Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions CHANGELOG.adoc
Original file line number Diff line number Diff line change
@@ -1,5 +1,13 @@
== Changelog

== 2026-11-17 — F-5's distinctness lemma moves into the graded cone

* F-5's distinctness lemma `maps-disagree-at-positive` moves from
`tests/unit/MorphismTest.agda` into `src/ResourceAlgebra/Morphism.agda`.
The ledger cites it as part of F-5, but in the unit tests it sat outside the
graded import cone (`All.agda`) and so outside the Agda receipt's inputs
manifest: a change to it would not have invalidated F-5's receipt.

== 2026-11-17 — `CI/gate` required check and declared coverage exemption

* `proofs.yml` gains a job named exactly `CI/gate` (the context the live
Expand Down
11 changes: 9 additions & 2 deletions docs/status/PROOF-STATUS.adoc
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
// SPDX-License-Identifier: CC-BY-SA-4.0
= resource-algebra — Proof Status
:revdate: 2026-11-11
:revdate: 2026-11-17

Every row is a theory theorem (`PROOF-NEEDS.adoc`). Records follow
`3-practice/PROOF-POLICY.adoc` in `hyperpolymath/standards`, which keeps
Expand Down Expand Up @@ -145,6 +145,13 @@ support. The receipt deed is written by hand from the run; no gate writes
or recomputes it yet. The run summaries that hold the per-leg digests are
kept by GitHub for a limited time.

One gap in the receipt at `f211fc1`: F-5 cites `maps-disagree-at-positive`,
which was then in `tests/unit/MorphismTest.agda`. The run checked it (the
gate checks the unit tests), but the Agda leg's inputs manifest lists only
`src/`, so a change to that file would not have invalidated F-5's receipt.
Comment thread
coderabbitai[bot] marked this conversation as resolved.
Resolved
The lemma is now in `src/ResourceAlgebra/Morphism.agda`, inside the cone, and
the next receipt covers it.

Earlier receipts: run `42629220745` on source commit
`9f175cc57f9c2c2db255c1b3161e24cfba4abc4d` (2026-11-09). They are
`superseded`: F-5's classifier, the strict flag set and `⊕-least` all
Expand Down Expand Up @@ -411,7 +418,7 @@ none goes to MaxPlus or to Loss.
`support`. Neither distinguishes `fin 1` from `fin 2`.
| `PROVEN`. Both are morphisms (`zero-test-hom`, `support-hom`) and every
morphism is one of them (`Loss→MaxPlus`). They are distinct
(`maps-disagree-at-positive`, tests/unit/MorphismTest.agda), and every
(`maps-disagree-at-positive`), and every
morphism collapses `fin 1` and `fin 2` (`Loss→MaxPlus-collapses`,
`no-injective-Loss→MaxPlus`).
| 1
Expand Down
4 changes: 4 additions & 0 deletions src/ResourceAlgebra/Morphism.agda
Original file line number Diff line number Diff line change
Expand Up @@ -254,6 +254,10 @@ no-injective-Loss→MaxPlus : (h : Hom Loss MaxPlus)
no-injective-Loss→MaxPlus h inj with inj (fin 1) (fin 2) (Loss→MaxPlus-collapses h)
... | ()

-- The two morphisms are distinct: they disagree at every positive finite grade.
maps-disagree-at-positive : ∀ n → zero-test (fin (suc n)) ≢ support (fin (suc n))
maps-disagree-at-positive n ()

----------------------------------------------------------------------
-- MaxPlus → Loss: exactly one strict morphism
support′ : Tropical → L.Grade
Expand Down
3 changes: 0 additions & 3 deletions tests/unit/MorphismTest.agda
Original file line number Diff line number Diff line change
Expand Up @@ -31,9 +31,6 @@ support-at-finite n = refl
support-at-infinity : Hom.map support-hom inf ≡ bot
support-at-infinity = refl

maps-disagree-at-positive : ∀ n → zero-test (fin (suc n)) ≢ support (fin (suc n))
maps-disagree-at-positive n ()

reverse-support-at-finite : ∀ n → Hom.map support′-hom (val n) ≡ fin 0
reverse-support-at-finite n = refl

Expand Down
Loading