📐cycle-db-commit

The groupant beside tools/cycle/src/db.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.

Facts

Components

Obernants

The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.

Model text

Descriptant

Cycle db commit

Descriptant

Prospective inspection identified this relation before additions: Merge current by id, write tmp, rename current, reread input, append history, append event log.

The production implementation is unchanged; only selected inline tests are appended. Source-bound syntax is generated twice and compared. Native checks exercise bounded helpers; entry orchestration, teacher calls, models and corpora remain source-derived. Partial syntax is not semantic proof. Preserve failure prefixes and ambiguity when refining this model.

Approximant

Forward relation

Check promotion lock and content-bound policy, read/validate one immutable incoming buffer, reject selected conflicts, merge by ID, write tmp and rename current, append history from the same buffer, append event log. Legacy cross-file effects remain ordered, not atomic.

Composition

This fragment isolates the stated helper relation; parent composition retains surrounding effects and failures.

Catalog

Typed observations: cycle-db-commit

Behavior complement

Successful unique-id replacement cannot append position; post-rename failure cannot imply unchanged current.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Reverse reconstruction

Old+final current admit overwritten intermediate incoming versions. Counts constrain multiplicity, history may retain bodies but may be partial.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Duplicate index

Initial map stores last index for duplicate id; incoming duplicates use updated map and append only once.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Rename boundary

Current commit precedes history open/write and log open/write; error effects partition by this order.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Validated input snapshot

One incoming read is validated before effects and reused for current and history; a later input-path replacement cannot bypass the reviewed guard. Concurrent database writers remain outside the quiescent guarantee.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Timestamp scope

One seconds timestamp per call; pre-epoch failure maps zero. Time is not unique event identity.

Scope: the bound source interval and stated assumptions. Evidence: source inspection plus the checks explicitly named in model.md#tests; no finite check establishes all inputs.

Tests

Checks and limits

db::tests::replace_and_append covers positions/counts/current/log. Mid-write failure and concurrent reread gaps.

Exact native outcomes and command metadata are retained in the external wave33 cycle receipt bundle. Selected names run twice in debug and release; syntax is independently regenerated twice. Compiled mutations and source restoration receipts distinguish behavior changes from assertion-only controls. These claims are finalized in ../model.md#evidence. No live teacher, real model/corpus, benchmark CLI or lesson-learning CLI is run.

Invariant and cost

Invariant and cost

Incoming repeated new id adds once then replaces; initial duplicate ids remain except last indexed occurrence. No rollback.

O(old+incoming bytes+events log S), O(old+incoming) memory plus policy validation. Current and history reuse the same incoming snapshot; nontransactional append failures can still leave prefixes.

No executable percentage surrogate is claimed: effects, categorical acceptance and report text have no justified approximate metric here. Future profiles require a useful metric, zero case, reference and evidence at caps 1/2/5/10%.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/cycle/src/db.rs.groupant/commit.groupant/