Elamur Groupantsmodelcheck-library-syntax-envelope
The groupant beside tools/modelcheck/src/lib.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
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | Forward |
| ast | partial | ../../syntax.json |
| cost | partial | Cost |
| descriptant | partial | Descriptant |
| invariant | partial | Invariant |
| tests | partial | Tests |
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.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial relational specification | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial relational specification | reverse_reconstruction |
integrity_before_selection | integrity_before_selection | partial relational specification | integrity_before_selection |
permutation_quotient | permutation_quotient | partial relational specification | permutation_quotient |
unselected_damage_dependency | unselected_damage_dependency | partial relational specification | unselected_damage_dependency |
gap_preservation | gap_preservation | partial relational specification | gap_preservation |
Model text
Subject: current manifest source SHA and exact UTF-8 range [27698,29521). The observation is one completed scoped call with adequate resources; source-derived relations are partial specifications. Filesystem checks describe sequential reads, not an atomic snapshot.
Forward
Absent syntax_nodes returns success. Otherwise confine and reread/decode the selected artifact through the file-tree reader, fully validate current source and all nodes, then require a nonempty distinct unsigned ID list. Lookup each node by validated index, accumulate min start/max end, and compare exactly with fragment byte_range.
Invariant
The selection cannot use duplicate/missing IDs or a wider/narrower range. Memo reuse is decoding only; integrity checks precede selected ID consumption on every call.
Cost
Artifact read/equality O(B), validation O(source bytes + N), selection O(S log S), with one O(B+N) decode on each memo miss. S selected IDs do not replace whole-artifact validation.
Descriptant
Wave34, 2026-10-31: deeper scope retained under its existing parent without replacing historical views. Decode reuse was proposed before edits and checked against actual reference decoding; behavioral controls and repeated native receipts remain finite evidence. No numerical surrogate is justified for categorical validation outcomes; no percentage accuracy or complete semantic proof is claimed.
Tests
Exact native test: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Parent tests additionally exercise integration and failure order. Current evidence is in ../../wave34-evidence.md; old evidence remains explicitly historical in wave33-evidence.md at the source root.
behavior_complement
Exclude success for opted-in anchors whose selected set is empty/nonunique, whose IDs do not exist, or whose range differs from extrema. Invalid unselected nodes also exclude success because validation runs before selection.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
reverse_reconstruction
Given envelope [a,b), compatible ID sets have minimum start a and maximum end b and are nonempty and distinct. Interior nodes, overlapping nodes and permutations remain ambiguous; root inclusion is allowed if it yields the envelope.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
integrity_before_selection
The observation order is artifact admission and full integrity, then IDs, then final equality. A stale projection paired with duplicate IDs reports artifact staleness first.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
permutation_quotient
For a fixed valid projection, permuting a distinct selected list preserves extrema and acceptance. Adding a duplicate changes acceptance despite preserving extrema.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
unselected_damage_dependency
Hold selected nodes fixed and corrupt another projection node. The full-artifact gate still rejects, so an envelope-valid subset cannot hide artifact damage.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
gap_preservation
A selection of disjoint nodes takes their enclosing interval, including gaps. This does not assert the model covers the omitted source or reconstruct internal statements.
Assurance: source-derived partial relation; fixed source revision and stated scope. Oracle: groupant_tests::fragment_syntax_anchors_bind_envelopes_and_current_source. Finite observations do not establish universal completeness.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/modelcheck/src/lib.rs.groupant/fragment-validation.groupant/syntax-envelope.groupant/