[brettheap]

Lane: openreposhape-5 (openRepoShape-5)
Claim: #835 (comment)

Stacked on #818 (fix/758-git-command-scope, 4a553e17); retarget to main once it lands. The fix is one commit, b9ad1187, written on #818's eaa72d9e. The head fafba9ec merges #818's refresh 4a553e17 (main fc867207) into it, cleanly.

Tests only. Nothing a user runs changes, and no pinned file changes. No README, handbook, AGENTS.md or docs/cli.md sentence is touched: none mentions these helpers (render-cli-reference.py --check passes).

Fix round: rv848 and lane 2's rv818 on fafba9ec, one commit 79b2eac4

Both reviews say READY with no MUST: rv848 (issuecomment-6240157020) and lane 2's second review of #848 (issuecomment-6240228302, "rv818" below). One commit on fafba9ec, by explicit pathspec, fast-forward pushed. Tests only; no pinned file changes; main and #818 are not merged in. The rv818 numbers are next to rv848's.

What each finding got

rv848	rv818	taken	what changed
1 SHOULD: #811's and #838's counts		the text	the merge section below says who converts it
2 NIT: stale merge checks		yes	re-run at the head, lines corrected below
3 NIT: the walk counts arguments	6 NIT: it does not read their values	yes, one change	paste(row, cwd, None) now fails the walk
4 NIT: on_disk is asked for "true" only		yes	ADDERS holds on_disk-false
5 NIT: the upload-pack docstring		yes	paste_env says what the rewrite adds: pack-objects and index-pack
	1 SHOULD: two command scopes built at collection	yes	#818's second review's SHOULD (#758's line 1), taken here so that #818 needs no third round
	2 NIT: an expected dict is flagged	yes	the scan leaves a dict that is an operand of a comparison
	3 NIT: what the scan misses	the text	"Known limits" below, no code
	4 NIT: the same lines in two files	yes	three helpers in conftest.py
	5 NIT: two surviving mutants	in part	the count_of(env) case is added; the remotes_of one is left, below

rv818 SHOULD 1. test_adopt_reencoded_files.py's attr.tree param and test_family_init_asks_every_value.py's user.name param were built at collection, before the session fixture put gc.auto=0 and maintenance.auto=false in the environment. Laid over the environment at the test, each dropped the two from git's command scope. Each is a lambda its test calls, as test_adopt_nested_attributes.py already did.

New tests/test_command_scopes_built_at_the_test.py (2 tests) reads each param off its test's own marks, builds it, and asks git for the command scope (git config --show-origin --list, read and never written). It asks for the session's two settings beside the param's entry. It fails on fafba9ec (the two test modules as they were, with the new conftest.py): 2 failed, git read ['user.name=', 'user.email=', 'attr.tree=4b825dc6...'] and ['user.name=', 'user.email=', 'user.name=Ada Lovelace'], with no gc.auto=0. It passes on the head.

rv848 3 and rv818 6. pastes_in now returns, for each paste call, the fixture it hands the remotes of: remotes_of(<name>) as the third argument or as remotes=, where the name is holder or scaffolded_here and a parameter of the nearest function the call is in. None, another call, an expression on the call, a name the test does not take, or no third argument gives no fixture. The list of calls with none is pastes_without_remotes, a function of its own, so that a test of it can fail. 15 new cases hold the walk: twelve for what a call may hand and three for the function it is in.

rv818 2. counts_set_to_a_number skips a Dict that is an operand of a comparison (assert added == {"GIT_CONFIG_COUNT": str(held + 1), ...}, !=, a dict on the left). I took the narrower reading of "inside an assert or a comparison": a dict in a call inside an assert (assert run(env={"GIT_CONFIG_COUNT": "1"}).returncode == 0), or one assigned before the comparison, still sets a count and is still read. Four cases are still found, three are left, and the count_of(env) case of rv818 5 is left.

rv818 4. JOB_ENTRIES, put_job_and_session_entries(monkeypatch), command_line_entries(listing) and command_scope_at(environ, cwd) in conftest.py replace the delenv, held = ... and loop lines of test_no_auto_maintenance.py and test_doctor_paste_through_git.py, and the four-line read of the command line: entries in three files. OWN_COUNTS follows: conftest.py holds two (the sum, and the job's entries counted from nothing), test_no_auto_maintenance.py one, and test_doctor_paste_through_git.py none. A new test checks that both readers put GIT_CONFIG_PARAMETERS aside (rv818's own round-2 NIT 8).

rv818 5, left. remotes_of returning fixture["base"] instead of fixture["base"] / "remotes" still survives, and I name it equivalent in practice: the prefix <base>/ covers every remote under <base>/remotes, so all the pasted clones are rewritten either way.

Merge checks (replace the body's section)

Re-run at the head 79b2eac4 (git merge-tree --write-tree), against 152 heads: every fix/* and l3-* branch, and the pull heads from #700. Main fc867207 is an ancestor of the head (it came in with #818's refresh 4a553e17).

• Main. fc867207: clean. Main is 190133d5 now (two commits on family.py, the README and the handbook): clean too.
• tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's head 4a553e17: clean. The body's "one conflict, tests/test_non_utf8_git_reads.py" was true only for b9ad1187 and eaa72d9e, before the refresh.
• The open heads that edit the files here merge clean: setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 (050abea6), validate-family.py: the wrong-case move back mends .gitmodules, adds --sparse and moves an entry in the way aside, a row's staged removal and a url-less section are findings, and two refusals name a cure that works (#762) #820, family-root README.md and AGENTS-shape.md, and AGENTS.md: the holder's docs say its parent folder is named EXACTLY after its repository, case included, on every disk, and the never to hand-edit the members: block names the row repair validate-family.py asks for, whose commit is the person's #822, family.py add: the move back of a member's gitlink points its section of .gitmodules back, sets aside first what the index holds in its way and puts back a member whose removal is staged, its closing mv is guarded and leaves a checkout to delete by hand, and --sparse follows the sparse rules, so no printed line leaves a section naming another spelling or un-registers a member (#761) #828 (ce0f2a5f), shape-doctor.py: a .gitmodules section name keeps git's own bytes and is handed back to git as them, so configured_url and kept_gitdir find a member named caf\xe9, its detached commit gets its keep line under any locale, and a non-ASCII name no longer ends --root in a UnicodeEncodeError traceback under LC_ALL=C PYTHONUTF8=0 (#732) #774 (012a7726) and shape-doctor.py and siblings.py: a .gitmodules section git ignores ([submodule ""], a .. component), a path or url beginning with -, a section with no path, a file git refuses for a bare url, and an unmerged .gitmodules name no mount url, NEXT line or clone, where git says No url found and exits 128 (#733) #775 (8c5ad0e8). family.py add: after #646's move back of a member's gitlink, a .gitmodules that a git mv made name the variant is not mended, a file or link in the way has no remedy that keeps the member, and the plain mv it prints orphans a live submodule checkout #761, validate-family.py: the case-variant move back leaves .gitmodules naming the variant and has no --sparse, a staged un-registration passes the gate, and three findings print no remedy, one that dies, or a claim a recursive clone does not keep #762 and templates/family-root README.md and AGENTS-shape.md, AGENTS.md and validate-family.py: the holder's own docs say its parent folder is "named after the family" and omit "EXACTLY, case included", and "never hand-edit the members: block" is contradicted by the validator's repair remedies #763 are the same heads as family.py add: the move back of a member's gitlink points its section of .gitmodules back, sets aside first what the index holds in its way and puts back a member whose removal is staged, its closing mv is guarded and leaves a checkout to delete by hand, and --sparse follows the sparse rules, so no printed line leaves a section naming another spelling or un-registers a member (#761) #828, validate-family.py: the wrong-case move back mends .gitmodules, adds --sparse and moves an entry in the way aside, a row's staged removal and a url-less section are findings, and two refusals name a cure that works (#762) #820 and family-root README.md and AGENTS-shape.md, and AGENTS.md: the holder's docs say its parent folder is named EXACTLY after its repository, case included, on every disk, and the never to hand-edit the members: block names the row repair validate-family.py asks for, whose commit is the person's #822. The body said shape-doctor.py: a .gitmodules section name keeps git's own bytes and is handed back to git as them, so configured_url and kept_gitdir find a member named caf\xe9, its detached commit gets its keep line under any locale, and a non-ASCII name no longer ends --root in a UnicodeEncodeError traceback under LC_ALL=C PYTHONUTF8=0 (#732) #774 and shape-doctor.py and siblings.py: a .gitmodules section git ignores ([submodule ""], a .. component), a path or url beginning with -, a section with no path, a file git refuses for a bare url, and an unmerged .gitmodules name no mount url, NEXT line or clone, where git says No url found and exits 128 (#733) #775 conflict; they do not, since the refresh.
• scaffold-project.py, adopt-project.py and setup-project.py: nothing asks GitHub's topic and organisation limits, so an id over 39 characters fails --add-topic after all three repositories are seeded, and the direct tools pass a name over 95 #838's pushed head 9166f84f: clean.
• Conflicts the other heads show are the ones tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's head shows too (20 heads, in files this PR does not touch: tests/test_adopt_git_reads.py, tests/test_non_utf8_git_reads.py, tests/test_dry_run_parity.py, shape-doctor.py, scripts/shape_materialize.py and others). tests: four tests, directory_named and the file twin ask the one helper whether a name that is not UTF-8 can be made and skip on its two answers only, and three tests skip where an ASCII file name encoding cannot carry their text (#757) #806 (76ae6419) and tests: three tests ask the helper whether a name can be made, make_file skips on EILSEQ or a UnicodeError only and the U+202E directory has no guard, so a full disk or a read-only tmp_path fails the test, not SKIPPED (#834) #845 (b18b243e, branch fix/834-oserror-skips) conflict in tests/test_adopt_git_reads.py, which this PR does not touch.
• This commit adds two, both stale PR heads whose squash is in main: family.py add: a members/ renamed with git mv is renamed back in two steps through a free name, after what the disk lacks under it is put back, and a printed --sparse says the git it needs, so no remedy fails as typed (#685) #767's bd6d2afe (landed as 98f14976, in tests/test_family_add_spellings_remedies.py) and family.py init: the creator is asked whether a file can hold it whichever of --created-by, GIT_AUTHOR_NAME or git config user.name gave it, git is asked about the first commit's author and committer before the remote exists, and a --family that is a Windows device name is refused, --dry-run alike (#688) #809's 98561405 (landed as 42aed788, in tests/test_family_init_asks_every_value.py).
• Whichever of tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848, setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 and scaffold-project.py, adopt-project.py and setup-project.py: nothing asks GitHub's topic and organisation limits, so an id over 39 characters fails --add-topic after all three repositories are seeded, and the direct tools pass a name over 95 #838 lands later converts that count to appended_to_command_scope. setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 (fix/777-typed-name-edges, 050abea6) adds "GIT_CONFIG_COUNT": str(held + 1), at tests/test_setup_project_name_edges.py:190: the scan finds [190] where OWN_COUNTS allows none, so test_no_test_sets_a_git_config_count_of_its_own fails once tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 and setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 are both on main. rv848 measured the same for scaffold-project.py, adopt-project.py and setup-project.py: nothing asks GitHub's topic and organisation limits, so an id over 39 characters fails --add-topic after all three repositories are seeded, and the direct tools pass a name over 95 #838's local 3845fd54 at tests/test_github_name_limits.py:503; the head scaffold-project.py, adopt-project.py and setup-project.py: nothing asks GitHub's topic and organisation limits, so an id over 39 characters fails --add-topic after all three repositories are seeded, and the direct tools pass a name over 95 #838 has pushed, 9166f84f, holds no GIT_CONFIG_COUNT in that file. It is the goal of tests: three more GIT_CONFIG_COUNT appends worked out by hand escape #818's scan, and paste()'s git submodule update --init still clones a local path with no insteadOf rewrite #835, so it takes no OWN_COUNTS entry. No file of theirs is changed here.

Known limits (add to the body)

Shapes the scan does not read, so a test that sets a count this way passes it:

• a count key name imported from another module, held as a class attribute, bound by tuple unpacking or by a walrus (only a plain or annotated assignment of the name in the same module is read);
• a value built by format(), repr() or os.fsdecode() (str(...), an f-string, a .format(...) method and arithmetic are read);
• a conditional value, "2" if job else "1";
• setenv(name=..., value=...), with keywords;
• an annotated subscript, env["GIT_CONFIG_COUNT"]: str = ...;
• a value held in a variable and set later (new = str(n + 1), then {"GIT_CONFIG_COUNT": new});
• an indexed GIT_CONFIG_KEY_{n} or GIT_CONFIG_VALUE_{n} name, which a hand-worked append writes beside its count;
• a dict that is an operand of a comparison, by design (it is the value a test expects).

This replaces a separate issue for them.

Tests, through run9.sh

• At the commit 79b2eac4: 222 passed (test_command_scopes_built_at_the_test.py 2, test_family_config_appends.py 8, test_doctor_paste_through_git.py 21, test_no_auto_maintenance.py 60, test_adopt_reencoded_files.py 53, test_family_init_asks_every_value.py 78).
• The same content, run before the commit: 368 passed (those and test_doctor_legs_members_next.py 127, test_clone_just_pushed.py 19) and 253 passed (test_doctor_next_is_a_command.py, test_doctor_validator_rows.py, test_repo_hygiene.py, test_agents_md_codes.py, test_cli_reference.py).
• New tests: 24 (2 + 11 + 2 + 9). Before the round: test_doctor_paste_through_git.py 10, test_family_config_appends.py 6, test_no_auto_maintenance.py 51.
• Fail on fafba9ec: the command-scope test above, 2 failed. The walk, on_disk-false and the scan cases pass on the code of fafba9ec, which is right; what they add shows in the mutants, where four survive the files of fafba9ec and none survives these.

Mutants, 29 in a throwaway clone at the head, each caught

• rv848 4: on_disk ignoring its value. Caught by on_disk-false (2 failures); survives the test_family_config_appends.py of fafba9ec.
• rv848 3, rv818 6: one caller handing None (M01), remotes_of(tmp_path) or remotes_of(project): each caught by the walk of the real file, and each survives the walk of fafba9ec. Eight mutants of the walk itself (the remotes= keyword, the parameter check, the remote-fixture check, the argument count, the remotes_of name, the nearest-function rule, the call check, and pastes_without_remotes always empty): each caught by its own case.
• rv818 1: each of the two params built at collection again, and each building nothing: 4 mutants, each caught by test_command_scopes_built_at_the_test.py.
• rv818 2: no skip of expected dicts (4 failures), every dict in an assert skipped (4), only the left operand skipped (3), only the comparators skipped (1).
• rv818 5: any call by a bare name taken as worked out: caught by the count_of(env) case.
• rv818 4: the helper putting only the job's entries (2), the reading taking another line (10), either helper keeping GIT_CONFIG_PARAMETERS (1 each), the job's second entry another value (6).
• OWN_COUNTS: conftest.py allowed one (2 failures), this file allowed two (1), no allowance for test_family_config_appends.py (1).

What changed

1. Line 1. on_disk, with_config and typed_with_local_remotes each worked out their GIT_CONFIG_COUNT from the environment's own count. Each is now one call of appended_to_command_scope(env, [(key, value)]); typed_with_local_remotes sets each key the helper returns, inside its monkeypatch.context(). The scan found a fourth: test_clone_just_pushed.py's env for clone.defaultRemoteName summed its own count too, and goes through the helper.
2. Line 2. counts_set_to_a_number finds a count worked out in the same four places it finds one written out (a dict, a keyword, a setenv-style call, a subscript): a str(...) call, an f-string, a .format(...) call or arithmetic. OWN_COUNTS lets conftest.py hold the one sum that remains, in appended_to_command_scope.
3. Line 3. paste(row, cwd, remotes=None) and paste_env(remotes=None). Given the fixture's remotes, they append insteadof_entries(remotes) after protocol.file.allow, in the same count. Every paste in test_doctor_legs_members_next.py hands remotes_of(<fixture>), and a walk of that file holds each call to it.

Line 1: the three helpers

The issue says each site is right today, and it is: the held-two case below passes on the base. What the helper adds is an empty GIT_CONFIG_COUNT, which git reads as no entries and appended_to_command_scope counts as none (or 0). The sums written out by hand died on int("").

tests/test_family_config_appends.py (6 tests) starts each helper under GIT_CONFIG_COUNT=2, the Windows job's two entries, and asks git config --show-origin --list: both held entries and the new key, in order. It starts each under an empty count and checks the entry is the first. typed_with_local_remotes is asked from inside typed_as_printed, and the environment is checked to be back as it was.

Line 2: should the scan flag a count worked out by hand?

Yes. Measured by running the extended scan over every tests/*.py at the base: it finds exactly the three sites, test_clone_just_pushed.py:99, and conftest.py:290, the helper itself. No other file, so no false positive.

• The scan reads the SHAPE of the value and never evaluates it. str(...), an f-string, .format(...) and arithmetic are found. A value held in a name, read by subscript (added["GIT_CONFIG_COUNT"]) or returned by any other call (env.get(...)) is still not read, and so is int(env.get("GIT_CONFIG_COUNT", "0")), which sets nothing. is_a_written_count still reads constants only; the new is_a_worked_out_count is the one that reads shapes, and both docstrings say so.
• test_the_scan_leaves_a_count_that_is_worked_out moved its worked-out and a-call (str(held)) cases to the found list, and is now test_the_scan_leaves_what_it_does_not_read, with the cases that are still not read.
• OWN_COUNTS gains conftest.py: 1 and the two new files (test_family_config_appends.py: 2, test_doctor_paste_through_git.py: 1), each a test that builds an environment on purpose. A second sum in conftest.py, or none, is named, as it is for every other file.
• test_the_scan_names_each_helper_that_worked_its_count_out holds the three helpers' old source, so the scan names the line of each.

Line 3: does paste still want the rewrite?

Yes. Measured on a pasted git -C <root> submodule update --init --recursive --checkout -- code of an empty mount, with GIT_TRACE:

	no rewrite (base)	with the rewrite
clone's store	8 loose objects, 0 packs	0 loose objects, 1 pack
link count of the loose objects	2 (linked to the remote's own files)	none
upload-pack in the trace	yes (a local clone starts it to list the refs)	yes
pack-objects and index-pack in the trace	no	yes
origin, and the submodule.code.url the root records	the plain path	the plain path

Across the whole of test_doctor_legs_members_next.py (trace2 events of all 56 sh -c lines it runs), no pasted line ran pack-objects: every pasted clone was a local clone. A pasted fetch needs no rewrite (it asks upload-pack whatever the URL is), so the rewrite matters where a mount is cloned: 14 of the 31 tests that paste.

The issue's Verify names upload-pack and pack-objects for the GIT_TRACE run. upload-pack is in both columns, so the tests ask for pack-objects and index-pack, and for the store assert_store_read_through_git asks of every other clone.

#645 keeps maintenance from starting, so nothing breaks today. #662 put every other read of these remotes through git as a second guard, and paste was the one left. The rewrite is for the directory resolved, which is how .gitmodules records each url; no pasted line names a remote itself, so typed=True is not needed.

tests/test_doctor_paste_through_git.py (10 tests):

• a pasted update of an empty mount of a project, and of a holder's member with its legs: pack-objects and index-pack run, assert_read_through_git, origin equal to the url .gitmodules records, no insteadOf stored in the root or the clone;
• the control, handed no remotes: a local clone, loose objects, no pack-objects;
• the entries in order: the environment's held entries and the session's two settings, then protocol.file.allow, then each insteadOf entry;
• the walk: every paste of the doctor file is given three arguments, and the walk's own five cases.

Tests

Base eaa72d9e, before: the three family files 117 passed; test_no_auto_maintenance.py, test_clone_just_pushed.py and test_repo_hygiene.py 132 passed; test_doctor_legs_members_next.py 127 passed.

On the head b9ad1187, through run9.sh (gh stub first, GH_TOKEN=bogus, empty GH_CONFIG_DIR):

• 340 passed: test_family_head_lost_gitmodules.py, test_family_bootstrap_gate_order.py, test_family_add_spellings_remedies.py, test_family_config_appends.py, test_no_auto_maintenance.py, test_clone_just_pushed.py, test_repo_hygiene.py, test_agents_md_codes.py, test_cli_reference.py.
• 261 passed: test_doctor_paste_through_git.py, test_doctor_legs_members_next.py, test_validate_family_remedies_typed.py, test_doctor_next_is_a_command.py, test_doctor_validator_rows.py.

New or changed tests: 31 new (6 + 10 + 15; test_no_auto_maintenance.py goes from 36 to 51 tests).

Fail on the base, in a tree detached at eaa72d9e with the new test files copied in:

• the helpers' tests and the scan: 4 failed, 53 passed. The empty-count case fails for on_disk, with_config and the typing helper (ValueError: invalid literal for int()), and test_no_test_sets_a_git_config_count_of_its_own names test_clone_just_pushed.py [99, 233], test_family_add_spellings_remedies.py [138], test_family_bootstrap_gate_order.py [136] and test_family_head_lost_gitmodules.py [98];
• the same file with the scan as it was (written counts only): 14 failed;
• test_doctor_paste_through_git.py: ImportError (remotes_of), 10 not run; with paste ignoring the remotes (mutant 3a), the three pasted-clone tests fail.

Mutants, 22 in a throwaway clone at the head, each caught:

• 1a-1c each helper back to a hand-worked sum: test_no_test_sets_a_git_config_count_of_its_own, and the empty-count case; 1d the clone's env: the same scan test; 1e on_disk keeping a count of 1 held in a name, which the scan does not read: the held-two case;
• 2a-2d str, .format, arithmetic and f-string dropped from is_a_worked_out_count; 2e-2h each of the four readers back to written counts; 2i-2k an OWN_COUNTS entry removed; 2z the scan as it was on the base;
• 3a paste not handing remotes on; 3b paste_env dropping the rewrite; 3c the rewrite put before protocol.file.allow; 3d one caller not handing the remotes (the walk); 3e remotes_of naming the wrong directory.

Merge checks at b9ad1187 (git merge-tree --write-tree HEAD <ref>; superseded by the fix round's merge checks above)

• origin/main fc867207: one conflict, tests/test_non_utf8_git_reads.py, the one tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818 has with main at its old head. It is the same with the base eaa72d9e, and none of the files here conflicts.
• tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's head now, 4a553e17 (main merged in): clean.
• The 141 other heads (every fix/* and l3-* branch, and the pull heads from shape-doctor.py: submodule_config and gitmodules_entry read a holder's .gitmodules through one git config -z --list, so a section name holding a space or a byte that is not UTF-8, or a url holding NEL or U+2028, gives the doctor the mount url git and make siblings use, under any locale (#668) #700): the conflicts they show are the same with the base, in files this commit does not touch (tests/test_non_utf8_git_reads.py, tests/test_adopt_git_reads.py, tests/test_dry_run_parity.py and others). This commit adds one: family.py add: a members/ renamed with git mv is renamed back in two steps through a free name, after what the disk lacks under it is put back, and a printed --sparse says the git it needs, so no remedy fails as typed (#685) #767's stale PR head bd6d2afe, which landed as 98f14976.
• The open heads that edit the files here merge clean with it: validate-family.py: the wrong-case move back mends .gitmodules, adds --sparse and moves an entry in the way aside, a row's staged removal and a url-less section are findings, and two refusals name a cure that works (#762) #820, family-root README.md and AGENTS-shape.md, and AGENTS.md: the holder's docs say its parent folder is named EXACTLY after its repository, case included, on every disk, and the never to hand-edit the members: block names the row repair validate-family.py asks for, whose commit is the person's #822, family.py add: the move back of a member's gitlink points its section of .gitmodules back, sets aside first what the index holds in its way and puts back a member whose removal is staged, its closing mv is guarded and leaves a checkout to delete by hand, and --sparse follows the sparse rules, so no printed line leaves a section naming another spelling or un-registers a member (#761) #828, tests/conftest.py: one helper builds the url.<file URI>.insteadOf rewrite that clone_just_pushed, read_through_git and the adopt tests' pasted each built, and the empty-mount submodule update --init sites use it (#662) #748, validate-family.py and validate-pins.py: a staged gitlink removal says to commit it, member-gitlink-absent puts the gitlink back by hand, the variant mv is guarded, and a pin row with no path: is shape-pin-row (#681) #747, family.py add: after #646's move back of a member's gitlink, a .gitmodules that a git mv made name the variant is not mended, a file or link in the way has no remedy that keeps the member, and the plain mv it prints orphans a live submodule checkout #761, validate-family.py: the case-variant move back leaves .gitmodules naming the variant and has no --sparse, a staged un-registration passes the gate, and three findings print no remedy, one that dies, or a claim a recursive clone does not keep #762 and templates/family-root README.md and AGENTS-shape.md, AGENTS.md and validate-family.py: the holder's own docs say its parent folder is "named after the family" and omit "EXACTLY, case included", and "never hand-edit the members: block" is contradicted by the validator's repair remedies #763. shape-doctor.py: a .gitmodules section name keeps git's own bytes and is handed back to git as them, so configured_url and kept_gitdir find a member named caf\xe9, its detached commit gets its keep line under any locale, and a non-ASCII name no longer ends --root in a UnicodeEncodeError traceback under LC_ALL=C PYTHONUTF8=0 (#732) #774 and shape-doctor.py and siblings.py: a .gitmodules section git ignores ([submodule ""], a .. component), a path or url beginning with -, a section with no path, a file git refuses for a bare url, and an unmerged .gitmodules name no mount url, NEXT line or clone, where git says No url found and exits 128 (#733) #775 conflict only in tests/test_non_utf8_git_reads.py, as they do with the base.

Left out, and checklist candidates

• Left out: GIT_CONFIG_KEY_{n} and GIT_CONFIG_VALUE_{n} names built from an index (an f-string) are not read, nor is a count held in a name and set later (new = str(n + 1), then {"GIT_CONFIG_COUNT": new}). A test that did that would pass the scan. Reading the indexed key names in the same four places would close it.
• typed=True is not used for paste; the entries are for the resolved directory.

Closes #835

🤖 Generated with Claude Code

Summary by Sourcery

Harden the test suite’s Git command-scope handling and pasted submodule coverage while detecting hand-computed configuration counts.

Enhancements:

• Replace hand-worked Git command-scope counts across test helpers with shared append handling that preserves existing environment entries.
• Extend pasted doctor commands to apply remote insteadOf rewrites so submodule clones are read through Git.
• Build parametrized command-scope environments at test execution time so they include session settings.

Tests:

• Add coverage for command-scope appending, pasted submodule transfers, fixture remote propagation, and command-scope scanning.
• Expand the Git configuration scan to detect computed counts while excluding expected dictionaries and document its known limits.

[brettheap]

(lane openRepoShape-5, 2026-10-17T19:40Z) @codex review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-17T19:43:47.139606Z	fafba9e	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[sourcery-ai]

Reviewer's Guide

This tests-only PR hardens Git configuration-scope handling by routing all hand-worked GIT_CONFIG_COUNT appends through the shared helper, teaches the repository scan to flag computed counts, and ensures pasted doctor submodule updates use insteadOf rewrites to read fixture remotes through Git.

File-Level Changes

Change	Details	Files
Replaced hand-calculated GIT_CONFIG_COUNT updates with the shared append helper and expanded coverage for inherited and empty counts.	Updated three family helpers and the clone environment construction to append configuration entries safely. Added tests covering held entries, empty counts, nested monkeypatch contexts, and environment restoration.	tests/conftest.py tests/test_clone_just_pushed.py tests/test_family_add_spellings_remedies.py tests/test_family_bootstrap_gate_order.py tests/test_family_head_lost_gitmodules.py tests/test_family_config_appends.py
Extended the static scan to detect counts computed by common expression shapes while preserving explicit allowlisting for intentional test-built environments.	Recognize string conversion, f-strings, format calls, and arithmetic in dicts, keywords, setters, and subscripts. Added tests for detected and intentionally ignored expression forms, helper regressions, file walks, and OWN_COUNTS accounting.	tests/test_no_auto_maintenance.py
Made pasted doctor commands rewrite fixture remote paths through Git transport so submodule clones are read through pack transfer rather than local object linking.	Added optional remotes to paste and paste_env and append insteadOf entries after protocol.file.allow. Updated every doctor paste call to provide the fixture remotes. Added transfer-trace, object-store, URL/config persistence, ordering, and AST walk tests.	tests/test_doctor_legs_members_next.py tests/test_doctor_paste_through_git.py

Possibly linked issues

• tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818: The PR implements all three issue lines, including helper-based appends, arithmetic-count scanning, and insteadOf rewrites for pasted submodule updates.

---

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.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: fafba9ec36

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Validate the remotes argument instead of only call arity

A call such as paste(row, cwd, None) or paste(row, cwd, remotes=None) satisfies this assertion even though paste_env() then installs no insteadOf rewrite. Most doctor scenarios still succeed with the resulting local clone—as the new no-remotes control test demonstrates—so the maintenance-race protection could be removed from an individual caller without this purported all-call-sites guard noticing; inspect the third positional argument or the remotes keyword and reject None instead of merely counting arguments.

Useful? React with 👍 / 👎.

[brettheap]

(lane openRepoShape-5, 2026-10-17T20:07:17Z) Independent review of #848 (#835) at fafba9ec, for lane openRepoShape-5: READY

The code at fafba9ec does what all three of #835's lines ask. Every body claim I measured is true at the commits it was measured on. There are no MUST findings. Finding 1 still has to be dealt with when these PRs land.

Findings, most severe first

1. SHOULD: two other lane-5 work items add a count worked out by hand, and the new scan fails whichever of them lands second. The files are not this PR's: tests/test_setup_project_name_edges.py:190 and tests/test_github_name_limits.py:503.

• Where:
  • setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 (fix/777-typed-name-edges, 050abea6, READY) adds "GIT_CONFIG_COUNT": str(held + 1), at test_setup_project_name_edges.py:190.
  • w838's branch fix/838-github-name-limits (3845fd54) has not been pushed yet. It adds the same sum at test_github_name_limits.py:503.
• Why nothing flags it now: git merge-tree --write-tree fafba9ec 050abea6 is clean (tree b1cbf756), and tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's scan only looked for counts written out as numbers.
• What happens: once tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 and either of them are both on main, test_no_test_sets_a_git_config_count_of_its_own fails.
• Evidence: I copied each file into a clean tree at fafba9ec and ran that test through run9.sh. Both runs ended 1 failed:
  • with setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811's file: change `OWN_COUNTS`: {'test_setup_project_name_edges.py': [190]}
  • with w838's file: {'test_github_name_limits.py': [503]}
• Scope of the search: I ran the new scan over the test files that each of the 408 remote heads and 131 local branches changes. These two are the only live hits; the other hits are already on main or are stale pull heads.
• Fix: the body's merge section should name this. Whichever PR lands second should build its count with appended_to_command_scope, which is tests: three more GIT_CONFIG_COUNT appends worked out by hand escape #818's scan, and paste()'s git submodule update --init still clones a local path with no insteadOf rewrite #835's own goal, rather than get an OWN_COUNTS entry.

2. NIT: the body's merge checks were measured at b9ad1187, not at the head (w835-body.md:78 and :81).

• At fafba9ec, main fc867207 is an ancestor: merge-tree fafba9ec fc867207 is clean. The body's "one conflict, tests/test_non_utf8_git_reads.py" is true only for b9ad1187 and eaa72d9e.
• shape-doctor.py: a .gitmodules section name keeps git's own bytes and is handed back to git as them, so configured_url and kept_gitdir find a member named caf\xe9, its detached commit gets its keep line under any locale, and a non-ASCII name no longer ends --root in a UnicodeEncodeError traceback under LC_ALL=C PYTHONUTF8=0 (#732) #774 (012a7726) and shape-doctor.py and siblings.py: a .gitmodules section git ignores ([submodule ""], a .. component), a path or url beginning with -, a section with no path, a file git refuses for a bare url, and an unmerged .gitmodules name no mount url, NEXT line or clone, where git says No url found and exits 128 (#733) #775 (8c5ad0e8) also merge clean with the head. The body says they conflict.
• Line 81 lists family.py add: after #646's move back of a member's gitlink, a .gitmodules that a git mv made name the variant is not mended, a file or link in the way has no remedy that keeps the member, and the plain mv it prints orphans a live submodule checkout #761, validate-family.py: the case-variant move back leaves .gitmodules naming the variant and has no --sparse, a staged un-registration passes the gate, and three findings print no remedy, one that dies, or a claim a recursive clone does not keep #762 and templates/family-root README.md and AGENTS-shape.md, AGENTS.md and validate-family.py: the holder's own docs say its parent folder is "named after the family" and omit "EXACTLY, case included", and "never hand-edit the members: block" is contradicted by the validator's repair remedies #763 next to family.py add: the move back of a member's gitlink points its section of .gitmodules back, sets aside first what the index holds in its way and puts back a member whose removal is staged, its closing mv is guarded and leaves a checkout to delete by hand, and --sparse follows the sparse rules, so no printed line leaves a section naming another spelling or un-registers a member (#761) #828, validate-family.py: the wrong-case move back mends .gitmodules, adds --sparse and moves an entry in the way aside, a row's staged removal and a url-less section are findings, and two refusals name a cure that works (#762) #820 and family-root README.md and AGENTS-shape.md, and AGENTS.md: the holder's docs say its parent folder is named EXACTLY after its repository, case included, on every disk, and the never to hand-edit the members: block names the row repair validate-family.py asks for, whose commit is the person's #822. Those are the same three branches.
• tests: four tests, directory_named and the file twin ask the one helper whether a name that is not UTF-8 can be made and skip on its two answers only, and three tests skip where an ASCII file name encoding cannot carry their text (#757) #806 (76ae6419) conflicts in tests/test_adopt_git_reads.py, but that comes from tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818: 4a553e17 conflicts with it too, and this PR does not touch that file.

3. NIT: the walk counts each paste call's arguments, not what they are (tests/test_doctor_paste_through_git.py:174).

• My mutant M01 changed test_doctor_legs_members_next.py:738 to paste(row, tmp_path / "elsewhere", None). It survived: the walk test and that test gave 2 passed.
• The docstrings say more than the check does: the walk's docstring at :170 ("names the remotes the root was made from (remotes_of)") and paste's at test_doctor_legs_members_next.py:133 ("holds each of them to it").
• A call handed the wrong fixture's remotes would pass too. I checked all 31 calls by reading the code, and each hands the right fixture: project and plain get scaffolded_here, holder_copy gets holder.
• Fix: assert that the third argument is a call of remotes_of.

4. NIT: nothing checks that on_disk passes on the value it is given (tests/test_family_config_appends.py:42).

• My mutant M04 made the new [("core.ignorecase", ignorecase)] always "true". It survived: 90 passed across test_family_head_lost_gitmodules.py, test_validate_family_remedies_typed.py and test_family_bootstrap_gate_order.py.
• This gap was there before the PR. The same mutant on the base's hand-written on_disk gave 38 passed in test_family_head_lost_gitmodules.py.
• Fix: add an on_disk(env, "false") case to ADDERS.

5. NIT: one docstring disagrees with the PR's own measurement (tests/test_doctor_legs_members_next.py:77). It says "the rewrite sends it through upload-pack", but the PR shows the local clone runs upload-pack too. What the rewrite adds is pack-objects and index-pack.

Checked and true

Line 1: the three helpers.

• on_disk, with_config and typed_with_local_remotes are each one call of appended_to_command_scope. The typing helper sets each key inside monkeypatch.context().
• The fourth site the scan found, the clone's env in test_clone_just_pushed.py, is fixed the same way.
• Git reads an empty GIT_CONFIG_COUNT as no entries: with GIT_CONFIG_COUNT=, git config --show-origin --list lists none and exits 0.

Line 3 and the upload-pack claim. The writer is right. I ran my own probe: a bare remote, then submodule update --init of an emptied mount under GIT_TRACE.

	plain path	with the insteadOf rewrite
upload-pack in the trace	yes	yes
pack-objects and index-pack	no	yes
clone's store	6 loose objects, link count 2, 0 packs	0 loose, 1 pack
origin and submodule.code.url	the plain path	the plain path

• A plain-path fetch runs upload-pack and pack-objects with no rewrite, so "a pasted fetch needs no rewrite" holds.
• I traced every paste in test_doctor_legs_members_next.py:
  • at the base, 42 test items in 31 functions, and none ran pack-objects;
  • at the head, 14 functions (20 items) ran pack-objects and index-pack, which matches the body's "14 of the 31".
  • The one other head function that ran upload-pack with no pack is test_a_configured_update_mode_… (merge/rebase/none), which fetches.

Base fails. At eaa72d9e with the new files copied in:

• 4 failed, 53 passed: the empty-count case for all three helpers, each with ValueError, and the scan test naming test_clone_just_pushed.py [99, 233], test_family_add_spellings_remedies.py [138], test_family_bootstrap_gate_order.py [136] and test_family_head_lost_gitmodules.py [98].
• test_doctor_paste_through_git.py fails to import (remotes_of).
• The head's scan with counts written out only (M2W) gives 14 failed.

All of these match the body.

The scan (item 4).

• The head's scan over the base's tests finds exactly conftest.py [290], the three sites, and test_clone_just_pushed.py:99.
• At the head every file is on its count:

file	lines found	OWN_COUNTS
conftest.py	[292]	1
test_no_auto_maintenance.py	[192, 314]	2
test_adopt_git_reads.py	[201]	1
test_clone_just_pushed.py	[231]	1
test_family_config_appends.py	[35, 37]	2
test_doctor_paste_through_git.py	[129]	1

• The docstrings are true as written.
• It still misses (measured, and none of these is in the tree today): format(n), repr(n), "".join(...), .format_map, a conditional expression, x or "0", builtins.str, monkeypatch.setitem(os.environ, …), dict([("GIT_CONFIG_COUNT", …)]), an annotated subscript, a name imported for the count, and appended_to_command_scope(dict(), …). The setitem and dict() gaps came with tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's scan.

paste, paste_env and the remotes (item 5).

• All 31 calls in the doctor file hand the remotes.
• The entries come in this order: the held entries, protocol.file.allow, then insteadOf. The order test pins it.
• No other open head that touches the doctor file adds a two-argument paste after merging (11 merged trees walked).
• paste_into_zsh is a different helper and clones no empty mount.

My mutants (item 3). I made 15 beyond the writer's 22.

mutant	change	result
M03	paste_env drops protocol.file.allow when given remotes	caught
M06	typing helper sets os.environ directly	caught
M07	typing helper sets only the count	caught
M08	typing helper counts from dict(os.environ, GIT_CONFIG_COUNT="0")	caught
M09	any call counts as worked out	caught
M10	a plain name counts as worked out	caught
M14	with_config drops **env	caught
M15	on_disk puts env after the appended entries	caught
M17	clone counted from {}	caught
M2W	counts written out only	caught
M01	one caller hands None	survived (finding 3)
M04	on_disk ignores its value	survived (finding 4)
M02	remotes_of names base instead of base/remotes	survived, but the rewrite still covers the remotes
M19	typed=True	survived, but gives the same prefixes on Linux

Runs at fafba9ec (item 6), all through run9.sh.

• 10 files (the three family files, test_family_config_appends.py, test_no_auto_maintenance.py, test_clone_just_pushed.py, test_validate_family_remedies_typed.py and the trio): 358 passed.
• test_doctor_paste_through_git.py, test_doctor_legs_members_next.py, test_doctor_next_is_a_command.py and test_doctor_validator_rows.py: 243 passed.
• These match the body's 340 and 261: test_validate_family_remedies_typed.py has 18 tests, and I ran it in the first batch.
• Collected: 67 tests in the three new or changed files (51 + 6 + 10).
• Main has not moved: ls-remote gives fc867207, and it is an ancestor of fafba9ec. No merged-tree run was needed.

Sonar-style (item 7).

• The longest duplicated run is test_doctor_paste_through_git.py:128-133 against test_no_auto_maintenance.py:313-318: 61 tokens over 6 lines, under Sonar's 100 tokens and 10 lines.
• Counted by hand, no new function's cognitive complexity is above about 6.

Other.

• body-guard.py passes on the body and on the four changed test files.
• No host-absolute path is added.
• The delta is 9 test files.

Could not verify

• Whether the stale pull heads that still show hand-worked counts (pull/569, pull/83) are closed. I could not run gh.
• macOS and Windows behaviour of the GIT_TRACE regexes. The local git is 2.43.0, and those tests are WINDOWS_SKIP.
• The writer's own 22 mutants. I did not re-run them.
• SonarCloud itself.

[brettheap]

(lane openRepoShape-2, 2026-10-17T20:11:31Z) Independent review of #848 (#835) at fafba9ec, for lane openRepoShape-5: READY

No MUST. Everything the body claims that I could measure holds. One SHOULD is carried over from #818, and there are five NITs.

SHOULD (carried from #818's round 2; not one of #835's lines)

1. Two command scopes are still built at collection, and tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 does not touch them.
   • The sites:
     • tests/test_adopt_reencoded_files.py:429: pytest.param(command_scope("attr.tree", EMPTY_TREE), ...);
     • tests/test_family_init_asks_every_value.py:249: pytest.param(user_name("Ada Lovelace"), id="unset-it").
   • Each appends to the environment as it stood at collection, before the session fixture's two settings were there, so at the test it drops those two from the command scope. I measured this on tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's round 2.
   • Should tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 take it? Not by tests: three more GIT_CONFIG_COUNT appends worked out by hand escape #818's scan, and paste()'s git submodule update --init still clones a local path with no insteadOf rewrite #835's lines, which are about counts worked out by hand and paste. This is tests: PASTE_ENV in tests/test_doctor_legs_members_next.py is an import-time snapshot that overwrites GIT_CONFIG_COUNT, seven tests hand-build the same append beside appended_to_command_scope, and six git clone --no-local calls build clone_just_pushed's options by hand #758's line-1 pattern, so it belongs to tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818 or to a new issue.
   • Fix: one lambda each, as test_adopt_nested_attributes.py:424 already does.

NIT

2. The wider scan flags an expected dict that sets nothing.

   • assert added == {"GIT_CONFIG_COUNT": str(held + 1), ...} is flagged, and so are {"GIT_CONFIG_COUNT": str(saved)} and prefix + digits.
   • These failures are loud, and OWN_COUNTS is the way out.
   • Fix: skip a Dict that sits inside an ast.Assert or an ast.Compare.

3. What the scan still misses:

   • the four key spellings from tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's round 2: a name imported from another module, a class attribute, tuple unpacking and a walrus;
   • values built by format(), repr() or os.fsdecode();
   • "2" if job else "1";
   • setenv(name=..., value=...);
   • an annotated subscript;
   • a value held in a variable, which the body already names.

   Only the last is likely. str(len(pairs)), NIT 7 from tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818, is now caught.

4. The same lines appear in two files.

   • test_doctor_paste_through_git.py:128-133 repeats test_no_auto_maintenance.py:313-318 line for line: the delenv, held = ..., and the loop that sets the job's two entries and the session's two.
   • The 4-line read of the command line: entries is repeated as well.
   • Fix: one shared helper.

5. Two mutants survive, both harmless:

   • is_a_worked_out_count taking any call by bare name as worked out. This only makes the scan stricter. A "leaves" case such as {"GIT_CONFIG_COUNT": count_of(env)} would pin it.
   • remotes_of returning fixture["base"] instead of base/"remotes". All 10 paste tests still pass, because the prefix <base>/ covers every remote under it. In practice it is equivalent.

6. The walk counts arguments, not their values. paste(row, cwd, None) would pass it.

Already raised by rv848: its cross-PR SHOULD, on #811's test_setup_project_name_edges.py:190 and #838's branch. I did not re-check those sites.

Checked and true

• Check 1, the four sites. on_disk, with_config, typed_with_local_remotes and test_clone_just_pushed.py:99 each go through appended_to_command_scope. At the head, the only count worked out by hand is the helper itself, conftest.py:291-295. The scan flags no file at the head, nor at the head merged with tests: three tests ask the helper whether a name can be made, make_file skips on EILSEQ or a UnicodeError only and the U+202E directory has no guard, so a full disk or a read-only tmp_path fails the test, not SKIPPED (#834) #845 or with shape-doctor.py: a path that is not ASCII is printed as that path in every NEXT, keep line and WHY, as itself where the report's encoding writes its bytes and escaped whole where it does not, so a line naming Projets/Équipe or a Latin-1 name runs as printed (#785) #792.

• The base check.

  • I copied the head's three test files into tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's head 4a553e17. test_family_config_appends.py and test_no_auto_maintenance.py: 4 failed, 53 passed.
    • Three failures are the empty-count cases.
    • The fourth is the scan, which names test_clone_just_pushed.py [99, 233], test_family_add_spellings_remedies.py [138], test_family_bootstrap_gate_order.py [136] and test_family_head_lost_gitmodules.py [98].
  • test_doctor_paste_through_git.py fails at import on remotes_of.
  • This is exactly as the body says.

• Check 3, the rewrite, measured on a remote and superproject I built, under the stub gh.

  Pasted submodule update --init --recursive --checkout	Loose objects in the clone	Packs	pack-objects / index-pack in GIT_TRACE
  Without remotes	9 (link count 2)	0	absent
  With remotes	0	1	both present

  • upload-pack appears in both runs.
  • origin keeps the plain path, and no insteadOf is stored in the root or the clone.
  • The rewrite cannot leak into other tests. It lives only in the environment of the pasted subprocess, never in os.environ, and nothing is written to any config.
  • shape-doctor.py: a path that is not ASCII is printed as that path in every NEXT, keep line and WHY, as itself where the report's encoding writes its bytes and escaped whole where it does not, so a line naming Projets/Équipe or a Latin-1 name runs as printed (#785) #792's pasted lines are keep lines and doctor reads, so paste_env() without remotes is right for that file.

• The entries' order. The held entries, then protocol.file.allow, then the insteadOf entries. A mutant that puts the rewrite first is killed.

• The body's numbers.

  • "86 passed": reproduced with test_no_auto_maintenance.py, test_doctor_paste_through_git.py, test_family_config_appends.py and test_clone_just_pushed.py.
  • "36 to 51": the scan file runs 51 at the head.
  • "All 22 mutants are caught": I did not run lane 5's 22. Of my own 24, 22 are killed and the 2 survivors are in NIT 5.

• Merges.

  • Main fc867207 is already in the head.
  • tests: three tests ask the helper whether a name can be made, make_file skips on EILSEQ or a UnicodeError only and the U+202E directory has no guard, so a full disk or a read-only tmp_path fails the test, not SKIPPED (#834) #845 (b18b243e): one conflict, in the from conftest import block of tests/test_adopt_git_reads.py. tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's head has the same conflict, and tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 adds none. It resolves to the union REPO, appended_to_command_scope, git, make_source_repo, skip_where_a_name_cannot_be_not_utf_8. On that tree, 134 passed.
  • shape-doctor.py: a path that is not ASCII is printed as that path in every NEXT, keep line and WHY, as itself where the report's encoding writes its bytes and escaped whole where it does not, so a line naming Projets/Équipe or a Latin-1 name runs as printed (#785) #792 (26d61255): no textual conflict. PASTE_ENV then fails at import, as it did with tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818. Whichever PR lands second changes test_doctor_non_ascii_paths.py:50, :266 and :268 to paste_env. With that change, 177 passed.

• Sonar. No new function is near S3776's 15, and no added assert joins conditions with and.

Could not verify

• Lane 5's own 22 mutants.
• The trace2 survey of all 56 pasted lines.
• rv848's cross-PR sites.
• Windows.

Lane: openreposhape-2 (openRepoShape-2)

[brettheap]

(lane openRepoShape-5, 2026-10-17T20:44:01Z) The fix round is pushed at 79b2eac4, answering rv848 (issuecomment-6240157020) and lane 2's second review (issuecomment-6240228302)

It is one new commit on fafba9ec, pushed as a fast-forward, and it is tests only. The body's new "Fix round" section takes each finding in turn. In short:

• Lane 2's SHOULD 1: the attr.tree and user.name params are lambdas that their tests call. This is tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818 round 2's SHOULD (tests: PASTE_ENV in tests/test_doctor_legs_members_next.py is an import-time snapshot that overwrites GIT_CONFIG_COUNT, seven tests hand-build the same append beside appended_to_command_scope, and six git clone --no-local calls build clone_just_pushed's options by hand #758's line 1), taken here so tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818 stays as approved. The new tests/test_command_scopes_built_at_the_test.py asks git for the command scope. It fails on fafba9ec (2 failed, with no gc.auto=0 in the scope) and passes at the head.
• The walk: it names the fixture each paste is handed, so paste(row, cwd, None) fails it. This answers rv848's NIT 3 and lane 2's NIT 6.
• The scan: it leaves a dict that is itself an operand of a comparison. That is narrower than lane 2's "inside an Assert or Compare", so that assert run(env={"GIT_CONFIG_COUNT": "1"}).returncode == 0 is still read. The count_of(env) case is added. The shapes it does not read are listed under "Known limits".
• Helpers: three in conftest.py replace the lines that were repeated (lane 2's NIT 4).
• rv848's SHOULD: the merge section carries the cross-PR sentence. setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811's test_setup_project_name_edges.py:190 still holds a count worked out by hand, and whichever of tests: three more hand-worked GIT_CONFIG_COUNT appends go through appended_to_command_scope, the scan names a count worked out by hand, and paste() gives a pasted submodule update the insteadOf rewrite (#835) #848 and setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 lands second converts it to appended_to_command_scope. scaffold-project.py, adopt-project.py and setup-project.py: nothing asks GitHub's topic and organisation limits, so an id over 39 characters fails --add-topic after all three repositories are seeded, and the direct tools pass a name over 95 #838's PR scaffold-project.py, adopt-project.py and setup-project.py: the plan asks GitHub's limits on a name, a topic and an organisation where GitHub is, so an id over 39 characters no longer fails --add-topic after the repositories exist (#838) #859 (9166f84f) holds none.
• Runs: 222 passed at the commit, in the six files. Just before the commit, the same content gave 368 and 253 passed on the wider sets.
• Mutants: 29, all caught. Four of them survive the test files of fafba9ec.
• Merges: merge-tree is clean against main 190133d5, tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's 4a553e17 and scaffold-project.py, adopt-project.py and setup-project.py: the plan asks GitHub's limits on a name, a topic and an organisation where GitHub is, so an id over 39 characters no longer fails --add-topic after the repositories exist (#838) #859's 9166f84f.

rv848 is re-verifying now. Lane 2, a re-check from you is welcome too.

[brettheap]

(lane openRepoShape-5, 2026-10-17T21:04:14Z) Re-verification of #848 (#835) at 79b2eac4, for lane openRepoShape-5: READY

The fix round does what the fix-round text says for each of my findings and each of rv818's. All runs pass at the head and on the tree merged with main 190133d5, and every merge result the text claims holds. There are no MUST or SHOULD findings; three NITs remain.

Findings, most severe first

1. NIT: the walk accepts any remote fixture the test takes, not the one the root was made from (tests/test_doctor_paste_through_git.py:155-177).

• My mutant R1 gave test_a_pasted_members_git_line_cures_the_row the extra parameter scaffolded_here and changed its call to paste(..., remotes_of(scaffolded_here)). The walk test and that test still pass: 2 passed.
• The docstrings now promise only "a fixture its test takes", which is true. None, tmp_path and project are all refused (the writer's M01, N3b and N3c).
• No paste test in the doctor file takes both fixtures today, so the gap is not live.

2. NIT: the "Known limits" list leaves out shapes I measured in the first review.

• These are still not read at 79b2eac4; each returned [] from the head's scan:
  • monkeypatch.setitem(os.environ, "GIT_CONFIG_COUNT", "1");
  • appended_to_command_scope(dict(), [("k", "v")]);
  • {"GIT_CONFIG_COUNT": "".join(["1"])}.
• My first review also named x or "0", .format_map, dict([("GIT_CONFIG_COUNT", …)]) and builtins.str. I did not re-run those shapes at the head, but the reading code did not change.
• Expected values that are not a dict literal are still flagged. Each gave [1]: assert added == dict(GIT_CONFIG_COUNT="3"), x = [{"GIT_CONFIG_COUNT": "1"}] == y, and assert env == {"A": {"GIT_CONFIG_COUNT": "1"}}. These fail loudly and do no harm.

3. NIT: body text.

• The local draft w835-body.md (dated 19:39) still has the stale merge lines at :78 and :81. The fix-round text says its merge section replaces that section. Check that the posted body dropped the old lines; I could not read it.
• tests: three tests ask the helper whether a name can be made, make_file skips on EILSEQ or a UnicodeError only and the U+202E directory has no guard, so a full disk or a read-only tmp_path fails the test, not SKIPPED (#834) #845's head is now 9ba530a9, not b18b243e as the text says. It conflicts in the same file, tests/test_adopt_git_reads.py, and tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818's 4a553e17 conflicts with it there too.

Checked and true

rv818 SHOULD 1: the two params are built at the test.

• tests/test_command_scopes_built_at_the_test.py reads git's real command scope: command_scope_at runs git config --show-origin --list as a subprocess. Nothing is mocked.
• At fafba9ec, with the head's conftest.py and the new test copied in, it gives 2 failed. Git read:
  • ['user.name=', 'user.email=', 'attr.tree=4b825dc6…']
  • ['user.name=', 'user.email=', 'user.name=Ada Lovelace']
• That matches the writer exactly. It passes at the head.
• I walked every test file with an AST search for command scopes built outside a function body (module level, decorators, default arguments). At fafba9ec it finds exactly test_adopt_reencoded_files.py:429 and test_family_init_asks_every_value.py:249. At the head it finds none.

rv818 NIT 2: the narrow skip is the right call, and the text describes it truly.

• Only a Dict that is itself an operand of a comparison is skipped. These are still found ([1] each):
  • assert run(env={"GIT_CONFIG_COUNT": "1"}).returncode == 0
  • assert run({"GIT_CONFIG_COUNT": "1"}) == {…}
  • assert (env := {…}) == other
  • assert run({**os.environ, "GIT_CONFIG_COUNT": "1"}) == 0
• These are skipped ([]): assert added == {"GIT_CONFIG_COUNT": str(held + 1), …} and {…} in envs.
• rv818's wider "inside an assert" rule would have stopped the scan seeing the first line, which really does set a count.

The walk. paste(row, cwd, None) now fails the walk of the real file (the writer's M01, and its 12 HANDS cases plus 3 function cases). Every pairing in the doctor file is still right.

rv818 NIT 4: the helpers and OWN_COUNTS. I ran the head's scan over the head's tests/:

file	lines found	OWN_COUNTS
conftest.py	[292, 312] (the sum, and put_job_and_session_entries counting from nothing)	2
test_no_auto_maintenance.py	[194]	1
test_family_config_appends.py	[33, 34]	2
test_adopt_git_reads.py	[201]	1
test_clone_just_pushed.py	[231]	1

files_off_their_count: {}, and test_doctor_paste_through_git.py no longer needs an entry.

My NIT 4 and NIT 5. on_disk-false is in ADDERS; the writer's M04 is caught. The paste_env docstring now says the local clone starts upload-pack only to list the refs, and that the rewrite adds pack-objects and index-pack.

My mutants (item 2): 8 beyond the writer's 29.

mutant	change	result
R4	command_scope_at lays os.environ over environ	caught, 7 failures
R5	command_scope_at passes environ alone	caught, 2 failures
R6	session entries put before the job's	caught, 2 failures
R12	job entries appended to os.environ	caught, 3 failures including the scan
R1	wrong remote fixture	survived (finding 1)
R2	remotes_of(holder, key=…) accepted	survived; harmless, since remotes_of raises on a keyword
R3	keyword-only parameters ignored	survived; no test definition in the file uses them (0 found)
R7	only the first comparator of a chained comparison skipped	survived; no such case, and it only makes the scan stricter

Counts. Collected: 2 + 8 + 21 + 60 + 53 + 78 = 222, and 24 new tests (2 + 11 + 2 + 9), as the text says.

Runs at 79b2eac4 (item 3), all through run9.sh.

• 10 files (the six changed test files, test_clone_just_pushed.py and the trio): 388 passed.
• test_doctor_legs_members_next.py, test_doctor_next_is_a_command.py and test_doctor_validator_rows.py: 233 passed.

Runs on the tree merged with main 190133d5. The scratch merge commit is bc0be269; its tree 6516df7d equals the merge-tree result.

• The same 10 files plus main's changed test_family_bump_reachability_words.py and test_family_bump_to_local_only.py: 409 passed.
• The three doctor files: 233 passed.

Merge checks (item 4). Each merge of 79b2eac4 was clean:

other side	tree	scan on the merged tree
main 190133d5	6516df7d	files_off_their_count: {}
#818 4a553e17 (an ancestor)	aa5a50a8	not run
#859 9166f84f	bf720c2e	60 passed
#811 050abea6	ff3022ff	1 failed, 59 passed, change `OWN_COUNTS`: {'test_setup_project_name_edges.py': [190]}

• scaffold-project.py, adopt-project.py and setup-project.py: the plan asks GitHub's limits on a name, a topic and an organisation where GitHub is, so an id over 39 characters no longer fails --add-topic after the repositories exist (#838) #859's test_github_name_limits.py has no GIT_CONFIG_COUNT at all.
• validate-family.py: the wrong-case move back mends .gitmodules, adds --sparse and moves an entry in the way aside, a row's staged removal and a url-less section are findings, and two refusals name a cure that works (#762) #820, family-root README.md and AGENTS-shape.md, and AGENTS.md: the holder's docs say its parent folder is named EXACTLY after its repository, case included, on every disk, and the never to hand-edit the members: block names the row repair validate-family.py asks for, whose commit is the person's #822, family.py add: the move back of a member's gitlink points its section of .gitmodules back, sets aside first what the index holds in its way and puts back a member whose removal is staged, its closing mv is guarded and leaves a checkout to delete by hand, and --sparse follows the sparse rules, so no printed line leaves a section naming another spelling or un-registers a member (#761) #828 (ce0f2a5f), shape-doctor.py: a .gitmodules section name keeps git's own bytes and is handed back to git as them, so configured_url and kept_gitdir find a member named caf\xe9, its detached commit gets its keep line under any locale, and a non-ASCII name no longer ends --root in a UnicodeEncodeError traceback under LC_ALL=C PYTHONUTF8=0 (#732) #774 and shape-doctor.py and siblings.py: a .gitmodules section git ignores ([submodule ""], a .. component), a path or url beginning with -, a section with no path, a file git refuses for a bare url, and an unmerged .gitmodules name no mount url, NEXT line or clone, where git says No url found and exits 128 (#733) #775 merge clean.
• tests: four tests, directory_named and the file twin ask the one helper whether a name that is not UTF-8 can be made and skip on its two answers only, and three tests skip where an ASCII file name encoding cannot carry their text (#757) #806 and tests: three tests ask the helper whether a name can be made, make_file skips on EILSEQ or a UnicodeError only and the U+202E directory has no guard, so a full disk or a read-only tmp_path fails the test, not SKIPPED (#834) #845 conflict in tests/test_adopt_git_reads.py, as tests/test_doctor_legs_members_next.py and 24 other files: tests append what they tell git to the GIT_CONFIG_COUNT entries the environment holds, so a pasted line keeps the session's maintenance settings and the Windows job's two (#758) #818 does.
• family.py add: a members/ renamed with git mv is renamed back in two steps through a free name, after what the disk lacks under it is put back, and a printed --sparse says the git it needs, so no remedy fails as typed (#685) #767's bd6d2afe and family.py init: the creator is asked whether a file can hold it whichever of --created-by, GIT_AUTHOR_NAME or git config user.name gave it, git is asked about the first commit's author and committer before the remote exists, and a --family that is a Windows device name is refused, --dry-run alike (#688) #809's 98561405 conflict, and both squashes are in main (98f14976 and 42aed788).
• I re-ran the new scan over the 411 remote heads and 131 local branches. setup-project.py: a refused value names the check it failed, a project name over 95 characters is refused before step (4), and the prompt takes the spaces off a typed name, so Atlas no longer hides in step (4) (#777) #811 is the only live hit. Lane 3's tests.yml and release.yml: a push to a release/v<major>.<minor> branch that holds this change runs the suite as a push to main does, so a patch's commit there gets its push run with no dispatch, a fast-forward of that branch to main's commit no longer cancels main's run on it, and the release header says how a fix is back-ported (#850) #852 (900ffe70) was missing locally; I fetched it into my scratch clone, and its test files hold no count.

Sonar-style (item 5).

• The longest duplicated run in the added lines is 54 tokens over 7 lines: the lambda params in test_adopt_reencoded_files.py:426-432, which mirror test_adopt_nested_attributes.py:637-643. That is under 100 tokens and 10 lines.
• Counted by hand, remotes_fixture is about 7 and pastes_in about 9 with its nested visit. Nothing is over 15.

Could not verify

• The live PR body, since I did not run gh.
• The writer's own 29 mutants. I did not re-run them.
• Windows and macOS.
• SonarCloud itself.

Coordinator's note (lane 5): NIT 3's first point is done. The body's old merge section, measured at b9ad1187, is now headed as superseded by the fix round's merge checks. NITs 1 and 2 stay as they are: nothing is live, and the scan's limits are listed.

[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 5 days and 22 hours by commenting @sourcery-ai review. Upgrade to get a review now.