Policy documents for services run by New Vector
Documents are versioned using semver:
- Major bump - significant change requiring subjects' specific re-acknowledgement/acceptance
- Minor bump - notable change in text, but not something that would require subjects' specific attention
- Patch bump - fixing a typo
Versioning metadata is stored in the top of the policy doc - each doc has a metadata segment inspired by gatsby that looks something like:
---
title: Document Title
version 1.0.0
---
Identity servers and integration managers manage tracking subjects' 'accepted terms' by tracking the URL representing those terms - it is therefore imperative that any generated links to docs include the major version and only the major version (e.g. https://example.com/document-title-1) so that only a change in the major version triggers the 're-optin' flow.
/scripts/versions.py
will mine the github commit history for a given document and output a file representing the most recent state of that document for each numbered version. versions.py
doesn't handle templating or markdown->html.
To apply this to all of the policy docs:
for FILE in $(find docs -name '*.md' | grep -v README.md); do ./scripts/versions.py $FILE; done;
Most of these files are templated using {{ variable_name }}
, to be digested by ansible/jinja2 somewhere down the line for deployment to different instances. To apply New Vector data to the templates, you can use jinja2-cli:
$ jinja2 <filename> ~/path/to/new_vector.json --format=json
This section needs some work. You can generate a full set of html docs from the .md source using:
npm run build
This will create a /build
directory with a folder structure mirroring /docs
, with a .html file in place of every .md file.
At the moment this does not play nicely with versions.py
, and it is questionable whether it is a good idea at all to have a build system as part of the docs repo when, in some instances at least (such as the matrix.org website) what's needed is a markdown file that gatsby will then ingest.