the family root's validate-family.py and bootstrap.py: with the work tree's .gitmodules gone, three remedies undo each other or refuse; a git mv'd checkout is never named; bootstrap says ok beside git's exit 128 #879

Goal: in the family holder's two pinned scripts, no printed remedy undoes another's write or refuses when typed as printed, the checkout a git mv moved is named, and bootstrap never says ok after git's update has stopped.

Filed by lane openRepoShape-5, from the five candidates (C1-C5) of rv876's independent review of #876 (#839), #876 issuecomment-6246717354. rv876 measured each one on git 2.43.0 and 2.47.1. Each is the same at #876's base 233a334e, so none is a regression of #876. path:line is at #876's head 77f8a6c8.

Lands after: #876 (fix/839-holder-bootstrap-texts), which changes these functions. Stack on it, or wait for its squash.

Lines

• C1. The wrong-case move's git config -f .gitmodules submodule.<variant>.path <row path> line (_move_context, templates/family-root/scripts/validate-family.py:689-704) commits a one-key file where the work tree's .gitmodules is gone.
  • Measured: OTHER recorded only at Members/IRSS, with HEAD's section giving that path; the work tree's .gitmodules deleted.
    • In members mode the move is a refusal printed alone. Typed as printed, HEAD's file keeps one key, and every other section is lost (P5e).
    • In --no-members mode, with MEMBER's url also gone, the move prints first. Its git add -- .gitmodules stages the one-key file. member-gitmodules-url-absent then prints git checkout -- .gitmodules, which restores that staged one-key file. Typed and committed, the gate reports member-gitmodules-absent and member-gitmodules-url-absent in every mode (P5b).
  • Direction: put the file back first, from where git reads it, as the family root's bootstrap.py and validate-family.py: the credential advice and a member's exit code say what happened, and the gate's git rm --cached, guarded mv and section lines type as printed where they failed (#839) #876's url-absent and absent findings do. And order the move's write with theirs, so that one put-back serves all.
• C2. member-gitlink-removal-staged prints git checkout HEAD -- .gitmodules where the work tree's file is gone (_put_back_lines, :2434-2435), and that undoes an earlier finding's write.
  • Measured: git rm --cached members/IRRS, the file gone, and OTHER's section gone from HEAD. The absent finding puts the file back and writes OTHER's section. The removal-staged line then puts HEAD's file back over it. The commit fails with nothing to commit, and member-gitmodules-absent returns (P5d).
  • Direction: read "section taken" from the file as git reads it, not from the empty work tree. Print no second put-back after a finding that already put the file back.
• C3. A committed value git refuses, plus a stray gitlink: the stray's git rm --cached exits 128.
  • Measured: the value finding's --unset-all lines leave .gitmodules unstaged. The stray's git rm --cached then refuses with please stage your changes to .gitmodules or stash them to proceed, and -f does not help (P5g).
  • Direction: print git add -- .gitmodules after the value finding's lines, before any git rm. Or order the stray's line first.
• C4. After the git mv state's lines are typed as printed, the checkout git mv moved stays at Members/<P>, untracked and never named.
  • Measured in all three git mv states: clean, a commit in the checkout, and an untracked file in it.
    • After typing, the gate passes in every mode and git status reads ?? Members/.
    • family.py add --dry-run then refuses with family-members-dir-variant and mv -- Members/IRRS <new name>.
    • After the update, git -C Members/<P> status reads the row's path, so the checkout's untracked file goes unseen.
    • Neither the STRAY finding nor member-gitlink-removal-staged says what that directory is, or that its commits are in the member's store. The wrong-case move's closing sentences say both.
  • Direction: name the checkout, with its git -C <path> status, before the update, as family.py add's closing sentences do after 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 and family.py add: a section astray beside its gitlink, a variant of a row's path under members/, a link or checkout in a move's way, an unmerged entry and a staged or unmerged .gitmodules are named first, so each remedy types clean (#840) #865.
• C5. Bootstrap says family bootstrap ok, with the credential advice, while git's update has exited 128 (lane 4's RV877, O1 on bootstrap.py: a .gitmodules blob with a line git cannot parse is read up to that line, as git reads it, and a parity test holds the value rules both family scripts write to one answer (#856) #877, issuecomment-6246457284).
  • The state: every member checked out, the work tree's .gitmodules gone, and a bad line such as branch = "unterminated committed in HEAD's file after the first section.
  • Measured: git's update exits 128 with No url found for submodule path 'members/B'. Bootstrap prints the advice, then family bootstrap ok, at 77f8a6c8 and at 233a334e alike.
  • Why:
    • unfetchable returns before init_stop when no row is left (templates/family-root/scripts/bootstrap.py:758-759).
    • git config --blob refuses the whole blob, while git's own reader takes the first section, so the sections come back empty and are not marked unreadable.
    • The ok follows bootstrap.py and validate-family.py: family bootstrap ok with no member checked out, an earlier row's member-uninitialized hiding a case-variant refusal, and member-gitlink-unnamed with no command under --no-members #648's rule that a run with every member checked out passes.
  • Direction: ask init_stop even when no row is left, and fail on a stop there.
  • Overlap: lane 3's bootstrap.py: a .gitmodules blob with a line git cannot parse is read up to that line, as git reads it, and a parity test holds the value rules both family scripts write to one answer (#856) #877 (bootstrap.py and validate-family.py (family root): a .gitmodules in the index or HEAD with one bad line is read as no sections where git reads those before it, so a no-credential run fails wrongly, and the value rules are written twice #856) changes the blob read. Settle the order with it.

Verify:

• Each state above, typed as printed, then the gate in every mode.
• A plain clone plus git submodule update --init --recursive exits 0 with every member checked out.
• For C5, bootstrap exits non-zero and names the stop.
• One mutant per line, each caught.
• Both scripts are pinned in every holder (FAMILY_COPIED_VERBATIM), so each holder takes one upstream-changed row per file.
• No Windows or macOS run, since CI is Linux only (ci: run the test suite on Linux only #597).

Overlaps: #876 (#839), lane 3's #877 (#856, stacked on #876), lane 4's #878 (stacked on #876) and lane 4's #870 all edit validate-family.py; #877 also edits bootstrap.py.

Parent: #459.

Milestone: v1.1.1. Each line needs the work tree's .gitmodules gone, a hand git mv, or a hand-committed bad line, and then the printed remedies typed in order. None is a regression, and the gate's verdicts in the plain states are unchanged.

🤖 Generated with Claude Code