytes where the clone was, and --work-dir names the free spelling as written (0xE9). The first run says the clone is no longer whole, clones it and keeps it. The second gives plan ok, with no NOTE. gh is asked once in all, and the clone is still there.

Fix round

RV1153, lane 3's review of 590ab96e, posted on this PR, found no MUST, one SHOULD and two NITs. This round, ff33ba12, takes all three:

• S1, SonarQube python:S3776: _recorded_source had a cognitive complexity of 16, over the 15 allowed, from the conditional in its comparison. The comparison moved into _at_the_record. Given both versions of the method, SonarQube's snippet analysis reports S3776 at 16 for the one at 590ab96e, and nothing for the one at this head or for the helper. Counted by Sonar's rules, they are 15 and 1.
• N1, taken: the helper compares with every path _recorded_path asked, as above, so the reverse case keeps its clone (measured below, and its own test). M2's case, a file at 0xE9, is unchanged, and M2 is still killed.
• N2: the Sourcery block after the last line is gone, and adopt-project.py: plan written to and cloning <repository> (read only) into name the caller's --out and --work-dir with _shown, as the other lines of the report name a path, so a bidi control, a line feed or a tab in either no longer reaches the terminal as it was typed (#1115) #1154 is in the merge list.
• A wording that N1 brings: say a symlink that leads nowhere sits at the 0xE9 target, while the UTF-8 bytes hold a file. Then adopt-work-dir-occupied now begins with the record's state, "is a clone plan made that is no longer whole (it is not a directory)", and its way out says to delete the target and run again. At 590ab96e the same case got the plain occupied wording. Both refuse, and neither asks gh (measured).

Measurements

Each check ran under a Latin-1 locale built with localedef, where a child reports iso8859-1 ISO-8859-1 0. The plan was written in UTF-8 mode for the clone at <run>/Thing<U+00E9>/source-readonly/Thing, recording cloned: true and repository: testorg/Thing. The clone was then moved away. Both runs used check --work-dir <run>/Thing<U+00E9>, with a stub gh that clones a repository on disk.

Code	Run 1	Run 2
#1106's head 504a1dd6	exit 0, NOTE "no longer exists ... cloned again", gh calls 1, clone removed at the end	exit 0, the same NOTE, gh calls 2, clone removed again
590ab96e	exit 0, the NOTE, gh calls 1, clone kept	exit 0, plan ok, no NOTE, gh calls still 1, clone kept
This head ff33ba12	exit 0, the NOTE, gh calls 1, clone kept	exit 0, plan ok, no NOTE, gh calls still 1, clone kept

The reverse case (N1): the same plan, a file at the UTF-8 bytes where the clone was, and check --work-dir <run>/Thing<0xE9>, the spelling as written, which is free.

Code	Run 1	Run 2
590ab96e	exit 0, NOTE "no longer whole", gh calls 1, clone removed at the end	exit 0, the same NOTE, gh calls 2, clone removed again
This head ff33ba12	exit 0, the NOTE, gh calls 1, clone kept	exit 0, plan ok, no NOTE, gh calls still 1, clone kept

With a file at the spelling as written (M2's case), #1106's head, 590ab96e and this head give the same result: each run clones again, gh calls reach 2, and nothing is left.

Out of scope

#1112 has no "Needs a ruling" or "Documented limit" section. Nothing else was changed.

Tests

Each file ran on its own at ff33ba12, under a stub gh that exits 97, GH_TOKEN=bogus, an empty GH_CONFIG_DIR, GIT_SSH_COMMAND=false, GIT_CONFIG_GLOBAL=/dev/null and EMAIL set, with /usr/bin/python3 -m pytest -q -p no:cacheprovider (Python 3.12.3):

Test file	Result
tests/test_adopt_one_byte_locale.py (-rs)	13 passed in 17.83s, none skipped
tests/test_adopt_unenterable_source_and_locale.py	13 passed in 10.15s
tests/test_adopt_c_locale_bytes.py	17 passed in 39.87s
tests/test_adopt_source_clone_record.py	100 passed in 38.11s
tests/test_scaffold_leg_path_collisions.py	104 passed in 4.84s
tests/test_repo_hygiene.py	100 passed in 2.82s
tests/test_shape_doctor.py	445 passed, 3 skipped in 767.13s (the counts of the run at 590ab96e)
tests/test_path_guard_hygiene.py, at the merge of this head into origin/main 038dafba	49 passed in 11.75s
tests/test_adopt_one_byte_locale.py (-rs), at that merge	13 passed in 17.43s, none skipped
tests/test_adopt_unenterable_source_and_locale.py, at that merge	13 passed in 9.02s

• The guard test: tests/test_path_guard_hygiene.py (tests/test_path_guard_hygiene.py and shape-doctor.py: a bare is_file/is_dir/exists guard in a shipped tool fails by file and line unless its _is_a or a reasoned allowlist covers it, and a project.yaml linked into an unsearchable folder is yaml-unreadable (#430) #1111) is on main and not on adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106's branch. It ran on a local merge commit of origin/main and this head, built with git merge-tree --write-tree and commit-tree and checked out detached. This change adds no is_file, is_dir, exists, is_symlink or os.path guard, only os.path.realpath and os.path.expanduser.
• pyflakes 4.0.0: no warnings on either file. No added line is over 79 characters, and nothing added is outside ASCII.

Mutants

Each mutant was applied to adopt-project.py, and tests/test_adopt_one_byte_locale.py was run. The file was then restored, and cmp found it identical.

Mutant	Result	Killed by
M1: _at_the_record compares with the path _recorded_path gave alone (asked = [path]), as on #1106's head	KILLED, 2 failed	test_a_clone_made_again_at_the_record_is_kept_under_a_one_byte_locale and test_a_clone_made_again_at_the_spelling_asked_first_is_kept (each second run said "cloned again")
M2: every path compared even where one is there (asked = paths)	KILLED, 1 failed	test_a_clone_made_again_beside_the_spelling_as_written_is_not_kept (the second run gave exit 2, REFUSED adopt-work-dir-occupied)
M3: _recorded_path returns the last path, not the spelling as written, where nothing is there	KILLED, 1 failed	test_the_record_is_asked_as_written_then_as_utf8[neither-latin1]
M4 (N1): where a spelling is there, that one alone (asked = [path] if there else paths), as at 590ab96e	KILLED, 1 failed	test_a_clone_made_again_at_the_spelling_asked_first_is_kept (the second run said "cloned again")

#1106's M4 (_keep_the_clone_at_the_record) is unchanged by this PR.

Merges

git merge-tree --write-tree at this head, ff33ba12:

• origin/main at 038dafba: clean.
• Every open PR that touches adopt-project.py or tests/test_adopt_one_byte_locale.py: each is clean. They are adopt-project.py: plan written to and cloning <repository> (read only) into name the caller's --out and --work-dir with _shown, as the other lines of the report name a path, so a bidi control, a line feed or a tab in either no longer reaches the terminal as it was typed (#1115) #1154, adopt-project.py: a --work-dir link into a loop that Python 3.12 reads past a link that does not loop names the loop, Windows' winerror 1921 counts as one, and a path read as the loop's link says so, with tests for r1065's mutants (#1074) #1143, adopt-project.py: a recorded source whose .git this user cannot read is refused source-unreadable, naming that .git, so check and execute no longer call a clone plan made with its .git at mode 000 no longer whole and clone it again (#1099) #1141, adopt-project.py: check and execute ask a plan's org: and id: right after its three repository names, so a bad one beside an unanswered entry, a moved source, a refused name or another repository is that one refusal in both, and a pins: entry that is no pin is said as _read_as says it (#1020) #1125, README.md, adopt-project.py and the handbook: a kept directory entry makefile/ is tested, README says make at the root reads a kept GNUmakefile or makefile instead, and a docstring says the shape's Makefile alone holds CONTRACTS_DIR (#1061) #1123, adopt-project.py: an org/repo --source too long for the disk is refused as a path is, naming no GitHub repository and offering no url, so --source <300 a's>/Thing is no longer offered https://github.com/<300 a's>/Thing, which can never clone (#1003) #1122, adopt-project.py: execute's NEXT prints python3 scripts/bootstrap.py where GNU make at the root reads a kept GNUmakefile or makefile, and its NOTE names that file and what it takes for make bootstrap to run there (#1047) #1121, adopt-project.py: plan's source line names the source's branch as _shown names a path, so a bidi control, a line boundary or a byte that is not UTF-8 in it is spelled and no longer reaches the terminal raw (#1035) #1120, adopt-project.py: check asks the split's own clone of the source in the work directory, so a clone this git will not make is adopt-git-unasked before any leg, and _resolved and _answers make their repositories as check's other throwaways do (#1083) #1119, adopt-project.py, shape_materialize.py and tests: a named bad link left out of the copy and pin findings takes no other path's finding with it, and the docstrings and the materializer's refusal name the #979 exception and what helps it (#1089) #1118 and shape_materialize.py and the pinned AGENTS-shape.md: where the root keeps a GNUmakefile or a makefile, which GNU make reads first, the Makefile collision follow-up no longer says the shape's make targets arrive with the merge, and the pinned file says make reads the first of the three and the block reaches it only once that file holds it (#1048) #1105. adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106's own head is clean with each as well (RV1153 checked it with all eleven).

Closes #1112

Lane: openreposhape-3 (openRepoShape-3)

🤖 Generated with Claude Code

[sourcery-ai]

Reviewer's Guide

Fix repeated cloning under one-byte locales by recognizing either recorded path spelling when neither path exists, while preserving the established precedence and refusal behavior when the spelling-as-written path is occupied; add focused two-run integration tests for both cases.

Sequence diagram for locale-safe clone record detection

sequenceDiagram
    participant Check
    participant Plan_recorded_source
    participant Recorded_paths
    participant Filesystem
    participant Clone

    Check->>Plan_recorded_source: _recorded_source(local_path, work_root)
    Plan_recorded_source->>Recorded_paths: _recorded_paths(local_path)
    Recorded_paths->>Filesystem: Check spelling as written
    Recorded_paths->>Filesystem: Check UTF-8 spelling
    Filesystem-->>Recorded_paths: Neither path exists
    Plan_recorded_source->>Clone: Source.clone(repository, work_root, unwritable)
    Clone-->>Plan_recorded_source: Clone at UTF-8 path
    Plan_recorded_source->>Plan_recorded_source: Compare clone real path with each recorded path
    Plan_recorded_source-->>Check: Keep clone at the record
    Check->>Plan_recorded_source: _recorded_source(local_path, work_root)
    Plan_recorded_source-->>Check: Existing clone is reused

Loading

File-Level Changes

Change	Details	Files
Add a shared path-resolution helper and use both locale-dependent spellings when determining whether a recreated clone is at the recorded location.	Enumerate the spelling-as-written path first, followed by the UTF-8-byte path when distinct, with home-directory expansion. Refactor _recorded_path to preserve its existing precedence and fallback behavior while using the shared enumeration. Update clone-retention logic to compare the recreated target against all candidate paths only when neither spelling currently names an entry; retain the existing single-path behavior when one does. Document the one-byte-locale edge case and the intentional behavior boundary.	adopt-project.py
Add end-to-end Latin-1 regression coverage for retaining recreated clones without changing occupied-record behavior.	Create a local gh stand-in that clones a repository and records invocations. Verify a recreated clone at the UTF-8 path is kept and reused on the second check. Verify a file at the spelling-as-written path causes the recreated clone to be removed and cloning to recur. Reuse locale, repository, and plan helpers to exercise real subprocess behavior.	tests/test_adopt_one_byte_locale.py

Possibly linked issues

• adopt-project.py: under a one-byte locale, check compares a clone it made again for a plan's clone record only with the spelling as written, so it removes that clone instead of keeping it and every run clones again #1112: The PR fixes the exact issue by checking both recorded spellings when neither names an existing path, so the clone is kept.

---

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, 2026-10-22T15:38:24Z) Lane 3's independent review of #1153 (RV1153), head 590ab96. PARTIAL: stopped at a usage checkpoint. No MUST so far.

SHOULD

• S1. New CRITICAL Sonar issue python:S3776 on _recorded_source (adopt-project.py:6567), cognitive complexity 16 > 15, caused by the conditional at :6681. The gate still passes. Fix: move :6677-6681 into a module-level helper _at_the_record(target, path, there, recorded) that computes realpath(target) and compares it with [path] if there else _recorded_paths(recorded); _recorded_source calls it. SonarQube's snippet analysis reports S3776 at 16 for the method as it is, and 0 issues with the helper.

NIT

• N1. The there-guard is stricter than its reason needs (adopt-project.py:6681, docstring :6608-6609). Take the UTF-8 spelling there but broken, and --work-dir at the free 0xE9 spelling: the clone is removed and every run clones again, as on the parent (measured: gh 2 at both). But _recorded_path asks 0xE9 first. Comparing with every path _recorded_path asked (paths[:paths.index(path)+1] if there else paths) keeps it, giving gh 1 and plan ok on run 2, and leaves M2's case unchanged (trial only; the test files were not run against it). Fold it in, or name the case under Out of scope.
• N2. Strip the Sourcery block after the robot line. Add adopt-project.py: plan written to and cloning <repository> (read only) into name the caller's --out and --work-dir with _shown, as the other lines of the report name a path, so a bidi control, a line feed or a tab in either no longer reaches the terminal as it was typed (#1115) #1154 to the merge list.

Checked

• Probe, check run twice: gone under Latin-1 went from gh 2 (parent) to gh 1 with plan ok (head); a symlinked work dir likewise. Gone under C, POSIX and UTF-8 mode, a file at 0xE9, a work dir elsewhere, the reverse case and a collision case (record Thing + U+00C3 U+00A9, work dir bytes C3 A9): the same key lines at parent and head. No new false acceptance.
• Other record comparisons: _keep_the_clone_at_the_record already asks after the clone exists (adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106 M4); the other _same_spelling callers compare runtime paths. Nothing to do.
• Tests at head, one at a time: one_byte_locale 12 passed with -rs, no skips; unenterable 13; leg_path_collisions 104; repo_hygiene 100. pyflakes clean; no line over 79 characters.
• Mutants M1, M2 and M3 all KILLED, each file restored and checked by cmp.
• merge-tree: clean with origin/main 038dafb (tree 6e3eb76b), and the head and parent are each clean with all 11 open adopt-project.py PRs (adopt-project.py: plan written to and cloning <repository> (read only) into name the caller's --out and --work-dir with _shown, as the other lines of the report name a path, so a bidi control, a line feed or a tab in either no longer reaches the terminal as it was typed (#1115) #1154, adopt-project.py: a --work-dir link into a loop that Python 3.12 reads past a link that does not loop names the loop, Windows' winerror 1921 counts as one, and a path read as the loop's link says so, with tests for r1065's mutants (#1074) #1143, adopt-project.py: a recorded source whose .git this user cannot read is refused source-unreadable, naming that .git, so check and execute no longer call a clone plan made with its .git at mode 000 no longer whole and clone it again (#1099) #1141, adopt-project.py: check and execute ask a plan's org: and id: right after its three repository names, so a bad one beside an unanswered entry, a moved source, a refused name or another repository is that one refusal in both, and a pins: entry that is no pin is said as _read_as says it (#1020) #1125, README.md, adopt-project.py and the handbook: a kept directory entry makefile/ is tested, README says make at the root reads a kept GNUmakefile or makefile instead, and a docstring says the shape's Makefile alone holds CONTRACTS_DIR (#1061) #1123, adopt-project.py: an org/repo --source too long for the disk is refused as a path is, naming no GitHub repository and offering no url, so --source <300 a's>/Thing is no longer offered https://github.com/<300 a's>/Thing, which can never clone (#1003) #1122, adopt-project.py: execute's NEXT prints python3 scripts/bootstrap.py where GNU make at the root reads a kept GNUmakefile or makefile, and its NOTE names that file and what it takes for make bootstrap to run there (#1047) #1121, adopt-project.py: plan's source line names the source's branch as _shown names a path, so a bidi control, a line boundary or a byte that is not UTF-8 in it is spelled and no longer reaches the terminal raw (#1035) #1120, adopt-project.py: check asks the split's own clone of the source in the work directory, so a clone this git will not make is adopt-git-unasked before any leg, and _resolved and _answers make their repositories as check's other throwaways do (#1083) #1119, adopt-project.py, shape_materialize.py and tests: a named bad link left out of the copy and pin findings takes no other path's finding with it, and the docstrings and the materializer's refusal name the #979 exception and what helps it (#1089) #1118, shape_materialize.py and the pinned AGENTS-shape.md: where the root keeps a GNUmakefile or a makefile, which GNU make reads first, the Makefile collision follow-up no longer says the shape's make targets arrive with the merge, and the pinned file says make reads the first of the three and the block reaches it only once that file holds it (#1048) #1105).
• SonarCloud: gate OK, duplication 0.0%. CI: all 4 checks pass.
• Body: draft, no milestone, Stacked sentence first, follow-up first under What changed, ends Closes adopt-project.py: under a one-byte locale, check compares a clone it made again for a plan's clone record only with the spelling as written, so it removes that clone instead of keeping it and every run clones again #1112 / Lane / robot, nothing above U+007F but the robot.
• Not yet: test_shape_doctor.py, and test_path_guard_hygiene.py plus this PR's test files on the local merge with main.

Lane 3's plan: S1 and N1 go into a fix round on #1153 in the next usage window: the _at_the_record helper, with N1's wider comparison folded in and a test for the reverse case if the measured behaviour holds there, and M1 and M2 re-pointed. N2: the Sourcery block comes off and #1154 joins the merge list. RV1153 then finishes test_shape_doctor.py and the merged-tree tests. #1153 lands only after #1106, after the v1.2.0 tag.

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

[brettheap]

(lane openRepoShape-3, 2026-10-22T19:46:27Z) Lane 3's independent review of #1153, delta round (RV1153), head ff33ba1. Verdict: READY. No new MUST or SHOULD.

• S1 fixed: the comparison moved into _at_the_record (adopt-project.py:1891); nothing else changes but docstrings. SonarQube snippet analysis of the helper and the new method finds no issue (S3776 was 16 at 590ab96). SonarCloud on adopt-project.py: under a one-byte locale such as Latin-1, a clone check makes again for a plan's clone record that neither spelling names is compared with both, so it is kept and the next run reads it rather than cloning again (#1112) #1153: no open issues, gate OK, duplication 0.0%. CI: all four checks pass on ff33ba1.
• N1 fixed: every probe scenario re-run at ff33ba1. gone (Latin-1, C, POSIX, UTF-8 mode), a file at 0xE9 (M2's case), --work-dir elsewhere, a symlinked --work-dir and the collide case give the same key lines as 590ab96. reverse went from gh 2 with the clone removed to gh 1 with the clone kept and plan ok on run 2.
• No new kind of false acceptance: the only newly kept clone is at the 0xE9 spelling, which _recorded_path asks first. adopt-project.py: under a one-byte locale such as Latin-1, check reads a plan written in UTF-8 whose source is there, asking the plan's UTF-8 bytes where the spelling as written names nothing, rather than refusing plan-source-gone (#1078) #1106's head already keeps a clone there when the UTF-8 clone is gone (measured: gh 1, kept). It is a fresh clone of the plan's own repository, held to commit: by plan-stale.
• The wording change in the dangling-symlink case is true. Both versions exit 2 with no gh call. Following the new way out (link moved aside) cloned and kept on run 1 and gave plan ok on run 2, with no new gh call. NIT, no change needed: "It is no longer a whole clone" is said of a symlink that leads nowhere.
• The new test test_a_clone_made_again_at_the_spelling_asked_first_is_kept passes. Mutants all KILLED, each file restored and checked by cmp: M1 (asked = [path], 2 failed), M2 (asked = paths, the beside test), M3 (paths[-1], neither-latin1), M4 (asked = [path] if there else paths, the new test).
• Tests at head, one at a time: test_adopt_one_byte_locale.py -rs 13 passed, none skipped; test_adopt_unenterable_source_and_locale.py 13 passed. pyflakes clean; no line over 79 characters; nothing added is non-ASCII. test_shape_doctor.py skipped as allowed (FX1153 ran it: 445 passed, 3 skipped).
• Merges: clean with main 961640c (only v1.2.0 release notes since 038dafb), and with all 14 open PRs touching adopt-project.py, including adopt-project.py and repo_shape.py: a name that holds a default-ignorable character, which a terminal draws as nothing, is spelled by its code point in plan's lines and in the unsafe-value refusal, so a branch ma<U+034F>in no longer prints as main (#1128) #1171, adopt-project.py: check names each case twin at the shape's copy, above it and at the manifest, before a link there or after it, and at a pin's file whose exact spelling holds nothing, as _KeptTree keeps every row, not the first (#1138) #1170 and adopt-project.py: check's rehearsal and _stored_as_written take a parent's -c settings, attr.tree too, and the source's objects, GIT_ATTR_SOURCE is refused by name before any question, and each throwaway reads its environment first (#1136) #1169, which opened since.
• Body: draft, no milestone; ends Closes adopt-project.py: under a one-byte locale, check compares a clone it made again for a plan's clone record only with the spelling as written, so it removes that clone instead of keeping it and every run clones again #1112, Lane, robot; no Sourcery block; adopt-project.py: plan written to and cloning <repository> (read only) into name the caller's --out and --work-dir with _shown, as the other lines of the report name a path, so a bidi control, a line feed or a tab in either no longer reaches the terminal as it was typed (#1115) #1154 in the merge list; only closing keyword Closes adopt-project.py: under a one-byte locale, check compares a clone it made again for a plan's clone record only with the spelling as written, so it removes that clone instead of keeping it and every run clones again #1112; nothing above U+007F but the robot.

Lane 3's plan: #1153 is READY. It lands only after #1106, which lands after the v1.2.0 tag on a later word. Then lane 3 brings main into this branch, moves its base to main, and lands it by the stacked-PR merges.

Lane: openreposhape-3 (openRepoShape-3)