P16 baseline cannot reach the DTC historical seal checks in a shallow checkout

Abierto
#408 0 comentarios 0 reacciones 0 asignados Ver en GitHub

Nadie ha tomado este issue todavía.

Evaluación

Dificultad
4/5
Tiempo estimado
3-5 días
Aptitud para principiantes
68/100
Tipo de issue
Error
Claridad
Bien especificado
Estado de actividad
Activo
Stack tecnológico
django, git, github-actions, python

Línea de trabajo

Read AGENTS.md, docs/PROCESS.md, docs/04-quality-gates.md, docs/03-playbooks.md, and .github/workflows/cross-repo-check.yml, then inspect C0.6 and the DTC DeploymentWorkflowContractTests guard. Verify the DTC checkout and run the specified single test in a disposable full-history donor checkout, followed by actionlint and plan checks. Done means the unchanged guard reaches its historical and final assertions, with raw evidence and existing semantics preserved.

Escrito por el modelo de indexación a partir del texto del issue.

Descripción

bug

Groomed package task: C0.6a Fetch DTC history for the P16 baseline guard.

The saved PR407 baseline at package head 01524f43/main c20e090 and DTC donor106c777 executed 4174
tests, then reported an ERROR in the unchanged operator-contract guard: historical git show exited
128 before its digest and final isolation assertions. Raw checkout receipts show depth1; native
DTC Django CI requests depth0. Local committed-object inspection proves the historical commit is
an ancestor and its workflow blob matches the unchanged expected digest. This supports a checkout
history setup diagnosis; the traceback does not preserve Git stderr, so no literal Git message is
claimed. PR407 P16 comparison green remains valid within its separate documented scope; it never
established that this baseline guard completed.

The following is the normative task draft for the implementing package PR. Current plan sources
are not modified by this grooming. No implementation, guard execution or fresh CI is claimed.

C0.6a Fetch DTC history for the P16 baseline guard

Repository: community-base. Depends on: C0.6. Freeze required: no.
Related issue: https://github.com/DataTalksClub/community-base/issues/408.

Goal: make the unchanged DTC historical workflow guard executable in the disposable P16 donor
checkout. Add full history to that checkout so its baseline can examine the historical workflow
blob and remaining assertions. Preserve site selection, baseline/linked comparison semantics,
all current seals and the exact existing link-artifact exceptions. D15 and D0.2 remain unchanged.

Read first

  • AGENTS.md, docs/PROCESS.md, docs/04-quality-gates.md and coding-standard.md.
  • docs/01-decisions.md D15 and docs/03-playbooks.md P16.
  • .github/workflows/cross-repo-check.yml and phase-0 C0.6.
  • DTC donor core/tests/test_deployment_workflow.py and .github/workflows/ci.yml through committed
    objects, initially donor 106c777d2cd39666afb5f0990c7301fe5ad418c3; record the actual donor used.
  • Preserved PR407 P16 raw baseline/linked evidence and independent qualification.

Ownership and scope

  • SWE owns only .github/workflows/cross-repo-check.yml, docs/03-playbooks.md's P16 checkout
    explanation, this phase-0 task entry and its generated STATUS row.
  • No site file, package runtime, UI, feature, dependency, migration, seal, assertion, guard,
    exclusion, failure-category normalization or CI trigger change belongs to this issue.
  • Only the DTC donor checkout needs fetch-depth: 0. Keep both package checkouts and the AISL
    checkout settings, donor ref/default-branch semantics and native site commands unchanged.

Steps

  1. Add this accepted task immediately after C0.6, run uv run python scripts/plan.py sync, and set
    its STATUS row to in-progress with issue408 until a PR exists. Confirm C0.6 is done and no
    existing row changes. One bounded bug PR may include this normative entry and implementation;
    it adds a task and fixes setup, without changing an existing decision or contradicting a step.
  2. Preserve the known failing receipt: PR407 baseline executed 4174 tests and the unchanged
    operator guard raised ERROR when git show 91490f0d3f172327a400c9edf5b441265890f897:.github/workflows/ci.yml
    exited128. Record its exact package/donor/command/category. Git stderr is unavailable in that
    traceback; do not invent its literal message or promote prior P16 comparison green to a fixed
    baseline. The earlier two linked seal FAILs are distinct link artifacts.
  3. Add fetch-depth: 0 to the existing Check out DataTalksClub/website step. Keep its repository,
    ref expression and sibling path. Add a concise P16 explanation that the full donor history
    supports unchanged historical Git assertions, matching native DTC Django checkout behavior.
  4. In a disposable donor checkout with reviewed baseline pin and full history, execute the actual
    unchanged guard at verbosity2 as a single non-skipped test. Record its Git source, module/test
    provenance, dependency pin and natural output/exit. It must reach the historical digest and
    final workflow-isolation assertions; a separate git show success alone is insufficient. Never
    link or alter a shared real site checkout. If the donor moves, inspect and bind its actual
    unchanged guard before interpreting results; do not rewrite assertions to make it pass.
  5. Run workflow lint and applicable package quality gates; publish only after the root releases
    the package publication window. Dispatch the sole designated OnCall for fresh exact-head
    required CI and both P16 consumers. Preserve raw baseline/linked logs, counts, skips,
    ERROR/FAIL categories, checkout refs, package install provenance and normalized failure sets.
    Independent review must establish the DTC historical baseline ERROR is resolved rather than
    hidden and qualify any linked seal failures by actual tracebacks and unchanged policy.

Verification

  • actionlint .github/workflows/cross-repo-check.yml -> no findings.
  • Committed diff -> DTC donor fetch-depth: 0 plus explanatory P16/plan tracking only; current
    triggers, ref expressions, commands, two exact FAIL exclusions and category handling unchanged.
  • Disposable baseline command, with the donor's documented test environment:
    DJANGO_SETTINGS_MODULE=website.settings.test PUBLIC_MEDIA_STORE_BACKEND=memory uv run python
    manage.py test core.tests.test_deployment_workflow.DeploymentWorkflowContractTests.test_gate_b_operator_contract_is_exact_and_workflow_isolated
    --verbosity 2 --parallel 1 -> exactly1 executed, zero skipped, passes naturally. Verify the
    imported unchanged method has no early return/skip and includes the historical digest and final
    assertions. Capture actual argv as one shell command; line wrapping above is documentation.
  • git rev-parse --is-shallow-repository -> false in the actual DTC full-history checkout;
    historical git show succeeds and SHA-256 matches the unchanged test constant. Initial donor
    expected digest is d6730d36c41866adcfd933ef733132e26ea67d292ddd0334caf42f9b2524850d.
  • Standard package gates applicable under docs/04-quality-gates.md section1 pass; record actual
    commands/counts and list genuinely unavailable evidence as Not run here, needs:.
  • uv run python scripts/plan.py check -> OK, no foreign tracking change.
  • Fresh exact-head sole OnCall -> package test/plan and both real consumer comparisons succeed.
    DTC raw baseline no longer has this historical Git ERROR; linked exceptions are limited to the
    existing two link-induced seal FAILs after independent qualification. Report all other donor
    failures accurately; no claim that a comparison pass makes a native site suite green.

Done when

  • The DTC P16 donor checkout fetches full history with existing selection/path semantics.
  • The actual unchanged baseline guard executes through historical and final assertions.
  • Fresh exact-head full baseline/linked consumer evidence is preserved and independently
    qualified without new exclusions, edited seals, skipped guard or category relabeling.
  • Applicable package/workflow/plan gates and required consumer CI pass; docs/tracking are updated.

Not run here, needs: implementation, actual guard execution, fresh exact-head required CI/P16,
merge and resulting-main observation. No site native-green, deployment, donor retirement, runtime
feature, tag/release or adoption acceptance is supplied by grooming or historical PR407 evidence.

Docs

  • docs/03-playbooks.md P16, docs/plan/phase-0.md and generated docs/plan/STATUS.md.
Lenguaje dominante
Python
Estrellas
0
Forks
0
Merge medio
1 h 5 min
PR fusionados (30 d)
3

Preparar el entorno

Este proyecto no incluye contenedor de desarrollo, Dockerfile ni guía de contribución, así que la configuración corre por tu cuenta: empieza por su README y consulta nuestra guía para la primera contribución para los pasos generales.

Primeros pasos

  1. Lee el issue completo y luego la guía de contribución del proyecto.
  2. Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
  3. Haz un fork del repositorio y trabaja en una rama.
  4. Abre un pull request que haga referencia al número del issue.

Más de DataTalksClub/community-base

Todos los issues de DataTalksClub/community-base

Issues similares

Más issues de Python

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.