📐algorithm-patience-diff-tests

The groupant beside algorithms/src/patience_diff/tests.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

Epoch memory — 2026-10-30

Original bounded byte patience_diff. Reference: Bram Cohen (2007), Patience Diff Advantages. Prospective equations preceded implementation. Exact semantics and observed checks are distinct from syntax binding. Round-trip cannot alone validate two mutually wrong codecs; retain hand-derived streams and malformed checks. No timing claim or percentage approximant.

model

Mathematical combiner

For source S, partial syntax P(S), forward relation F and typed relations O, the sidecar records (hash(S),P(S),F,O,I). Source identity binds descriptions to bytes but does not prove F. Tests intersect compatible implementations with finite observations. Composition with callers requires domain caps and exact model identity; no global Mova accuracy claim follows. A new byte string alone cannot recover old or script: inserting a into empty and keeping a both yield a. With script, Equal/Delete projection uniquely recovers old and Equal/Insert projection uniquely recovers new.

approximant

Forward, invariant, and cost

Strip equal prefix/suffix. Match bytes occurring exactly once in both remaining slices; choose LIS of new positions (strict updates keep earliest predecessor and endpoint), recursively diff gaps. No anchors means all deletes then all inserts. Apply validates consumed old bytes.

O(n^3) conservative time and O(n^2) conservative recursive storage for n<=4096; anchors per segment <=256.

Diff inputs cap 4096; apply source cap 4096, edit cap 8192, full consumption and identity before output budget min(user,65536). No floating arithmetic or callbacks. Recursive depth <=input length; allocator failures not caught.

Invariant: A successful patch producing a byte sequence other than its Equal/Insert projection is excluded; its Equal/Delete projection must exactly equal old. Minimal edit count is not guaranteed: abab->baba emits 8 changes although 2 suffice.

Original implementation informed by Bram Cohen (2007), Patience Diff Advantages; no copied implementation. Exact categorical semantics; no useful certified percentage surrogate is attached. Patience minimality is explicitly unsupported. Finite fixtures do not constitute universal proof.

models

Typed test relations

behavior_complement

A passing transcript excludes any tested output failing the explicit expected value, reconstructed bytes, or independent oracle assertion. It excludes neither untested inputs nor environmental allocation failure. For production semantics: A successful patch producing a byte sequence other than its Equal/Insert projection is excluded; its Equal/Delete projection must exactly equal old. Minimal edit count is not guaranteed: abab->baba emits 8 changes although 2 suffice.

reverse_reconstruction

A green transcript identifies the equivalence class of programs agreeing on the finite fixtures; many incorrect programs remain compatible outside those fixtures. A new byte string alone cannot recover old or script: inserting a into empty and keeping a both yield a. With script, Equal/Delete projection uniquely recovers old and Equal/Insert projection uniquely recovers new.

oracle_independence

Myers compares a full edit grid; LZ77 independently enumerates feasible lengths and nearest tie; codecs use round-trip plus hand-derived streams/malformed cases; BPE uses specified first-merge/tie values; patience checks both script projections and a nonminimal witness. Round-trip alone is insufficient.

boundary_partition

Finite empty, cap, malformed-code, overlap and resource witnesses sample contract branches; exhaustive binary strings do not cover all 256 symbols or allocator behavior.

mutation_sensitivity

A compiled production mutation must fail the named assertion; successful detection distinguishes that fault only. The original source is restored byte-for-byte and rerun.

tests

Checks

Run cargo test -p mova-algorithms patience_diff::tests in debug/release twice. Preserve normalized transcripts, compiled negative control, and exact restoration under $MOVA_DATA/runs/octopus-wave25-sequences-2026-10-30. No benchmark evidence yet.

prospective

Prospective finite checks

Independent oracle or enumerated fixtures, empty and invalid boundaries, reconstruction, and compiled production mutation; test success is finite evidence only.

validation

Observed validation — 2026-10-30

The five-module batch adds 14 native tests. All pass twice debug and twice release; sorted named test records match across all four runs. Initial integrated library run: 632 passed. Strict library Clippy passed. Each production file has a compiled mutant rejected by a named assertion, followed by byte-for-byte restoration and a passing module rerun: Myers score+1/exhaustive_dp; LZW flipped decoded byte/known_and_malformed; LZ77 match length-1/greedy_overlap_and_invalid; BPE tie comparison >=/ties_and_model_identity; patience anchor Insert instead of Equal/unique_anchor_and_nonminimal_witness.

Each of ten source/test files has paired identical partial syntax projections, refreshed after all mutants were restored. Receipts: $MOVA_DATA/runs/octopus-wave25-sequences-2026-10-30/{focused-debug-1.log,focused-debug-2.log,focused-release-1.log,focused-release-2.log,normalized-,mutations.log,mutant-,restored-*,syntax.log,clippy.log}. No timing or speedup evidence; allocator failures untested. These finite observations and lexical source bindings do not establish a semantic proof or global Mova quality improvement.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/algorithms/src/patience_diff/tests.rs.groupant/