Repository navigation
proofs(agda): Compose group in Agda, reconciled (PWG-2i for 12); UsesK pragma fix - #64
hyperpolymath wants to merge 3 commits into
Conversation
The mathesis census reads every Agda file with no `OPTIONS --safe` as an estate-policy break (proof-holes.sh), so this fixture counted as one. Its pragma now sets --safe alone. It still leaves out --without-K and --cubical-compatible, which implies --without-K. Under Agda's defaults, which have K, the file is accepted, as checked locally with no flags. The gate's refusal therefore still shows that its own command line runs with --without-K. The correspondence doc says so. The receipt is unaffected, because fixtures are not in the input set. Agda-source session (estate proof policy), Kernel group follow-up. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VTeVWDEJgdYKsUNr4uhnXX
|
No actionable comments were generated in the recent review. Summary
Merge Risk:
|
proofs/agda/Marches/Compose.agda is the Agda source of Marches/Compose.lean. It has the same constructors and the same theorems: - the two-region positions, labels and interface; - stratified preference (`CBetter`, `ctrans`, `cirrefl`); - the typed step and trace judgments; - `no-back-cross` and `no-a-label-in-B`; - `step-worsens`, `trace-cbe`, `trace-worsens` and `wt-cycles-are-empty`. Prelude gains `⊥-elim`. A new fixture, tests/reject-agda/BackCross.agda, offers a crossing from B back into A, and it must be refused. Agda reports UnsupportedIndexedMatch warnings for ctrans, cirrefl and trace-worsens. Those concern computation under Cubical transports, not the statements, so they stay visible, as in januskey. Agda-source session (estate proof policy), Compose group. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VTeVWDEJgdYKsUNr4uhnXX
docs/agda-lean-correspondence.adoc gains the Compose section. It compares the eight theorems binder by binder and adds residual R4: Agda's UnsupportedIndexedMatch warnings concern computation under Cubical transports, not the statements. PROOF-STATUS.adoc: - the Agda receipt is retaken at 2c4f9a2, digest 6e9394f…, with five input files; - the lineage: the Kernel-only receipt at 660e1c6 is superseded, and its digest still holds on main at 08f6a79; - the Lean re-check of Compose reports no axioms; - the census is PWG-2i 12 and ungraded 199, of 211. README, ROADMAP and STATE are updated to match. Agda-source session (estate proof policy), Compose group. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VTeVWDEJgdYKsUNr4uhnXX
Second group of the Agda port, plus a small fixture fix. They share a PR because this session works on a single branch.
1. UsesK sets
--safein its pragma (9a1feaf)The mathesis census flags an Agda file without
{-# OPTIONS --safe #-}as an estate-policy break (proof-holes.sh). So after #63 the marches row showedpolicy 1, fromtests/reject-agda/UsesK.agda.--safealone, still without--without-Kor--cubical-compatible.--without-K.2. Compose group (
2c4f9a2,9392318)proofs/agda/Marches/Compose.agdaportsMarches/Compose.lean.Pos,CLabel,Interface,CBetter,Step,Trace.ctrans,cirrefl,no-back-cross,no-a-label-in-B,step-worsens,trace-cbe,trace-worsens,wt-cycles-are-empty.Preludegains⊥-elim.tests/reject-agda/BackCross.agdaoffers a crossing from B back into A, and it must be refused.UnsupportedIndexedMatchforctrans,cirreflandtrace-worsens. The cause is constructor injectivity, which Cubical Agda cannot yet compute along transports. This concerns computation, not the statements. The warnings stay visible, as in januskey, and are recorded as residual R4.docs/agda-lean-correspondence.adocgains a Compose section, compared binder by binder.2c4f9a2, inputs-sha2566e9394f6…6f7d, five files;660e1c6is superseded, and its four files give the same digest4f589df…onmainat08f6a79(an echo);PWG-2i12 andungraded199, of 211.Evidence (local)
scripts/check-agda.shat2c4f9a2ends with[check-agda] OK. Three fixtures are refused for their stated reasons, and the--safecontrol is refused too.#print axiomsreports none for all eight Compose theorems.Marches/is unchanged, tree7ef83e4.Next
GaoRexfordgroup. Its Lean leg rests onpropext, soPWG-2ithere needs that axiom's disposition (c).🤖 Generated with Claude Code
https://claude.ai/code/session_01VTeVWDEJgdYKsUNr4uhnXX