Skip to content

docs(0.6): align bounded InMemory API with published 0.5.0 boundary - #93

Draft
agustin18 wants to merge 22 commits into
stablefrom
docs/92-060-release-boundary-audit
Draft

agustin18 wants to merge 22 commits into
stablefrom
docs/92-060-release-boundary-audit

Conversation

@agustin18

Copy link
Copy Markdown
Owner

Purpose

Post-merge documentation audit for #85: the source on stable now implements InMemoryBackend(max_entries=...), but the last published fastapi-cache-stable==0.5.0 wheel/sdist is still unbounded. Avoid telling pip users to pass a constructor parameter unavailable in their pinned distribution.

Scope

Verifications required

  • All exact-head CI/changelog/docs site and link/anchor checks successful.
  • Independent review confirms no contradictory 0.5.0/0.6.0 usage or release-status statements and no broken site links.
  • Any new sample marked source-only until actual 0.6.0 publication.
  • No public deployment, PyPI tag, version bump, or release performed by merging this docs PR.

Release strategy: This is a small docs-only follow-up to the separately approved production PR #85; it does not require freezing 0.6.0 or publishing now. Refs #54 #86 #92.

Copy link
Copy Markdown
Owner Author

Documentation counteraudit package — exact HEAD CI-clean (2026-11-21)

Review SHA: 1aa2b2151eb09f93bae8c7f7b79083e7bc201e64 (PR remains DRAFT, not yet merged; stable currently contains #85 at 190d1059). GitHub compare: 22 commits ahead / 0 behind; 17 changed files, documentation plus release-boundary regression and changelog only. No version, code API, runtime dependency or release changed.

All exact-head workflows SUCCESS:

  • CI 29/29: lint/pre-commit, changed-file typecheck, release-boundary tests, Python 3.8–3.13, distribution inspection.
  • Documentation site: 15 generated pages and 588 local links validated, including version-labelled feature matrix; no Pages deployment.
  • CodeQL, Changelog, PR title: success.

Reviewer attention:

  1. Can a developer on fastapi-cache-stable==0.5.0 accidentally copy any max_entries snippet without the warning that it is unavailable? Published 0.5.0 references must remain unbounded.
  2. Is the post-feat(inmemory): optional max_entries LRU capacity without changing defaults (#54) #85 merged into stable, not yet published status consistent in README, ROADMAP, COMPATIBILITY, QUICKSTART, site docs, the backend matrix and version history?
  3. Is max_entries documented correctly as positive integer/None, fixed/read-only after constructor, per-instance entry count (not payload bytes/global multiworker capacity), LRU+expired-first with indexed heap, no background sweeper and bulk-clear scan?
  4. Are feat(observability): optional cache events after Redis modernization #50 event callbacks still not shipped and deferred by default, while security(cache): residual hidden tenant/identity context and header-cardinality safety after 0.5.0 #53 residual security, test(0.6): two genuinely independent consumer exercises and evidence gate #91 external consumers (0/2), docs: build versioned documentation site after PR #52 quickstart #48 Pages and release(0.6): exact-artifact acceptance, security and independent adoption gates #92 release gates remain clearly separate?
  5. Do not alter tagged historical 0.5.0 docs, 0.5.0 PyPI metadata, or run Pages. Source docs are previewed with Zensical only.

Provisional decision: documentation technical PASS, open for maintainer/counteraudit review. No automatic merge/tag/PyPI deployment.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant