Skip to content

ci: fmt under stable 1.96 to make rust-ci green (follow-up)#47

Merged
hyperpolymath merged 9 commits into
mainfrom
claude/new-session-znxgm7
Jun 28, 2026
Merged

ci: fmt under stable 1.96 to make rust-ci green (follow-up)#47
hyperpolymath merged 9 commits into
mainfrom
claude/new-session-znxgm7

Conversation

@hyperpolymath

Copy link
Copy Markdown
Owner

Summary

Follow-up to the merged CI-green PR. With rust-ci now running, cargo fmt --all -- --check fails on main under the CI toolchain (stable rustfmt 1.96) — the toolchain-pin landed before the formatting fix. This brings the Rust sources to rustfmt-clean so rust-ci goes green.

Changes

  • Rust hygiene: cargo fmt (src/codegen/parser.rs) so cargo fmt --all -- --check passes under stable 1.96.

RSR Quality Checklist

Required

  • Tests pass (cargo test --locked --all-targets)
  • Code is formatted (cargo fmt --all -- --check)
  • Linter is clean (cargo clippy --locked --all-targets -- -D warnings)
  • No banned language patterns
  • SPDX license headers present on modified files
  • No secrets, credentials, or .env files included

Testing

Verified locally with the CI toolchain (rustc/clippy/rustfmt 1.96.0): cargo fmt --check, clippy -D warnings, cargo check --locked, cargo test --locked all pass.

🤖 Generated with Claude Code


Generated by Claude Code

claude and others added 8 commits June 27, 2026 19:29
Raise the Idrisiser Idris2 ABI to Layer 2 with a genuine, machine-checked
semantic proof of the repo's headline domain property (proven-correct
wrappers via dependent types).

Module Idrisiser.ABI.Semantics models a length-indexed SafeArray (Vect n a)
with a total accessor safeIndex : SafeArray n a -> Fin n -> a, and proves:

- InBounds i n is inhabited exactly when i < n (carries an LT i n proof);
  the out-of-range case has no constructor.
- Uninhabited (InBounds i 0): no index exists into an empty array, so
  out-of-range access is unrepresentable.
- decInBounds : sound + complete Dec (InBounds i n), backed by isLT.
- finToNatNatToFinLT / safeIndexUsesExactSlot: the proven index round-trips
  exactly to the requested raw position (no off-by-one, no runtime failure).
- certifyBounded + certifyBoundedSound: certifier into the ABI Totality
  witness, returning Total only when genuinely in bounds.
- Positive control (idx2InBounds reads "gamma") and negative controls
  (Not (InBounds 5 3), Not (InBounds i 0)), all machine-checked.

No believe_me / postulate / assert; the headline property is non-vacuous
(a deliberately-false InBounds 5 3 witness is rejected by idris2 0.7.0).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6PSzJWpRxtzGDjUCEh7Mx
Add a second, deeper machine-checked theorem over the SAME Layer-2 model
(SafeArray/Vect/Fin), raising the Idris2 ABI to Layer 3.

New module Idrisiser.ABI.Invariants proves McCarthy's store/select array
axioms — distinct from Layer-2 in-bounds totality:
  - safeReadAfterWriteSame: read i after write v@i = v (write-then-read)
  - safeReadAfterWriteOther: write@j leaves slot i/=j unchanged (read-other)
  - predictReadOverWrite + safeReadOverWriteCorrect: sound decEq-driven
    decision procedure with correctness proof in both branches

Vect-level lemmas (indexReplaceAtSame, indexReplaceAtOther) proved by hand
via induction, not delegated to a library lemma. Defines safeWrite as the
formal dual of safeIndex. Positive controls (write-then-read, other-slot
preserved, prediction agrees) plus a negative/non-vacuity control
(writeReallyChangedSlot: the write genuinely changed the slot). %default
total, no believe_me/postulate/assert. Clean build, zero warnings;
adversarial false-equality check rejected.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6PSzJWpRxtzGDjUCEh7Mx
Add Idrisiser.ABI.FfiSeam proving the resultToInt encoding is sound:
- resultRoundTrip: total decoder intToResult faithfully recovers every
  Result (lossless C-integer encoding).
- resultToIntInjective: distinct ABI outcomes never collide on the wire,
  derived cleanly from the round-trip via cong + Just-injectivity.
- positive controls (concrete decodes = Refl) and a machine-checked
  non-vacuity control (Ok and Error map to distinct wire integers).

Genuine proof: no believe_me/postulate/assert_total/idris_crash. %default
total, zero warnings. Registered in idrisiser-abi.ipkg.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6PSzJWpRxtzGDjUCEh7Mx
Assemble the existing per-layer ABI proofs into one inhabited end-to-end
certificate in Idrisiser.ABI.Capstone:

- record ABISound with one field per discharged layer:
  * flagshipControl  : InBounds 2 3 (Layer 2 flagship positive control)
  * layer3Invariant  : write-then-read law instance (Layer 3 invariant)
  * ffiSeamInjective : resultToInt injectivity (Layer 4 FFI seam)
- abiContractDischarged : ABISound built ONLY from the real exported
  witnesses (Semantics.idx2InBounds, writeThenReadDelta,
  resultToIntInjective). If any prior layer were unsound this value
  would not typecheck.

Non-vacuity: a bogus certificate (Layer-3 field claiming the old value
"beta" instead of "delta") is rejected by the type checker.

Genuine composition only: no believe_me/postulate/assert/%hint.
%default total, SPDX header, zero build warnings.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6PSzJWpRxtzGDjUCEh7Mx
…ble fix); port ABI-FFI gate Python->Bash (Python is estate-banned)

Resolves the standing baseline CI reds (rust-ci toolchain error, governance
Language/anti-pattern, governance workflow-lint) without altering the proven
ABI. The Bash gate reproduces the former Python gate's verdict verbatim
(validated across all -iser repos) and catches the same drift classes.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A6PSzJWpRxtzGDjUCEh7Mx
@hyperpolymath hyperpolymath marked this pull request as ready for review June 28, 2026 09:23
@hyperpolymath hyperpolymath merged commit 798961d into main Jun 28, 2026
4 checks passed
@hyperpolymath hyperpolymath deleted the claude/new-session-znxgm7 branch June 28, 2026 09:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants