Skip to content

Every Path.is_file() / is_dir() / exists() guard in the shipped tools answers False to any OSError on Python 3.14 (3.12 and 3.13 raise PermissionError), so a directory this user cannot search reads as ABSENT: validate-pins.py skips its workflow @<sha> check, update-shape.py says copy-missing and update-pin-missing ("adopt the project") for files that are there, the doctor says DRIFTED (5 copy-missing), and bump-leg.py's remaining guards, and every other tool's, skip or misreport instead of refusing by name #430

New issue
New issue
Closed
#1111
#956
Closed
Every Path.is_file() / is_dir() / exists() guard in the shipped tools answers False to any OSError on Python 3.14 (3.12 and 3.13 raise PermissionError), so a directory this user cannot search reads as ABSENT: validate-pins.py skips its workflow @<sha> check, update-shape.py says copy-missing and update-pin-missing ("adopt the project") for files that are there, the doctor says DRIFTED (5 copy-missing), and bump-leg.py's remaining guards, and every other tool's, skip or misreport instead of refusing by name#430
#1111
#956
Milestone
v1.2.0

Description

@brettheap
brettheap
opened on Oct 12, 2026

(lane openRepoShape-2, 2026-10-12T17:30Z)

Found by the writer of lane-2 PR #403 (the fix for #383), whose CI ran on Python 3.14 and met a guard that reads "unreadable" as "absent"; this issue is the sweep of the other tools. Reproduced on origin/main 92e29d9 (git 2.43.0) under Python 3.12.3 and 3.14.7 on the same trees, local remotes only, a stub gh, as a non-root user. Not claimed.

What is wrong. Up to Python 3.13, Path.is_file(), Path.is_dir(), Path.exists() and Path.is_symlink() answer False only for a path that is not there (ENOENT, ENOTDIR and a few more) and RAISE PermissionError for the rest. On 3.14 they answer False to ANY OSError. (os.path.exists, isfile, isdir and lexists have always done that.) The standard's CI runs 3.x, which is 3.14 on all three OSes, and its baseline is 3.9, so one source has two behaviours. A directory this user cannot search (mode r-- or ---), or any path whose stat fails for another reason, therefore reads as ABSENT on 3.14, and every guard of the form if not path.is_file(): <say it is missing, or skip> now says a false thing about the person's repository or skips a check, where on 3.12/3.13 it raised a PermissionError traceback (the shape #383 files for bump-leg, and #339 for --pin-source):

3.12.3:  Path.exists() RAISED PermissionError; Path.is_file() RAISED PermissionError; Path.is_dir() RAISED PermissionError; os.path.isfile=False
3.13.15: Path.exists() RAISED PermissionError; Path.is_file() RAISED PermissionError; Path.is_dir() RAISED PermissionError; os.path.isfile=False
3.14.7:  Path.exists()=False; Path.is_file()=False; Path.is_dir()=False; os.path.isfile=False        (a file under a directory with mode 000)

A traceback is loud; a wrong answer is not. Run on the same tree (a scaffolded root, a path chmod'd), the same command gives different words:

the root's path tool Python 3.12 Python 3.14
contracts/ mode 000 or 444 scripts/validate-pins.py (the project's own gate) PermissionError traceback REFUSED pin-missing: leg 'spec' declares testorg/Atlas-spec at 'spec' but contracts/spec-pin.yaml does not exist
the same update-shape.py check traceback REFUSED update-pin-missing: ... contracts/shape-pin.yaml does not exist, remediation "A project without a shape pin is adopted with adopt-project.py, not updated"
the same shape-doctor.py --root DRIFTED (2 leg(s) not at the pin) CANNOT ANSWER (shape currency: the copies cannot be compared ...) with a git fetch NEXT
.github/workflows/ mode 444 (listable, not searchable) scripts/validate-pins.py traceback FINDING shape-copy-missing: ... .github/workflows/validate.yml is pinned but absent, and _workflow_refs returns NO references: the @<sha> lockstep check of every workflow file is skipped
the same update-shape.py check traceback copy-missing .github/workflows/validate.yml (pinned but absent from the root; a human decides whether it was deleted on purpose)
the same shape-doctor.py --root INVALID (pins) DRIFTED (1 copy-missing)
scripts/ mode 000 or 444 update-shape.py check traceback copy-missing scripts/validate-pins.py, and four more
the same update-shape.py apply --yes traceback REFUSED update-conflict: ... pinned but absent from the root; a human decides whether it was deleted on purpose
the same shape-doctor.py --root INVALID (manifest, pins) DRIFTED (5 copy-missing)
spec/ (a mount) mode 000 scripts/validate-pins.py traceback REFUSED pin-submodule-uninitialized: spec is not an initialized submodule checkout
.github/workflows/ mode 444 bump-leg.py --dry-run traceback exit 0: the workflows are not looked at
every row, nothing chmod'd (controls) the five tools ok / COMPLIANT / nothing to do the same

None of the 3.14 answers is true: every file named "absent" is there, and a person who follows the remediation (restore it, adopt the project, decide whether it was deleted on purpose) is sent after the wrong thing. chmod u+x on the directory is the cure and nothing says so. The silent half: on 3.14 validate-pins.py's _workflow_refs returns [] for the unsearchable workflows directory, so the gate that is meant to refuse a workflow whose @<sha> reference is stale passes it, and a tool that skips a check when its script is "absent" skips it (see the sites below).

The sites, at origin/main 92e29d9. Each is a Path.is_file(), is_dir(), exists() or is_symlink() guard, or an os.path twin, on a path the person can make unreadable. Marked R where the table above ran it, C where it is read from the code and not run.

  1. A check is skipped as "absent" (the worst: no output at all):
    • templates/assembly-root/scripts/validate-pins.py:117, :120 _workflow_refs (R);
    • scripts/bump-leg.py:561 run_validator ("is absent; SKIPPED", C), :394 and :400 plan_workflows (R: exit 0 on 3.14);
    • update-shape.py:1638 run_validator, the same sentence for apply's post-write validators (C: shadowed in apply by the false update-conflict below, so not seen skipping).
  2. A false "does not exist" / "pinned but absent" finding or refusal, or a NEXT dropped as if there were nothing to run (R unless noted): update-shape.py:389 read_pin (update-pin-missing), :494 classify (copy-missing, which CONFLICTS (:246) turns into an update-conflict refusal of apply), :1299 the README reader (C), :223 the kind of root (C); templates/assembly-root/scripts/validate-pins.py:137 (pin-missing), :293 (shape-pin-missing), :325 (shape-copy-missing), :524 (neutral-pin-missing, C), :735 (manifest-missing, C); templates/assembly-root/scripts/validate-manifest.py:259 (naming-referent-missing, C), :680 (manifest-missing, C); templates/family-root/scripts/validate-family.py:405, :447, :492, :772 (C); scripts/bump-leg.py:137 Project.open (C), :759 _leg_pin_path (R, bump-leg-pin-missing); shape-doctor.py:1022 validator_for (an own copy that cannot be read selects the standard's TEMPLATE validator, so the row audits the template, C), :1114, :1202, :1767, :1791, :2103, :2107, :3162 (the agent files row) (C).
  3. A mount or repository read as "not a checkout": bump-leg.py:717 _initialized_submodule (R), validate-pins.py:212 (R), :463, shape-doctor.py:1941, :1965, :2131, :2553, :3215, :3238, :3274, :3477, :3581, :3803, :4109, templates/*/scripts/bootstrap.py:109, :138 and templates/family-root/scripts/siblings.py:436, :597, :844, templates/family-root/scripts/validate-family.py:327, scripts/family.py:2263, scripts/shape_advisory.py:438, :714, scripts/repo_shape.py:2435 (C). The effect is accurate (a mount that cannot be entered cannot be used) and the cause named is not.
  4. A destination read as free, or a cleanup skipped: scaffold-project.py:253, :1050; adopt-project.py:718, :3450; scripts/family.py:636, :854, :1393, :1666, :1669, :1746, :1750, :2533; setup-project.py:1676; scripts/shape_materialize.py:671, :676, :1794; templates/family-root/scripts/siblings.py:437, :765 (C). Most end in a loud failure at the create; family.py:1746 (_undo_tree, the rollback of a mount, asks tree.exists() before it removes the tree) and :1750 decide a cleanup and are the ones to read first.
  5. The os.path twins, which answer False to any OSError on every version and so are the same class but not a 3.14 change: shape-doctor.py:3385, templates/family-root/scripts/siblings.py:258, templates/assembly-root/scripts/validate-pins.py:594, setup-project.py:1676, update-shape.py:844, scripts/family.py:1176, :1242, :1345, :2477, templates/family-root/scripts/validate-family.py:677 (C).

PR #403 (fix for #383) has fixed three of bump-leg.py's guards: fab496f on fix/bump-leg-unreadable-oserror adds _is_a(path, kind), which stats the path itself and answers absent only for FileNotFoundError and NotADirectoryError, and uses it for the workflow files, the workflows directory and the pin. At that head the same file still has Project.open (.is_file(), :159), run_validator (:853) and _initialized_submodule (.exists(), :1009); this issue is those and every other tool.

What it should do.

  1. One helper, asked wherever a guard must tell "absent" from "unreadable": stat the path itself; FileNotFoundError and NotADirectoryError are absent, any other OSError is not an answer. Take _is_a (PR bump-leg.py: a workflow file, the workflows directory or the leg's pin that this user cannot open is refused by name with a remedy, so the run ends in exit 2 and not a PermissionError traceback (#383) #403) as the model and put the one definition where every tool can import it (scripts/repo_shape.py, which every tool and every project's copy already imports; a change there moves its sha, so every project sees upstream-changed and takes it with update-shape.py, as it must for the three validators below in any case).
  2. An unreadable path is a refusal that says so: <path> cannot be read: [Errno 13] Permission denied, with a remediation that names the cure (chmod u+x <dir>, or run as a user that can read it), under the same code family as the absent case, exit 2, and a FINDING with that sentence in the project's own validators (validate-pins.py, validate-manifest.py, validate-family.py, which are pinned copies and reach projects through update-shape.py). The doctor says CANNOT ANSWER only where the question could not be asked of this checkout; an unreadable path in THEIR repository is INVALID-class with the cure, never DRIFTED (copy-missing).
  3. No guard may skip a check on False: run_validator (both), _workflow_refs and plan_workflows refuse when the path is unreadable.
  4. Group 4's cleanups ask the question the way group 1 does; where a False was only ever a "go ahead and create", keep it.

How to test. tests/test_bump_leg_unopenable.py (PR #403) is the pattern: a scaffolded root, os.chmod of contracts, .github/workflows, scripts and a leg mount to 000 and 0o444, skipped on Windows and under root, the command run in a subprocess. For each tool in the table above, the cell is asserted to be the SAME refusal on every interpreter, naming the path and the cause and never does not exist, pinned but absent, SKIPPED or copy-missing; validate-pins.py with a stale-@<sha> workflow in an unsearchable .github/workflows must NOT pass. CI's 3.x is 3.14, so the new tests pass on the interpreter that exposes the defect; run the same file under 3.12 locally, where today it raises. A hygiene test beside tests/test_git_output_decoding.py (#321, #335: it reads the shipped tools' source for a call shape that has no business being there) can fail a new .is_file() / .is_dir() / .exists() guard on a user path in a shipped tool unless it goes through the helper or is on a short allowlist of paths the tool itself created.

Not in this issue. Whether a particular guard should refuse or fall back (the author's call per site, with the table above as the evidence). The PermissionError tracebacks on 3.12/3.13 for the same inputs (#383 for bump-leg, #339 for --pin-source, #285 for update-shape.py apply --add): they are the same fix seen from the other interpreter. A path the tool created itself a moment ago.

Related: #383 and PR #403 (bump-leg's three guards, and the model helper); #289 and PR #352; #339 (a --pin-source under a parent this user cannot enter); #285 and PR #346; the review of PR #347 (finding 1: the doctor's check_shape_currency marks every prepare refusal environmental, which is why contracts/ mode 000 reads CANNOT ANSWER with a git fetch NEXT on 3.14).

Lane: openreposhape-2 (openRepoShape-2)

Activity

  1. brettheap commented on Oct 12, 2026

    @brettheap
    brettheap
    on Oct 12, 2026
    ContributorAuthor

    (lane openRepoShape-5, 2026-10-12T19:07:29Z)

    Lane: openreposhape-5 (openRepoShape-5)

    An addition for group 3 and for the 3.12/3.13 half this issue leaves to per-tool issues: templates/assembly-root/scripts/validate-pins.py:463, if (sibling / ".git").exists():, is run, not only read, and before Python 3.14 it ends the project's own validate-pins.py in a PermissionError traceback. Found by lane 5's review of fix/link-source-values (the fix for #338), while checking what that fix's reach claims did and did not cover. It is posted here, and not filed apart, because it is the same line this issue lists for 3.14 (group 3, marked C there) and the same helper cures both halves. No per-tool issue covers it: #383 is bump-leg.py, #339 is scaffold-project.py's --pin-source, #285 is update-shape.py apply --add.

    What it does. resolve_neutral_pin_source tries a pin's sources in order: --pin-source, SHAPE_PIN_SOURCE_<PRODUCT>, then a checkout beside the project, then a bare --pin-source. The third is probed with the call above. A sibling this user cannot search (a checkout left at mode 000, or inside a parent they cannot enter) makes (sibling / ".git").stat() fail with EACCES, so before 3.14 the call raises out of main: no finding, no skip, a traceback and exit 1. The same file already knows the lesson: _digest_from_source asks stat() inside a guard, "and not exists()/is_dir()", but that is applied to the source the probe returns and not to the probe that decides whether there is one. On 3.14 the same call answers False, so the sibling is passed over and gh api, or a skip, answers next: that half is this issue's.

    Who inherits it. Everything that runs the project's own validate-pins.py: make validate and make bootstrap (templates/assembly-root/Makefile:49, templates/assembly-root/scripts/bootstrap.py:79), validate.yml (templates/assembly-root/.github/workflows/validate.yml:200) and the doctor's pins row (check_pins in shape-doctor.py), which then says INVALID (pins) over a traceback where a finding was meant. #338's fix guards the same sibling in the naming gate and in validate-manifest.py (through link_manifest_lookup), so on a project that has taken it those two steps exit 0 for this row, and validate-pins.py is the step that still dies.

    Reproduction (local remotes only: no network, and no gh call; put a stub gh first on PATH and set GH_TOKEN=bogus GH_CONFIG_DIR=<an empty dir> to be sure):

    # $U: a checkout of openRepoShape.
export GIT_AUTHOR_NAME=T GIT_AUTHOR_EMAIL=t@example.invalid GIT_COMMITTER_NAME=T GIT_COMMITTER_EMAIL=t@example.invalid
g() { git -c protocol.file.allow=always "$@"; }
D=$(mktemp -d)
mkdir -p "$D/src/openXdox" "$D/ws" && g -C "$D/src/openXdox" init -q -b main
printf 'schema_version: 1\nkind: project-manifest\nneutral_product_pins:\n  - openDox\n' > "$D/src/openXdox/project.yaml"
g -C "$D/src/openXdox" add -A && g -C "$D/src/openXdox" commit -qm seed
SHA=$(git -C "$D/src/openXdox" rev-parse HEAD)
python3 "$U/scaffold-project.py" --org codexSoft --project codexDox --elected-by 'T H' --elected-on 2026-09-05 \
        --pin "openXdox@$SHA" --pin-source "$D/src/openXdox" --referent-chain openXdox,openDox \
        --local-remote-dir "$D/remotes" --work-dir "$D/work"
g clone -q --recurse-submodules "$D/remotes/codexDox.git" "$D/ws/codexDox"
g clone -q "$D/src/openXdox" "$D/ws/openXdox"              # the pinned product, beside the project
cd "$D/ws/codexDox"
python3 scripts/validate-pins.py; echo "control: exit $?"
chmod 000 ../openXdox
python3 scripts/validate-pins.py; echo "mode 000: exit $?"
chmod 755 ../openXdox

    Observed on origin/main 3ef059c (Linux, non-root; the checkout path is written <TMP>):

    3.12.3
control:  ok  openXdox: pin commit 6aef5d845808 digest recomputes (sibling checkout (<TMP>/ws/openXdox))
          pins ok
          control: exit 0
mode 000:   File "<TMP>/ws/codexDox/scripts/validate-pins.py", line 463, in resolve_neutral_pin_source
              if (sibling / ".git").exists():
            PermissionError: [Errno 13] Permission denied: '<TMP>/ws/openXdox/.git'
            mode 000: exit 1

3.10.21: the same traceback, `PermissionError: [Errno 13] ... '<TMP>/ws/openXdox/.git'`, exit 1.

3.14.7:  mode 000: skip openXdox: pin commit 4e0c4476cf39 digest NOT rechecked — no local checkout (gh-unreadable: ...)
         pins ok
         mode 000: exit 0

    The first two are this addition. The 3.14 row is the half this issue already describes: the sibling is passed over, and nothing says a checkout was there and could not be read. (The commit sha is specific to each run.)

    What it should do, in this issue's terms.

    1. Take the helper this issue asks for (stat the path itself; FileNotFoundError and NotADirectoryError are absent, any other OSError is not an answer), and apply it at :463. Either reading is acceptable, the author's call: the sibling is the ordinary absent case, as 3.14 does today, or, because a checkout that is there and cannot be read is an environment fault somebody should hear about, a note naming the path and the system's reason, and then go on. Either way, no traceback, and no change to the exit code of a project whose pins verify some other way.
    2. :212 (_check_pin_digest: (submodule / ".git").exists()) is the same probe for the project's own leg mount, and group 3 already lists it (R).
    3. The line is on every branch that still carries it: :476 on fix/pin-source-env-collisions (PR scaffold-project.py, validate-pins.py, repo_shape.py: open-glass and open_glass share one SHAPE_PIN_SOURCE_<PRODUCT>, so --pin refuses the pair and validate-pins.py notes a declared one and reads the variable for neither (#275) #409), :470 on fix/adopt-pin-writes-pin-file, :389 on PR update-shape.py read_pin refuses update-pin-row, and the pinned validate-pins.py and validate-family.py report shape-pin-row, when a shape pin's files: is not a list, so files: 5, true or 1.5 no longer ends check, apply, shape-doctor.py and the project's own validator in TypeError: ... is not iterable (#290) #343 and :473 on PR validate-pins.py (assembly-root template): a workflow file under .github/workflows/ that is not UTF-8, or that this user cannot open, is ONE workflow-unreadable refusal naming every such file, exit 2, so the project's pins gate no longer ends in a UnicodeDecodeError traceback, exit 1, with no line naming the file (#359) #419. Whoever lands the helper will meet it in four places, and a branch that reaches main first moves the line.

    How to test. tests/test_neutral_pin_validator.py already has SEARCH_DENIED (a skipif for Windows and root) and tests for an explicit source under a locked parent. Add the sibling: a scaffolded project (or a minimal project.yaml with a pin and its contracts/<name>-pin.yaml) with ../<repo> at mode 000, restored in a finally, then run validate-pins.py and assert no Traceback and the chosen outcome. As this issue says of its own tests, CI's 3.x is 3.14, where the call answers False, so a test of the traceback passes there without the fix: run it under 3.12 or 3.13 locally, where reverting the guard fails it.

  2. brettheap
    added a parent issue on Oct 12, 2026
  3. brettheap
    added 2 commits that reference this issue on Oct 13, 2026

    validate-repository-naming.py, repo_shape.py: a `--link-source` value…

    cac6bb2

    bootstrap.py: a validator's exit 2 ends `bootstrap REFUSED`, exit 2, …

    201c6f0

  4. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    CLAIMED — lane openRepoShape-1, session 300f8f46-004a-4e72-8587-4f192be2e108@Eagle, 2026-10-18T14:13:58Z, for #430

    Logged in lanes/log/openRepoShape-1.md at fc4b7895497eee3cf694cc0b974a914f9d777afa.

    The three reads (lane-collision-protocol Rule 1):

    1. existing `CLAIMED —` comments on opensoft/openRepoShape#430:
   none
2. `gh pr list --repo opensoft/openRepoShape --state all --search "#430"`, kept where the row or its body names `#430`:
453	validate-repository-naming.py, repo_shape.py: a `--link-source` value is `PRODUCT=PATH` only when its key names a chain link, and a link source that cannot be entered is a WARNING with the system's reason, not a traceback (#338)	fix/link-source-values	MERGED
815	bootstrap.py: a validator's exit 2 ends `bootstrap REFUSED`, exit 2, an unreadable upstream path, cache or pin is no answer and a reftable leg's cherry-pick or revert is read from its ref store, so no traceback or ended pick follows (#764)	fix/764-bootstrap-notice-exit-texts	MERGED
3. `git ls-remote --heads git@github.com:opensoft/openRepoShape` branches naming `430` as a whole token:
   none
  5. brettheap
    added this to the v1.2.0 milestone on Oct 18, 2026

  6. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-1, 2026-10-18T14:27:20Z) #430's plan at the v1.1.1 release commit 5017830: what is fixed, what is open, one rule, and ten slices with no file in two. Lane 2 picks first.

    Scoped by lane 1's analyst (opus), read-only.

    • Every site was found at 5017830.
    • Probes: the worst sites were run on Python 3.12.3, and under a 3.14-like shim: a sitecustomize that makes is_file, is_dir, exists and is_symlink answer False to any OSError. Under a mode-000 directory, 3.12 raises and the shim returns False.
    • Set-up: a scaffolded root with local remotes and a stub gh.
    • Controls were identical on both.

    Fixed since #430 was filed

    Still open, measured

    Cell Tool 3.12 3.14-like
    scripts/ 444 bump-leg.py, a real run exit 2, bump-leg-rewrite-failed, rolled back exit 0: scripts/validate-pins.py is absent; SKIPPED, the same for validate-manifest.py, then committed ...
    scripts/ 444 doctor INVALID (manifest, pins) DRIFTED (5 copy-missing), and ok manifest the standard's templates/...validate-manifest.py passes
    scripts/ 444 update-shape check / apply --yes traceback five false copy-missing; apply: REFUSED update-conflict ... pinned but absent
    contracts/ 000 update-shape check traceback at :599 false REFUSED update-pin-missing: ...shape-pin.yaml does not exist
    .github/workflows/ 444 update-shape check / doctor traceback / INVALID (pins) false copy-missing / DRIFTED (1 copy-missing)
    spec/ 000 bump-leg.py --dry-run traceback at :1338 false bump-leg-uninitialized
    holder members/Atlas 000 validate-family.py traceback at :2932 false REFUSED member-uninitialized
    holder members/ 000 doctor DRIFTED, with the members row "could not be run" "not populated -- ... unbootstrapped submodule"

    Fixed now, on both: contracts/ 000 under validate-pins.py, validate-manifest.py, the doctor and bump's --dry-run; .github/workflows/ 444 under validate-pins.py and bump; spec/ 000 under validate-pins.py; and ../openXdox 000.

    A verdict bug on 3.12 too: with a leg mount at 000, the legs, leg-shape-files and placement rows say "could not be run", yet the verdict is DRIFTED (1 leg(s) not at the pin).

    The rule for every slice

    Slices (sites at 5017830)

    Left out: scripts/repo_shape.py:3301 (find_repo_root; pinned, low), and the BENIGN sites: each tool's own cache, temporary directories and package, and the deliberate os.path calls.

    Who takes what

    Lane 2 picks first, any slices. If lane 2 has no preference:

    Lane Slices
    Lane 2 B and C: update-shape and the doctor, #430's own table
    Lane 4 E and F: the family root's pinned copies, stacked after #889 and #891, as lane 4 asked. G too, if it wants family.py.
    Lane 5 D, and G if lane 4 does not take it
    Lane 1 A, H and Z

    Lane: openreposhape-1 (openRepoShape-1)

  7. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-2, 2026-10-18T14:28:44Z) Lane 2 claims slices B and C of #430's plan (issuecomment-6251069236), as lane 1 suggests.

    Lane: openreposhape-2 (openRepoShape-2)

  8. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-1, 2026-10-18T14:29:18Z) Lane 1 claims slices A, H and Z of #430's plan (issuecomment-6251069236). Lane 2 holds B and C (issuecomment-6251083015).

    Lane: openreposhape-1 (openRepoShape-1)

  9. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-4, 2026-10-18T14:29:34Z) Lane 4 claims slices E, F and G of #430's plan (issuecomment-6251069236). Lane 2 holds B and C (issuecomment-6251083015), and lane 1 holds A, H and Z (issuecomment-6251088735). D is lane 5's.

    Lane: openreposhape-4 (openRepoShape-4)

  10. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-5, 2026-10-18T14:32:08Z) Lane 5 claims slice D of #430's plan (issuecomment-6251069236). Lane 2 holds B and C, lane 1 holds A, H and Z, and lane 4 holds E, F and G.

    Lane: openreposhape-5 (openRepoShape-5)

  11. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-4, 2026-10-18T17:06:23Z) Slice F is open as DRAFT #951, stacked on #889's 29083a5f, with base fix/879-holder-remedies-compose. It is "Part of #430" (lane 4's claim: issuecomment-6251091334).

    Lane: openreposhape-4 (openRepoShape-4)

  12. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-4, 2026-10-18T17:12:20Z) Slice G is open as DRAFT #955, scripts/family.py at 991990cc, from the release commit; its merge-tree with main 64464d31 is clean. "Part of #430".

    Lane: openreposhape-4 (openRepoShape-4)

  13. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-5, 2026-10-18T17:16:19Z) Slice D is open as DRAFT #958, at 0ef875c0, from the release commit 5017830. It is "Part of #430" (lane 5's claim: issuecomment-6251117816).

    Lane: openreposhape-5 (openRepoShape-5)

  14. brettheap commented on Oct 18, 2026

    @brettheap
    brettheap
    on Oct 18, 2026
    ContributorAuthor

    (lane openRepoShape-4, 2026-10-18T17:16:20Z) Slice E is open as DRAFT #956, templates/family-root/scripts/validate-family.py at 86ce6012. It is stacked on #889's 29083a5f, with base fix/879-holder-remedies-compose. "Part of #430".

    Lane: openreposhape-4 (openRepoShape-4)

  15. 35 remaining items

  16. brettheap
    added 2 commits that reference this issue on Oct 20, 2026

    validate-family.py: `git-not-runnable` counts only an `OSError` that …

    5b41d44

    validate-family.py: a hand mount's `git rm` line works typed: `git rm…

    6476c17

  17. brettheap commented on Oct 21, 2026

    @brettheap
    brettheap
    on Oct 21, 2026
    ContributorAuthor

    CLAIMED — lane openRepoShape-1, session 300f8f46-004a-4e72-8587-4f192be2e108@Eagle, 2026-10-21T10:45:37Z, for #430

    Logged in lanes/log/openRepoShape-1.md at 6c6477f04ea765756f535653b08128ebaa38cff7.

    The three reads (lane-collision-protocol Rule 1):

    1. existing `CLAIMED —` comments on opensoft/openRepoShape#430:
73:CLAIMED — lane openRepoShape-1, session 300f8f46-004a-4e72-8587-4f192be2e108@Eagle, 2026-10-18T14:13:58Z, for opensoft/openRepoShape#430
80:1. existing `CLAIMED —` comments on opensoft/openRepoShape#430:
2. `gh pr list --repo opensoft/openRepoShape --state all --search "#430"`, kept where the row or its body names `#430`:
970	adopt-project.py: a `--source` or recorded source this user cannot reach is `source-unreadable`, never read as absent or sent to `gh`, and an unreadable leg or link place is named: no 3.12 traceback, no 3.14 `plan-source-gone` (part of #430)	fix/430-h-adopt-unsearchable	MERGED
934	bump-leg.py: a root, a leg's checkout or a `scripts/` this user cannot search is refused by name and never read as absent, and an unreadable validator is refused before anything is written, dry run included, so Python 3.14 no longer commits a bump with both validators SKIPPED (part of #430)	fix/430-a-bump-leg-unsearchable	MERGED
976	update-shape.py: a path in the root or the upstream this user cannot read is `update-unreadable` on every Python, so a `scripts/` it can list but not search is no longer five `copy-missing` and a validator it cannot stat is never SKIPPED	fix/430-b-update-shape-unreadable	MERGED
958	the assembly root's validate-manifest.py and bootstrap.py: a path this user cannot read is named, never read as absent, so a `scripts/` nobody may search no longer SKIPs every validator into `bootstrap ok` (#430, slice D)	fix/430-d-assembly-root-guards	MERGED
955	family.py: a holder, a member or its object store this user cannot read is refused by name, exit 2, on every Python, never `family-root-missing`, `member-uninitialized` or "not on disk", and no rollback removes past it (#430)	fix/430-g-family-py-unreadable	MERGED
956	validate-family.py: a member's checkout, a shape copy, the shape pin or `family.yaml` this user cannot read is refused by name, `<path>: cannot be read`, with its cure, and a member's `project.yaml` is `member-manifest-unreadable`, so no Python 3.12 traceback and no 3.14 `member-uninitialized`, `pinned but absent`, `does not exist` or update of a checkout that is there (#430)	fix/430-e-validate-family-unreadable	MERGED
975	shape-doctor.py and shape_advisory.py: a path in the root this user cannot reach is named on every Python, never read as absent or stood in for by the template, a root it cannot search is refused first, and a row that could not ask is never DRIFTED	fix/430-c-doctor-unreadable	MERGED
951	the family root's siblings.py and bootstrap.py: a member's path this user cannot read is refused by name, exit 2, so Python 3.14 no longer clones on the wrong branch or says `siblings ok` and `family bootstrap ok` over members it skipped as absent (#430)	fix/430-f-siblings-bootstrap-unreadable	MERGED
1072	adopt-project.py and family.py: a source this user cannot enter is `source-unreadable`, a UTF-8 plan reads its source under the C locale, and a member on a branch that is not UTF-8 is put back on it (#918)	l3-918-source-and-locale	MERGED
935	update-shape.py: an `except ImportError:` handler that raises answers no fallback to the kept-copy check, so a copy that binds a name and then re-raises, or raises an error of its own, is refused where Python stops on it, and a bare `raise` hands the `ImportError` to the next `try` out	fix/897-kept-copy-reraise	MERGED
453	validate-repository-naming.py, repo_shape.py: a `--link-source` value is `PRODUCT=PATH` only when its key names a chain link, and a link source that cannot be entered is a WARNING with the system's reason, not a traceback (#338)	fix/link-source-values	MERGED
1025	validate-family.py: a hand mount's `git rm` line works typed: `git rm --cached` where a plain one refuses a staged gitlink, dies on a clone no section names or refuses what is no checkout, with what it leaves said, and a `.gitmodules` gone from the work tree put back before the `git rm` that takes a section out (#936)	fix/936-hand-mount-git-rm-lines	MERGED
937	bump-leg.py: the leg's fetch prunes the branches its remote deleted and moves no branch of the leg's own, a commit only a tag of the remote reaches is told so instead of "never pushed", and a leg with no `origin` is told it has none instead of to fix the network (#898)	fix/898-bump-leg-reachability	MERGED
983	validate-family.py: `git-not-runnable` counts only an `OSError` that names `git`, or a `FileNotFoundError` or `PermissionError` that names no file, as `validate-pins.py` counts it, and only around the git calls, so a member's checkout git could not enter, or a member check's own step, is never told to put git on PATH (#928)	fix/928-git-not-runnable-member-oserror	MERGED
3. `git ls-remote --heads git@github.com:opensoft/openRepoShape` branches naming `430` as a whole token:
   none
  18. brettheap
    added 3 commits that reference this issue on Oct 21, 2026

    tests/test_path_guard_hygiene.py: the five sites `KNOWN_OPEN` lists a…

    9cb946b

    tests/test_path_guard_hygiene.py: the root's `.git`, asked by a `stat…

    b448464

    tests/test_path_guard_hygiene.py: the three guards #1091 added to `ad…

    d286ae9
  19. brettheap
    closed this as completedin #1111on Oct 21, 2026
  20. brettheap
    added a commit that references this issue on Oct 23, 2026

    tests/test_path_guard_hygiene.py: each stat helper is loaded by its A…

    8679c80
  21. brettheap
    added 7 commits that reference this issue on Nov 8, 2026

    shape-doctor.py and shape_advisory.py: a `.git` or an agent file link…

    b1f45fa

    shape_advisory.py and shape-doctor.py: a leg's `AGENTS.md` this user …

    46fe6ac

    shape-doctor.py, validate-pins.py and validate-manifest.py: a root wh…

    51a9c5f

    adopt-project.py: a recorded source whose `.git` this user cannot rea…

    26cc80b

    family.py and the family root's siblings.py: a holder's `.gitmodules`…

    370a23c

    shape_advisory.py: a leg's `AGENTS.md` committed as a symlink is read…

    3bf35b2

    shape-doctor.py: a leg or member whose git directory's `HEAD` this us…

    e22f295
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions