Struct LogRow 

pub struct LogRow {

    pub ts: DateTime<Utc>,
    pub ts_seq: u64,
    pub event: LogEvent,
    pub ref_: Option<String>,
    pub source: Option<String>,
    pub result: LogResult,
    pub license: Option<String>,
    pub size_bytes: Option<u64>,
    pub store_path: Option<String>,
    pub capability: Capability,
    pub session_id: String,
    pub error_code: Option<String>,
    pub schema_version: String,
    pub canonical_digest: Option<String>,
    pub prev_hash: String,
    pub this_hash: String,
}

One row of the provenance log (PROVENANCE_LOG.md §3).

The on-disk wire field names match the spec table; struct-field order is not load-bearing for the hash because canonicalization sorts keys lexicographically (see PROVENANCE_LOG.md §4).

Schema version: this struct is the v2 row shape (ADR-0024). Every v2 row carries schema_version = "v2" literally; the canonical_digest field carries the ADR-0021 §1 audit identity of the fetch on rows where one applies (Fetch / Resolve / StoreWrite) and is None on session bookend rows (SessionStart / SessionEnd / CapabilityResolved) that have no ref. v1 rows (pre-Slice-4) lack both fields and MUST be migrated via migrate_v1_to_v2 before the v2 binary can read them — the deny_unknown_fields + non-defaulted schema_version shape ensures v1 rows fail to parse loudly rather than producing silent hash-chain mismatches.

Fields

ts: DateTime<Utc>

RFC3339 UTC timestamp of the append (millisecond precision).

ts_seq: u64

Per-session monotonic sequence number, starting at 1.

event: LogEvent

Event class (see LogEvent).

ref_: Option<String>

Optional reference (DOI / arXiv id). Wire field name is ref.

source: Option<String>

Optional source name (e.g. unpaywall).

result: LogResult

Result (see LogResult).

license: Option<String>

OA license string (event=fetch, result=ok); None otherwise.

size_bytes: Option<u64>

Bytes written / fetched, on success rows.

store_path: Option<String>

Path to the stored payload, relative to the store root (event=fetch, result=ok); None otherwise.

capability: Capability

Capability under which the row was written (REQUIRED, every row).

session_id: String

26-char ULID identifying the process invocation (REQUIRED).

error_code: Option<String>

Stable error code on failure rows.

schema_version: String

Row schema version. Always LOG_SCHEMA_VERSION ("v2") for new rows written by this build (ADR-0024). v1 rows lack this field; they MUST be migrated via migrate_v1_to_v2 first.

canonical_digest: Option<String>

Canonical-digest of the fetch’s audit identity (ADR-0021 §1) as 64 lowercase hex chars. Present on rows with a ref (Fetch, Resolve, StoreWrite); None on session bookend rows. The digest is computed from a crate::CanonicalRef whose resolver_profile matches this row’s source field for migrated v1 rows; new v2 rows MAY pass an explicit resolver_profile distinct from source.

prev_hash: String

64 lowercase hex chars, OR the literal string "GENESIS" for the first row of a fresh log file.

this_hash: String

64 lowercase hex chars. SHA-256 of canonical JSON of THIS row with the this_hash field removed. See module docs.

Trait Implementations

impl Clone for LogRow

fn clone(&self) -> LogRow

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for LogRow

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for LogRow

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for LogRow

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.