Skip to content

About these docs

Liferay Docker Manager is a community project maintained by Peter Richards. It is not an official Liferay product and is not covered by Liferay Support.

This documentation is written with the help of AI coding agents, with the maintainer accountable for what ships. Each page shows when it was last updated and when it was last reviewed (reviews are carried out by the agents as part of keeping the docs current). Changes reach this site through pull requests the maintainer merges.

Found something wrong or out of date? Open an issue.

What those statements are worth

A disclosure is only useful if you can check it. Here is what enforces each claim above, and โ€” just as importantly โ€” what does not.

The claim What enforces it
Every page carries a last-updated and last-reviewed date check_docs_review.py, run on every pull request. A page with no footer fails the build, not just a stale one.
Those dates are never more than 180 days old The same check, with --max-review-days 180 and --max-gap-days 180.
A date is never moved by automation The timestamp hook is append-only: it adds a footer where none exists and never rewrites an existing date. Bumping it on a real edit is the editing agent's job.
Changes ship only through a reviewed pull request A branch ruleset on master requiring a pull request, passing status checks and linear history. Pushing straight to master is not possible, for anyone.

Two statements on this page are not enforced by anything, and should be read as disclosures rather than guarantees:

  • that the documentation is written with AI agents. There is no check for this, and deliberately so โ€” the only way to enforce it would be to fail the build when a human writes a page by hand, which would mandate AI involvement rather than disclose it.
  • who performs the reviews. The build can tell that a page was reviewed recently. It cannot tell who did the reviewing. That reviews are carried out by the agents is a statement of practice, recorded in the project's AGENTS.md, not something a machine verifies.

Both are claims made against the project's interest rather than in its favour, which is why they are stated plainly and left unenforced. The claims you might actually rely on are the ones in the table.

Why the dates matter more than they look

Twice this project has removed a date that implied a review nobody had done. An earlier version of the timestamp hook rewrote "last updated" on every full run, so a commit touching one file stamped reviewed today across dozens of untouched pages and hid real staleness. Separately, the man page's version stamp was advertised as freshness when a release bump had moved it.

Both were corrected, and the rule that came out of them governs this page too: a date here means someone looked, or it means nothing at all.


Last Updated: 2026-10-15 | Last Reviewed: 2026-10-15