Polish release docs and add CI checks

- Correct stale references and headline-path wording in READMEs and Lean docstrings
- Document build steps, comparator requirements, sweep scope, and coverage
- Unify Apache headers across Lean and C++ sources
- Add a bwrap-free axiom audit for the headline theorems
- Add a Gitea workflow running lake build, the axiom audit, C++ tests, and coverage
- Match LLVM coverage tool versions to the compiler and recreate stale build dirs
This commit is contained in:
andrew committed 2026-11-17 21:29:16 -05:00
1 parent 73e371fb40
commit 273ad72f0c
34 files changed
+431 -133

No files matched your search

@@ -20,8 +20,8 @@ last reported key, `q = beginSlack` and `r = reportedBelowBegin`, put `C = q + 1
(so `C ≥ 1` in the clamping cell). The upper endpoint is `begin.key` when the interval
`[keyAfter L, begin.key)` holds `C` keys (or `C + 1` keys when `end.key = ⊥`), and
`keyAfter L` otherwise. The witness `res = [[5], [6]]`, `end.key = [1]`,
`begin.key = [6,0,0]`, `begin.offset = -3`, `end.offset = 3` is kernel-checked in
`Counterexample.lean` (`nch_M_not_minimal`): every compatible database then has
`begin.key = [6,0,0]`, `begin.offset = -3`, `end.offset = 3` refutes the rule:
every compatible database then has
`begin.key = [6,0,0]` strictly above `keyAfter last = [6,0]`, but the extra key `[6,0]`
lies in no sound range, so the `max` rule over-collects it.