📐mlab-builtins-io

The groupant beside mlab/src/builtins/io.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.

Parts

• mlab-builtins-io-format-parser
• mlab-builtins-io-atom-cycle

Model text

descriptant

mlab-builtins-io

Output registration and C-like format parser/cycler; scalar numeric formatting and num2str/mat2str conversions.

Modeling epoch — 2026-10-31

Prospective source read preceded edits. Bounded synthetic checks and selected existing safe tests exercise actual code; paired debug/release results and compiled negative-control assertions are retained externally. Frozen scenario, corpus, external measurement, full-school and training effects are unexecuted. Partial lexical syntax supplies no macro-expanded or semantic proof.

Cross-review refinement — 2026-10-31

Repeated flag/segment scans multiply conversion cost by expanded atom count; output length alone does not bound formatting work. The scalar-fid/char-format branch alone ignores a file identifier, and the numeric-first fallback returns no byte-count value. These source-derived qualifications change models only; retained runtime evidence and source hashes remain applicable.

approximant

Forward relation

sprintf_impl parses format, flattens real numeric matrices column-major to atoms, then cycles segments until atoms consumed. Strings under numeric conversions expand to Unicode scalar codes. fprintf formatted-string branch writes the completed string through Interp.out and returns UTF-8 byte count only for nargout>0. A scalar nonchar first argument is ignored as fid only when followed by a char format. The numeric-first fallback formats all arguments with %g\n and returns no values even for nargout>0.

Domain

Valid Rust value/storage invariants and available bounded resources; exact branch and effect observations only.

Executable approximant gap

No useful percentage-error surrogate is established for these categorical, textual or effectful observations. Supported future caps are 1%, 2%, 5%, 10%; no executable approximation or production substitution is claimed.

invariant

Invariants and costs

Formatting builds a String before calling out; string precision counts Unicode scalars, output count uses UTF-8 bytes. %g preserves negative zero via fmt_num, shortest canonicalizes it.

Failure ordering

Argument extraction and atom conversion precede output. Unknown format specs remain literal, excessive width/precision can exhaust resources rather than return a structured format error.

Cost

O(format bytes + expanded atoms + repeated segment/flag scans + emitted bytes); repeated conversion/flag scans can require O(format length * expanded atoms) work even with small output. String-to-code expansion and requested width/precision can dominate allocations. No resource bound or useful percent surrogate for exact text.

composition

Composition

sprintf_impl parses format, flattens real numeric matrices column-major to atoms, then cycles segments until atoms consumed. Strings under numeric conversions expand to Unicode scalar codes. fprintf formatted-string branch writes the completed string through Interp.out and returns UTF-8 byte count only for nargout>0. A scalar nonchar first argument is ignored as fid only when followed by a char format. The numeric-first fallback formats all arguments with %g\n and returns no values even for nargout>0.

Numeric atoms follow m.re storage, not printed row order; char matrices are one string atom. num2str default aligns columns while mat2str assembles row syntax.

format-parser

Scan Unicode chars; %% adds literal percent, flags/width/precision/length are collected before supported conversion membership; unsupported prefixes stay literal.

Boundary: Literal spans are unescaped when flushed; supported conversions separate literal and spec segments.

atom-cycle

Compute segments and atom queue; zero specs returns literals, otherwise cycle. Initially empty queue pads each spec once; exhausted noninitial queue returns immediately at the next spec.

Boundary: Every productive numeric/string conversion removes an atom or expands a finite string to code atoms; iteration ends when queue empties.

Interaction limit

Local fragment relations compose only with valid Value storage, caller error propagation and delegated providers. They do not establish whole-interpreter or host-filesystem guarantees.

models

Typed models — mlab-builtins-io

Source-derived partial relations; finite checks do not prove universality.

behavior_complement

Unsupported atom types exclude any successful formatted output even for a literal-only format because atoms is called first. Empty argument queue excludes numeric substitution; empty specs can still contribute width padding.

reverse_reconstruction

Rendered strings admit many inputs: integer/float atoms with same formatting, varying matrix shapes with identical real storage, ignored imaginary components and fid values in the qualifying scalar-fid/char-format branch. No inverse recovers those discarded components.

state_invariant

Formatting builds a String before calling out; string precision counts Unicode scalars, output count uses UTF-8 bytes. %g preserves negative zero via fmt_num, shortest canonicalizes it.

failure_precedence

Argument extraction and atom conversion precede output. Unknown format specs remain literal, excessive width/precision can exhaust resources rather than return a structured format error.

representation_transport

Numeric atoms follow m.re storage, not printed row order; char matrices are one string atom. num2str default aligns columns while mat2str assembles row syntax.

resource_cost

O(format bytes + expanded atoms + repeated segment/flag scans + emitted bytes); repeated conversion/flag scans can require O(format length * expanded atoms) work even with small output. String-to-code expansion and requested width/precision can dominate allocations. No resource bound or useful percent surrogate for exact text.

tests

Evidence and limitations

Selected named outcomes: builtins::io::tests::g_format, builtins::io::wave30_tests::cycling_unicode_and_empty. Each ran twice debug and twice release with sorted identical nonempty outcomes and numeric exit 0. Source-specific compiled control exercised "builtins::io::wave30_tests::cycling_unicode_and_empty": compile exit 0, assertion-red exit 101, exact-restored rerun exit 0. External receipts: octopus-wave30-io-2026-10-31 (native-commands.json, positive-receipts.json, controls.json).

Scope: Formatting builds a String before calling out; string precision counts Unicode scalars, output count uses UTF-8 bytes. %g preserves negative zero via fmt_num, shortest canonicalizes it.

AST projection generated twice to fresh files and compared byte-for-byte. Runtime evidence is finite and does not establish every branch, allocation failure, backend implementation or effect path. No optimization or speed claim.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/mlab/src/builtins/io.rs.groupant/