Domain model

These pages describe what Adriano stores and the rules it applies, as implemented in app.domain (pure logic) and app.storage (persistence). They are written for developers changing that logic. How authors write ADR and erratum files is covered for authors in Users.

        flowchart LR
  file["ADR or erratum file<br/>(bytes from a source or CI)"]
  parsed["ParsedRecord<br/>metadata, links, issues"]
  digest["Digest<br/>sha256-v1"]
  row["record row<br/>published or candidate"]
  rev["revision<br/>canonical snapshot"]
  event["decision_event<br/>accepted or rejected"]
  state["Approval state<br/>of the current digest"]

  file -- "parse_record()" --> parsed
  file -- "content_digest()" --> digest
  parsed -- "upsert" --> row
  digest -- "names" --> rev
  row -- "current digest" --> rev
  event -- "identity + digest" --> state
  row -- "identity + current digest" --> state
    

• Records
  • From file to row
  • Identity
  • Published and candidate rows
  • Revisions and change tracking
  • Relationships
  • Validation issues
  • When a record can be decided
• Content hashing
  • Algorithm sha256-v1
  • Test vectors
  • Why paths and refs are excluded
  • Bytes that are not valid UTF-8
  • What the digest does not prove
  • Deferred: a signed audit export
• Approval
  • The policy
  • States
  • Candidates, published rows and promotion
  • Recording a decision
  • Self-approval
  • Diffs
  • Freshness
  • Errata
• Errata
  • Model
  • Rules
  • Counting errata on a decision
  • Issue codes
• Database schema
  • Migrations
  • Tables
  • Constraints and indexes
  • Triggers