Docs pass: ARCHITECTURE.md + README.md catch up to v0.3.21 (part of #80) #84
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "docs-pass"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
First part of #80 — the Cairn repo docs. Site and license-server docs follow in their own PRs.
§5.4 renamed to "Authorization and data governance" and given the four features that shipped API-first and were undocumented:
-rogroup-name convention,storage.ErrReadOnly, and that the UI badge is advisory while scope enforces.retainMonthsprune, what is and isn't recorded (never List/Stat), and that it covers public share reads via the unscoped store.Decorator order is now written down, which it wasn't anywhere. I verified it against
cmd/cairnd/main.gorather than trusting memory — and it's worth flagging that it is base → encrypt → quota → trash → hold → audit → scope, i.e. audit sits above trash (a user delete logs asdelete, not the rename trash does beneath it) and hold sits under audit. My own working notes had recorded "hold → audit → trash", which was wrong; the doc follows the code.§7 gains the
admin/auditandadmin/holdsendpoints, and the stale "Planned additions: quota enforcement" line is corrected (enforcement shipped in v0.3.1).§9 scaling posture: the role-split is no longer hypothetical — documented the shipped
--target=all|api|crypto(#33) and theremoteUrl/remoteTokenoffload.§10 milestones: v0.3 now lists the governance features, crypto split, and auto-purge; v0.4 marks license server + site as live and leaves only the GitHub mirror pending.
README gains one Data-governance feature bullet.
Scope note: I deliberately left §4.3's pre-implementation "design task before any implementation" encryption bullets untouched — that's a separate contradiction (the section intro already reflects the shipped probe) and outside this issue's list. Flagging it rather than fixing it blind. SYNC.md I reviewed in full and it is accurate to current behaviour (app passwords,
.249, per-user homes) — no change needed; the "interop matrix" I worried about in the issue text doesn't live there.