Proof debt/standards 134 coq preservation admit#110
Closed
hyperpolymath wants to merge 2 commits into
Closed
Conversation
…Refs standards#134)
`formal/Semantics.v` carries `Qed.` for `preservation` at L3327 with an
in-file comment claiming the proof was "FULLY CLOSED (2026-04-27). Zero
Admitted." That comment is unsubstantiated — `coqc` (8.18.0) rejects the
proof script with:
Error: (in proof preservation): Attempt to save an incomplete proof
(there are remaining open goals).
Consequence: `rust-ci.yml`'s "Coq proofs" job has been failing on every
push to main (3 most-recent runs on main all `conclusion: failure`,
back to 2026-05-19), and PR #87 transitively propagated the bogus
"Qed, closed 2026-04-27" claim from the comment into ROADMAP and
PROOF-NEEDS.
This change marks `preservation` as `Admitted.` so the formalisation
builds and the documentation accurately reflects proof state:
- `formal/Semantics.v` — `Qed.` → `Admitted.` on `preservation`;
in-file comment rewritten honestly (the supporting lemmas
`region_env_perm_typing`, `region_add_typing`,
`region_shrink_preserves_typing` remain Qed).
- `ROADMAP.adoc` — Coq status row + "v0.1.0 close `preservation`"
checkbox both flipped to reflect `Admitted`; admitted-proofs counter
0 → 1.
- `PROOF-NEEDS.md` — "current state" row + admitted counter corrected;
"what needs proving" now lists `preservation` as the top item with a
precise pointer to L3215–L3326 of the proof script; the prior
reconciliation note rewritten to record the propagated-lie failure
mode honestly; priority raised MEDIUM → MEDIUM-HIGH.
Verified locally on Coq 8.18.0:
cd formal && coq_makefile -f _CoqProject -o Makefile.coq && make -f Makefile.coq
→ exit 0; Syntax.vo + Typing.vo + Semantics.vo all produced.
Refs standards#134 (NOT Closes — joint-close on agreement; #134 also has
ephapax#88 open for the 14 SPDX headers, which is independent of this).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
The companion change to the `Qed.` → `Admitted.` honesty fix. Running `Show. Show Existentials.` immediately before the `Admitted.` reveals that the proof script leaves **910 goals open**, which is exactly **35 (step rules) × 26 (typing rules) = 910** — the *full combinatorial after `induction Hstep; inversion Htype; subst;`*. The chain of `try solve [...]` tactics in `Semantics.v` L3228–L3326 closes **zero** of them. This contradicts the in-file comment block at L3269–L3296 which claims "only ONE case remains open: S_Region_Step + T_Region_Active." That comment was either always optimistic, or captured a transient earlier state that silently regressed (no CI gate caught it because the Coq job has been failing for ≥3 runs — see ephapax#92's parent issue). `formal/PRESERVATION-HANDOFF.md` records: - How the diagnostic was produced (`Show.` recipe + restore step). - The exact open-goal count (910) and combinatorial decomposition (35 × 26). - Why the existing `try solve` chain fails (cross-case `discriminate` has no equation after `inversion Htype; subst`; diagonal-case `eassumption` rarely matches because the IH gives body typing but the constructor needs step-result typing). - A realistic remediation plan (move `remember` out of `try (...)`; rewrite cross-case closure arm-by-arm; write per-step-rule diagonal proofs; the `S_Region_Step + T_Region_Active` case may need a region-env *weakening* lemma which does not yet exist). - Effort estimate: days–weeks, not "one more tactic." - Anti-patterns to avoid (mass-`Admitted.` per case; replacing `induction` with `inversion`). - The checklist for unwinding the honest-mark when `preservation` is finally closed. The supporting lemmas needed already exist (`subst_preserves_typing`, `region_env_perm_typing`, `region_add_typing`, `region_shrink_preserves_typing`, `values_dont_step`) — they all have Qed. The missing piece is the glue + a region-env weakening lemma for non-values. Refs standards#134. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Owner
Author
|
Closing as superseded. This branch had two commits:
Both targets discharged on main. Closing without cherry-pick. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.