Commit Graph
261 Commits
Author SHA1 Message Date
kevinveenbirkenbachandClaude Opus 4.7 525699a9ab test(lint/cache-read): extend scope to utils, cli, plugins, library
Modules imported by tests run inside the same pytest process as the
tests themselves. Raw ``path.read_text(...)`` calls in those modules
bypass the cache the tests just populated. ``roles/`` stays excluded
(deployed scripts have no ``utils/`` on PYTHONPATH); ``utils/cache/``
stays excluded (cache implementation itself).

Five conversions to ``utils.cache.files.read_text``:
- ``utils/handler/yaml.py`` (YAML loader, read-only)
- ``utils/annotations/summarize.py`` (CI log scanner, with
  ``UnicodeDecodeError`` fallback replacing ``errors="replace"``)
- ``cli/meta/roles/applications/resolution/combined/__main__.py``
  (single-shot CLI JSON input)
- ``cli/core/help.py`` (read_folder_title + read_folder_description;
  same README is read many times across CLI help iteration)

Eight ``# nocheck: cache-read`` markers, each annotated inline:
- ``utils/claude/settings_sort.py`` (read then rewrite sorted)
- ``utils/docker/version_updater.py`` (read then rewrite config)
- ``cli/administration/deploy/development/network.py`` (tests mock
  ``Path.read_text`` directly; caching would bypass the mock)
- ``cli/administration/deploy/development/log_utils.py`` (append flow)
- ``cli/contributing/changelog/archive/changelog_md.py``,
  ``__main__.py``, ``package_mirror.py`` (file mutated in same flow)

Two callers refactored to variable extraction so ruff's 88-char wrap
keeps the ``# nocheck`` marker on the ``read_text(...)`` AST span line
(``cli/.../network.py`` and ``package_mirror.py:105``).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-15 14:23:32 +02:00
kevinveenbirkenbachandClaude Opus 4.7 bd1a3a69ad fix(deploy/development/inventory/planner): split per-round include from union purge_set
`3c1d5d821` solved the cross-round backbone-abriss problem by making
`include` equal the union of every variant's closure for every round.
Side-effect: every matrix round then ALSO deployed the full union, so
`08_deploy_performance.sh` saw `dashboard.infinito.example` still
answering HTTP 200 after a matomo-only `deploy-fresh-purged-apps` —
the deploy step was no longer matomo-only at all.

Split the two concerns instead:

- `include` (PlanEntry[3]) goes back to the per-round variant-closure.
  Both `init`/`builder` (`--include` for the inventory bake) and
  `deploy` (`--id` for ansible) consume it. Each round therefore
  deploys ONLY its own variant-closure again — the pre-`3c1d5d821`
  semantics.
- `purge_set` (PlanEntry[4], new) carries the union across all rounds.
  `deploy.py` uses it for the inter-round wipe instead of the previous
  round's include. Result: between rounds the WHOLE matrix footprint
  is purged, not just the previous round's variant slice — a variant
  pinning `services.<X>.shared: false` cannot leave its bundled
  provider on the host for a sibling round that wires the same service
  shared. The backbone-abriss `3c1d5d821` was hunting stays fixed:
  whichever round needs the backbone re-deploys it immediately after
  the wipe (purge runs at the start of each round, not the end).

`previous_round_include` tracking is gone. The first iteration skips
the purge (nothing to wipe yet), which also covers `total_rounds == 1`
and `--variant N` pinned single-folder modes cleanly.

Unrelated, but staged together because it is part of the same
investigation:

- `roles/dev-nix/templates/nix.conf.j2`: enable `keep-derivations` /
  `keep-outputs`, so `nix-collect-garbage -d` (run by the aggressive
  prune step) does not drop the build closure of pkgmgr's flake; this
  was making every deploy re-fetch the build inputs from
  cache.nixos.org and overshoot pkgmgr's update timeout.
- `roles/pkgmgr/tasks/01_core.yml`: raise the pkgmgr update timeout
  from 600s to 1800s and lower the retry count from 6 to 3, so a slow
  fetch on a single try no longer pushes total wallclock past three
  hours.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-15 00:44:47 +02:00
kevinveenbirkenbach 3c1d5d8214 fix(deploy/development/inventory/planner): include union of all variant closures per round
A primary app's variant MAY pin `services.<X>.shared: false` to test
the bundled-provider path. That removes <X>'s shared provider from
that variant's closure; the inter-round purge then tears down a
backbone the next variant still expects, leaving consumers in the
following round on an empty dashboard (Index of /), a missing
postgres (5432 Connection refused), or a torn-down mailu.

Pre-compute every variant's closure once, then union them and use the
union as the include for every round. Variants keep changing HOW each
consumer is wired (baked host_vars per round); the deployed app set
stays the superset so purge → redeploy stays round-independent.
2026-05-14 17:12:56 +02:00
kevinveenbirkenbachandClaude Opus 4.7 a2b5269244 refactor(inventory): drop unused variant-only matrix mode
Commit 16f323319 added a strict variant-only include rule + conflict
detector in `inventory/variants.py`, gated behind the presence of any
primary's `meta/variants.yml`. In practice the rule never paid off:

* The CI failure that motivated it (matomo's matrix transition leaving
  a stale dashboard nginx vhost -> 502) is already addressed by the
  nginx-purge primitive shipped in commit bfd04e4a1.
* The existing `CombinedResolver` path already implements the "any
  role says true or `in group_names` -> pull" semantics via
  `utils.roles.applications.services.registry.is_explicit_truth`, so
  every needed service still gets pulled and services that all roles
  mark False stay out of the include.
* The strict literal-equality conflict detector was a false-positive
  factory: 79/80 variant-equipped primaries failed it against the
  current repo state because roles legitimately disagree on `enabled`
  flags for shared services.

This commit removes the entire variant-only branch and restores the
single-path planner. Round-index clamping for non-primary deps is left
exactly as before 16f323319 -- only the additions are reverted.

The file-split refactor from 16f323319 is intentionally kept; only the
`variants.py` submodule, the corresponding `test_variants.py`, and the
variant-only gate test class in `test_planner.py` are removed. The
`__init__.py` re-exports + `__all__` are pruned to match.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-13 04:25:07 +02:00
kevinveenbirkenbachandClaude Opus 4.7 f102c229f0 refactor(inventory): split monolith into SRP submodules + variant-only mode
Two changes ship together because the variant-only matrix-deploy logic
was written into the legacy 697-line inventory.py and then re-split:

* Source package (max 240 lines per submodule, well under the 300 budget):
  - spec.py            DevInventorySpec + PlanEntry data types
  - payload.py         _resolve_variant_payloads + _bake_overrides
  - legacy_resolver.py CombinedResolver fallback (default-mode path)
  - variants.py        strict variant-only include resolution
  - planner.py         plan_dev_inventory_matrix + filter_plan_to_variant
  - builder.py         build_dev_inventory + build_dev_inventory_matrix
  - __init__.py        public-API re-exports + back-compat shims

* Test package mirrors the source 1:1 with a shared _fixtures.make_spec
  to avoid duplicating the DevInventorySpec defaults across files.

* New variant-only matrix mode (gated on any primary shipping
  meta/variants.yml): only services whose merged variant payload pins
  both enabled and shared to literal True become inventory groups;
  Jinja flags never qualify (group_names does not exist yet at build
  time); run_after only orders the include, never expands it;
  cross-primary disagreement on a (service-key, round) raises ValueError.
  Default mode (no variants.yml on any primary) keeps the legacy
  CombinedResolver path untouched.

* Docs in design/{inventory,variants}.md updated to point at the new
  per-submodule paths; test_nginx_vhosts.py's Test classes now also
  inherit directly from unittest.TestCase so the AST-level lint in
  test_all_test_files_have_tests.py picks them up (transitive base via
  NginxVhostsTestBase alone is not enough).

Coverage: 458 cli unit tests, 349 utils unit tests, 99 dev-deploy tests,
all green; ruff check clean.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-13 02:17:26 +02:00
kevinveenbirkenbachandClaude Opus 4.7 576fad84ff refactor(images): move role images from defaults/main.yml to meta/services.yml
Consolidates image declarations under a single schema: every role's
`<svc>.{image,version}` lives in `meta/services.yml` alongside compose
service entries. The `image` lookup now reads from disk per role,
`iter_role_images` collapses to a single source, and the mirror chain
(resolver + apply_mirror_overrides) drops the dead `images:` branch.

`test-e2e-playwright` resolves `TEST_E2E_PLAYWRIGHT_IMAGE` directly via
`lookup('image', 'playwright', 'ref')` in vars/main.yml; the runtime
set_fact + override-empty-string dance is gone.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-10 22:06:32 +02:00
kevinveenbirkenbachandClaude Opus 4.7 c55ad69c41 refactor: drop requirement-file references from code, add doc rule + lint
Strip every `req <NNN>`, `req-<NNN>`, `Requirement <NNN>`,
`requirement-<NNN>`, and `docs/requirements/<NNN>-...md` mention from
non-Markdown source files. Comments anchored to requirement numbers
rot when requirements are renumbered or rewritten and conflate
present-state behaviour (which the comment should describe) with the
audit trail of how the code got there (which belongs in commit
messages, PR descriptions, and the requirement file itself).

* Document the rule in `docs/contributing/documentation.md` under
  "Comments". Code MUST describe the *why* in present-state language;
  contract enforcement lives in tests, not folklore.
* Clean 254 source files across `roles/`, `tests/`, `plugins/`,
  `cli/`, `scripts/`, `utils/`, `group_vars/`, and the repo-root
  `eslint.config.js`. Pure comment / docstring / log-message edits;
  no behavioural change.
* Add `tests/lint/repository/test_no_req_references_in_code.py` to
  enforce the rule going forward across `.js`, `.py`, `.yml`,
  `.yaml`, `.j2`, and `.sh` files. The legacy
  `test_no_stale_requirement_refs.py` is exempt as the only file
  whose purpose is to detect these very tokens.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-10 16:25:22 +02:00
kevinveenbirkenbachandClaude Opus 4.7 85a6961a13 fix(cli/inventory-validate): correct PROJECT_ROOT depth after move
The script lives at cli/administration/inventory/validate/__main__.py
(5 levels below repo root), so the sys.path bootstrap needs parents[4],
not parents[3]; the old value pointed one directory too deep and
broke project-package imports.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-09 15:40:21 +02:00
kevinveenbirkenbachandClaude Opus 4.7 a26bfc4e94 refactor(cli): reorg admin/contributing tree + per-folder help system
Module moves (with imports, paths, argv literals, scripts, docs, and
parents[N] formulas updated in lockstep):

- cli/deploy             -> cli/administration/deploy
- cli/validate/inventory -> cli/administration/inventory/validate
- cli/vault              -> cli/administration/secrets/vault
- cli/administration/inventory/encrypt -> cli/administration/secrets/seal
- cli/mirror             -> cli/contributing/mirror

The new cli/administration/secrets cluster collocates the whole-file
ansible-vault wrapper next to the inventory-aware per-field sealer.
The user-facing CLI subcommand paths follow the package layout, e.g.
"infinito administration deploy dedicated <inv>" and
"infinito administration inventory validate <inv>".

Help system redesign in cli/core/{discovery,help,app}.py:

- Discovery now distinguishes runnable commands (directories with
  __main__.py) from category folders (directories that only group
  further commands). Empty support packages are filtered out.
- Each category folder MAY carry a README.md; its H1 heading is used
  as the title and its first paragraph as the one-line description in
  the help output.
- "infinito" / "infinito --help" prints only the top-level entries
  (one level deep). "infinito <category>" navigates into a folder and
  shows its sub-folders and commands one level deep, with the README
  title as the page heading and the invocation as a dim subtitle.
- New "--tree [N]" flag walks the tree under the (optional) path
  prefix recursively. N caps the depth (1-based); omit N for an
  unbounded walk. The output uses Unicode tree connectors, an emoji
  per entry (folder for categories, gear for commands), a globally
  aligned description column, and multi-line wrapped descriptions
  with tree-continuity bars on continuation lines.
- Category and command listings use the full invocation
  ("infinito meta callorder") rather than the bare leaf name, so the
  output is copy-pasteable.

Convention doc + 17 CLI READMEs:

- New docs/contributing/artefact/files/cli/readme_md.md describes the
  H1+paragraph layout the help system reads, and how it differs from
  role/docs READMEs.
- README.md added for cli/{administration, build, contributing, meta}
  and their immediate sub-folders, plus cli/meta/roles/applications
  and its resolution sub-folder. Each follows the "title with trailing
  emoji + single descriptive paragraph" convention.

Mirror resolver/sync description strings (the help system reads them
from argparse --help output) added so they no longer surface as
"options:" in the global index. cli/mirror/cleanup already had a
description and is unchanged.

Fixes uncovered by the move:

- import unittest.mock added to four dedicated-deploy tests so
  unittest.mock.patch resolves under stdlib unittest discovery
  (Python 3.13 no longer surfaces unittest.mock through "import
  unittest" alone).
- parents[N] index in the new
  cli/administration/inventory/validate/__init__.py corrected from
  3 to 4 to match the deeper layout.
- test_paths.INVENTORY_VALIDATOR_PATH assertion and test_runner
  substring assertions updated to match the new validator path.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-08 20:17:09 +02:00
kevinveenbirkenbachandClaude Opus 4.7 7dc6d6e491 chore(ruff): enable TC + N rules
TC (flake8-type-checking): typing-only imports moved into
``if TYPE_CHECKING:`` blocks across 8 source files via auto-fix
(117 hits). Reduces runtime import surface and rules out circular
imports for type-only deps. Project-wide ``from __future__ import
annotations`` (FA, already enabled) keeps annotations as strings,
so the moved imports never need to load at runtime.

N (pep8-naming): 32 hits handled.

Renames (semantic + lint):
- ``isSmallerThenMaximumBackupSize`` -> ``is_larger_than_maximum_backup_size``
  (snake_case + fixes both the typo "Then -> Than" and the
  semantic mismatch: the function returns ``current > maximum``)
- ``isDirectoryDeletable`` / ``deleteVersion`` /
  ``getAmountOfIteration`` / ``deleteIteration`` -> snake_case
- 8 test functions with ALL-CAPS env-var fragments
  (``test_..._DOMAIN_PRIMARY_...`` etc.) -> snake_case
- Mock alias ``VH`` -> ``vh`` in test_host_vars.py

Exception class renames (``N818`` Error suffix):
- ``TimeLimitExceededException`` -> ``TimeLimitExceededError``
- ``AttemptException`` -> ``AttemptError``
- ``PathNotFound`` -> ``PathNotFoundError`` (propagated across
  4 callers under tests/integration/lookups/config/)
- ``_ExitJson`` / ``_FailJson`` -> ``*_Error``

Refactor:
- ``METADATA = {...}`` local constant in two functions of
  ``plugins/filter/invokable_paths.py`` lifted to module-level
  ``_METADATA_KEYS``.
- ``from typing import TYPE_CHECKING`` in
  ``tasks/utils/migrate_meta_layout/__main__.py`` moved above
  the ``sys.path.insert`` hack to satisfy E402 (collateral from
  the TC auto-fix that landed it after the path mutation).

Per-file ignores in pyproject.toml for the legitimate cases the
rule cannot accept:
- ``pull-specific-host.py`` -> ``N999`` (Ansible-invoked role-level
  script, hyphenated filename pinned by the surrounding stack).
- 4 matrix-deploy files -> ``N806`` for the ``include_R`` /
  ``raw_include_R`` round-index notation; the uppercase ``R``
  here mirrors the formula in the docstring rather than naming
  a constant.
- ``plugins/filter/value_generator.py`` -> ``N806`` for the
  lazy-loaded class binding ``ValueGenerator =
  _value_generator_cls()``, which retains the imported class
  identifier per PEP 8.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-08 18:57:07 +02:00
kevinveenbirkenbachandClaude Opus 4.7 a080dc4f98 refactor(cli): inventory tools under administration/, drop role scaffolder + dead SSO indirection
Move:
- cli/build/inventory/full      -> cli/administration/inventory/devices
- cli/create/inventory          -> cli/administration/inventory/provision
- cli/create/credentials        -> cli/administration/inventory/credentials
- cli/encrypt/inventory         -> cli/administration/inventory/encrypt

Delete:
- cli/create/role + templates/roles/ (stale scaffolder)
- roles/web-app-baserow/templates/javascript.js.j2 + dead sso_warning snippet
- js_application_name across akaunting/baserow/opencloud/opentalk/taiga,
  javascript service disabled for baserow + taiga

Wire-through:
- argv lists in cli/deploy/{container,development}/* + 2 shell scripts
  now invoke 'infinito administration inventory provision'
- subprocess module names + docs (contributing/* + administration/deploy.md)
- discovery resolver test + folder-item-limit whitelist updated
- ruff fixes in devices/command.py (RET504 inline + Path.open over open)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-08 18:41:04 +02:00