Claude/happy mendel b4 ie u - #344
Merged
Merged
Conversation
Adds the four standards documents and CI scaffolding the testing survey flagged as missing: - docs/standards/TESTING.adoc (DOC-10) — authoritative test taxonomy (alcotest unit / alcotest E2E / fixture / golden / smoke / microbench), PR-level expectations, deferred-regression-test protocol. Pointers go through CAPABILITY-MATRIX for the live gate count rather than hard-coding it. - docs/standards/PANIC-ATTACK.adoc (DOC-11) — SOP for the estate compliance scanner. Disposition vocabulary (fix/accept/suppress/ defer); standing-suppression record format. Companion workflow .github/workflows/panic-attack.yml runs weekly on Sunday 03:00 UTC + workflow_dispatch; uploads the log as artifact. Per-PR invocation is intentionally out of scope (codeql/semgrep/secret- scanner already cover per-PR at finer granularity). - bench/ scaffold (DOC-12, half) — dune stanza under @bench alias, 4 phase microbenches (lex / parse / typecheck+quantity / codegen-wasm) over 3 shared fixtures. just bench / just bench-record recipes. CI job bench-visibility uploads bench-output.log as artifact + step summary. Visibility-only — no merge-blocking threshold. - bisect_ppx coverage (DOC-12, half) — added as with-test depend in dune-project + affinescript.opam. just coverage recipe (HTML to _coverage/, gitignored). CI job coverage-visibility uploads _coverage as artifact + summary. Visibility-only — no enforced floor (ratchet prerequisite: calibrated baseline). - Stale TESTING-REPORT.adoc + .scm retired in-place with redirects to CAPABILITY-MATRIX.adoc + standards/TESTING.adoc. The 2025-12-29 "47 PASS / 27 FAIL / 63.5%" snapshot has been contradicted by the live gate (>250 / 0 fail) for the entire first half of 2026; a DOC-09 over-claim hazard if left in place. - STATE.a2ml drift-flag refreshed: live gate is 260/260 at 2026-05-19 reconstruction, lifted by subsequent borrow-checker PRs (#240 → 263, return-escape → 271/274, &mut surface → 278/281); exact live number comes from dune runtest --force per commit, not this file. Session-note hygiene-uplift-2026-05-23 recorded. - TECH-DEBT.adoc DOC-10/11/12 entries added. - NAVIGATION.adoc + docs/README.adoc updated to list the new standards and mark TESTING-REPORT as RETIRED. Not yet verified locally (no OCaml toolchain in this remote execution environment); rely on CI for build + test + bench verification. Refs #175, #176
Three single-session Stage D items grouped into one commit: (1) INT-01 follow-up (Refs #178): `Mod::fn(x)` in value-expression position. lib/parser.mly adds one production — `upper_ident COLONCOLON lower_ident` → `ExprField(ExprVar Mod, lower_ident)`, the same AST shape `Mod.fn` already produces. Disambiguated from the existing `Type::Variant` rule by token class (lower vs upper). No resolver change required: Resolve.lower_qualified_value_paths handles both `.` and `::` syntaxes identically because the lowering pattern-matches on the ExprField shape, not on the source separator. Two new fixtures + two alcotest cases pin both `use Mod; Mod::fn(x)` and `use Mod as M; M::fn(x)`. The ledger note "Mod::fn(x) in expression position is a parse error … `::` reserved for Type::Variant" is now closed. (2) BUG-005 deferred regression test landed. The original [[closed-bug]] record had `regression-test-status = "deferred — fixture needs a known-unknown function name in a WasmGC compile path"`. test_gc_unbound_function_loud_fail does exactly that: parses `fn main() -> Int { return totally_undefined_callee(42); }`, feeds it directly to Codegen_gc.generate_gc_module (no resolve step), asserts the emitted error contains the unknown name — which UnboundFunction's format string includes verbatim. A regression to silent drop+null would fail this gate. (3) docs/STDLIB-EXTERN-AUDIT.adoc — triage of all 135 `extern fn` + 24 `extern type` across 11 stdlib files into 4 classes (built-in / typed-boundary-bridge / adapter-Deno / adapter- Node) with per-row status and unblock condition. STDLIB-04 ledger entry now points here; implementation PRs close one row at a time, audit doc is the canonical long-tail roadmap. STATE.a2ml: stage-d-batch-2026-05-23 session note recorded; BUG-005 regression-test-status flipped from "deferred" to "landed 2026-05-23". Not yet verified locally (no OCaml toolchain in remote env). Parser change is low-risk: single production, no Menhir conflict delta expected — `upper_ident COLONCOLON lower_ident` is unambiguous against the prior `upper_ident COLONCOLON upper_ident` rule by token class. CI is the verification surface. Refs #178, Refs STDLIB-04
For DOC-10/11 + STDLIB-04 the human-readable .adoc carries rationale and prose; the .a2ml is the structured form bots / agents / tools consume to decide what counts. Each .a2ml names its human companion in [metadata].human-companion and mirrors its parameters under typed keys. Additions: - docs/standards/TESTING.a2ml — gate command + workflow + job, the six test kinds with their homes/drivers/add-when, the five PR-level rules with their triggers, deferred-regression-test reasons, bench + coverage visibility-only status, cross-refs. - docs/standards/PANIC-ATTACK.a2ml — three schedules (cron / pre- release / ad-hoc) with their blocking severities, the four-key disposition vocabulary (fix/accept/suppress/defer), standing- suppression record format, rule set mirror of panic-attack.toml. - docs/STDLIB-EXTERN-AUDIT.a2ml — totals (135 extern fn / 24 extern type / 11 files), the four classes with their real-impl criterion, per-module rows with status/done/partial/stub-only lists and next-row pointers, implementation policy. Pure additions; no behavioural change. The .adoc files remain authoritative for prose, the .a2ml for parameters. Refs DOC-10, DOC-11, STDLIB-04
Until this commit the AffineScript ↔ typed-wasm interface was
*referenced* from docs/ECOSYSTEM.adoc §"actual AffineScript ↔
typed-wasm contract" + code comments in lib/codegen.ml,
lib/tw_interface.ml, lib/tw_verify.ml — it had no dedicated home.
Adds the dedicated home:
- docs/specs/TYPED-WASM-INTERFACE.adoc (human; formal spec)
- docs/specs/TYPED-WASM-INTERFACE.a2ml (machine-readable companion)
Captures:
* Enforced surface: L7 (aliasing), L10 (linearity), L13 (module
isolation, negative form). L1–L6 / L14–L16 spec-level only.
* Carrier: affinescript.ownership v1 binary layout (LE):
u32 entry_count
per entry: u32 func_index | u8 param_count
| u8[n] param_kind | u8 ret_kind
kind ∈ {0 Unrestricted, 1 Linear, 2 SharedBorrow,
3 ExclBorrow}
v1 is unversioned (accident-of-history); ADR-020 proposed.
* Producer machinery: file/line pointers to lib/codegen.ml
emit symbols (ownership_kind_of_param, build_ownership_section,
ctx.ownership_annots, ...).
* Consumer machinery: lib/tw_verify.ml (per-module) +
lib/tw_interface.ml (cross-module boundary, with
LinearImportCalledMultiple + LinearImportDroppedOnSomePath).
* CLI surface: verify / verify-boundary / verify-bridge /
interface. All `done`; compile-pipeline does NOT auto-gate on
verify today (`--typed-wasm-gate` proposed in roadmap).
* Test coverage: test_tw_isolation.ml (5), E2E Boundary Verify
(3), ownership-section round-trip (3). Honest gap: no C5.1
cross-compat fixtures shipped to hyperpolymath/typed-wasm yet
(INT-12 / CONV-05; planned for Stage E).
* Multi-producer responsibilities: AffineScript = reference
producer, ephapax = second producer (same format), typed-wasm
= spec + Rust verifier + C5.1 cross-compat. ABI changes are
multi-producer; no unilateral changes.
* Stability promise: v1 layout frozen; any change requires
coordinated ADR landings across AffineScript + ephapax +
typed-wasm.
* Deliberate non-features (NF-001..NF-005) catalogued with
blockers + roadmap anchors.
This document supersedes the §"actual AS ↔ typed-wasm contract"
paragraph in ECOSYSTEM.adoc as the source of truth (ECOSYSTEM stays
the index for the spine and satellite registry). The companion
roadmap doc — what is needed to make typed-wasm natural and
optimal — lands in the next commit.
Refs CONV-01..05, INT-12What would have to land here (and what would have to be coordinated with hyperpolymath/typed-wasm + hyperpolymath/ephapax) for typed- wasm to become a first-class compilation target rather than the narrow but correct one it is today. Adds: - docs/specs/TYPED-WASM-ROADMAP.adoc (human) - docs/specs/TYPED-WASM-ROADMAP.a2ml (machine-readable companion) Structure — five tranches, ordered by what unblocks what: Tranche A — Ergonomics (no ABI change, AffineScript-only): A1 `--typed-wasm` compile-time flag (auto-gate) A2 `report-typed-wasm` JSON interface descriptor A3 extract ownership-section emission to lib/tw_ownership_section.ml A4 diagnostic span parity for ownership errors Tranche B — Schema hygiene (additive single-byte, ABI): B1 ADR-020 (proposed): 0xAF sentinel + u8 version prefix on affinescript.ownership; coordinated landing across the three repos. Cheap now, expensive once widening lands. Tranche C — Compiler-side foundations for L1–6 / L14–16: C1 Region inference (NLL/Polonius) — CORE-01 Phase 3 residual C2 Session-type / capability tracking C3 Async-boundary recogniser — CONV-02 / ADR-016 S2..S4 C4 Capability-typed extern annotations Tranche D — Multi-producer ABI for widening (THE big one): D1 ADR-021 (proposed): multi-producer coordination model D2 affinescript.regions (L1–L6 carrier) D3 affinescript.capabilities / .session / .choreography (L14–L16 carriers) Tranche E — Cross-compat closure: E1 INT-12 / CONV-05: AffineScript fixtures into typed-wasm's crates/typed-wasm-verify/tests/cross_compat.rs (C5.1) E2 Rust-verifier parity (out of scope here; typed-wasm-repo work) Sequencing recommendation embedded: A1 → A3 → A2 → E1 → B1 → A4 → D1 → C1..C4 → D2/D3 → E2 The roadmap is NOT a commitment to any timeline. It is the ordered list of what would have to be true for the answer to "how close to end of Stage E" to become "done". Not a behaviour change to the compiler — pure documentation. ADR-020 and ADR-021 are filed as PROPOSED; owner ratification required before either lands. The work-item shells track these in machine-readable form. Refs CONV-01..05, INT-12, CORE-01
Files two ADRs as PROPOSED status in META.a2ml. Both require owner ratification before they bind; neither lands code. ADR-020 — Ownership-section schema versioning Move affinescript.ownership from v1 (unversioned) to v2 (0xAF sentinel + u8 version + entry_count + entries). Backward- compat dispatch: v1 readers fail cleanly on v2 sections (bad entry-count); v2 readers see the sentinel and dispatch. v2.0 entry shape unchanged from v1 — this ADR introduces no new fields, just version discrimination. Why now: v1 is unversioned by accident-of-history. The first multi-producer ABI widening (L1–6 / L14–16 carriers) will need version discrimination; adding the version field WHILE STILL AT V1 is cheap, retrofitting it under load is expensive. Coordinated landing across hyperpolymath/affinescript + hyperpolymath/typed-wasm + hyperpolymath/ephapax. Verifier ships parse-support first; producers flip together. A deprecation window then removes v1 parse. ADR-021 — Multi-producer ABI coordination model Four-axis formal model replacing today's ad-hoc cross-repo coordination: (1) Spec authority — OCaml verifier today; flips to typed-wasm Rust crate on C5.1 closure (INT-12 / CONV-05). (2) Coordinated landing protocol — verifier ships parse-support first; producers flip together. (3) Test parity protocol — each ABI change requires fixture contributions to the cross-compat suite from each producer. (4) Conflict resolution — in-flight ABI proposals serialised via a coordination queue owned by typed-wasm. Why now: every prior ABI conversation has been one-off and ad-hoc; the failure mode is parallel incompatible landings. Cost is cheap before widening pressure arrives. Both ADRs cross-reference docs/specs/TYPED-WASM-ROADMAP.adoc §"Tranche B / D" which lists them as B1 / D1 work-items. ADR-020 is itself the first test of the ADR-021 protocol — its landing exercises the coordinated landing model. Refs CONV-04, CONV-05, INT-12
Wires the new TYPED-WASM-INTERFACE + TYPED-WASM-ROADMAP into the documentation lattice so every entry point reaches them: - docs/ECOSYSTEM.adoc §"The actual AS↔typed-wasm contract": the intro IMPORTANT block now names TYPED-WASM-INTERFACE.adoc as authoritative formal spec, ROADMAP.adoc as the widening plan, and ADR-020 / ADR-021 as the proposed ADRs. - docs/CAPABILITY-MATRIX.adoc typed-wasm row: adds L13 to the enforced-level list (was just L7+L10), links INTERFACE.adoc + .a2ml + ROADMAP.adoc. - docs/TECH-DEBT.adoc Stage E section header: links INTERFACE + ROADMAP + names the two proposed ADRs. CONV-02 / CONV-04 entries cross-reference roadmap tranches. Section A gets DOC-13 (INTERFACE), DOC-14 (ROADMAP), DOC-15 (ADR-020/021 PROPOSED). - docs/README.adoc §"Language Specifications": adds entries for TYPED-WASM-INTERFACE, TYPED-WASM-ROADMAP, SETTLED-DECISIONS. - NAVIGATION.adoc: adds "Typed-WASM interface (Stage E runway)" subsection with both spec links. - STATE.a2ml: session-note typed-wasm-formalisation-2026-05-23 records the spec/roadmap/ADR triad and the cross-link work. Pure documentation. No compiler code touched. Refs DOC-13, DOC-14, DOC-15, CONV-04
🔍 Hypatia Security ScanFindings: 132 issues detected
View findings[
{
"reason": "Stray AI.a2ml in root -- use 0-AI-MANIFEST.a2ml only",
"type": "banned",
"file": "AI.a2ml",
"action": "delete",
"rule_module": "root_hygiene",
"severity": "high"
},
{
"reason": "Superseded by 0-AI-MANIFEST.a2ml",
"type": "banned",
"file": "AI.djot",
"action": "delete",
"rule_module": "root_hygiene",
"severity": "high"
},
{
"reason": "Issue in quality.yml",
"type": "missing_workflow",
"file": "quality.yml",
"action": "create",
"rule_module": "workflow_audit",
"severity": "high"
},
{
"reason": "Issue in security-policy.yml",
"type": "missing_workflow",
"file": "security-policy.yml",
"action": "create",
"rule_module": "workflow_audit",
"severity": "medium"
},
{
"reason": "Action hyperpolymath/standards/.github/workflows/governance-reusable.yml@main needs attention",
"type": "unpinned_action",
"file": "governance.yml",
"action": "pin_sha",
"rule_module": "workflow_audit",
"severity": "high"
},
{
"reason": "Action actions/checkout@v6 needs attention",
"type": "unpinned_action",
"file": "publish-jsr.yml",
"action": "pin_sha",
"rule_module": "workflow_audit",
"severity": "medium"
},
{
"reason": "Action denoland/setup-deno@v2 needs attention",
"type": "unpinned_action",
"file": "publish-jsr.yml",
"action": "pin_sha",
"rule_module": "workflow_audit",
"severity": "medium"
},
{
"reason": "TypeScript file detected -- banned language",
"type": "banned_language_file",
"file": "/home/runner/work/affinescript/affinescript/affinescript-deno-test/example/smoke_driver.ts",
"action": "flag",
"rule_module": "cicd_rules",
"severity": "critical"
},
{
"reason": "TypeScript file detected -- banned language",
"type": "banned_language_file",
"file": "/home/runner/work/affinescript/affinescript/affinescript-deno-test/cli.ts",
"action": "flag",
"rule_module": "cicd_rules",
"severity": "critical"
},
{
"reason": "TypeScript file detected -- banned language",
"type": "banned_language_file",
"file": "/home/runner/work/affinescript/affinescript/affinescript-deno-test/mod.ts",
"action": "flag",
"rule_module": "cicd_rules",
"severity": "critical"
}
]Powered by Hypatia Neurosymbolic CI/CD Intelligence |
Uh oh!
There was an error while loading. Please reload this page.
hyperpolymath pushed a commit
that referenced
this pull request
May 24, 2026
Sixth operational finding, captured this session: when a PR is squash-merged the squashed commit on main has a new SHA, distinct from the source-branch commits. If you reset your local branch to main, `git status` reports N "ahead of origin/branch" — but those N are just main-side commits the obsolete remote-tip never saw, not unpushed work. Adds: * Recognition checklist (PR closed/merged, local matches main, remote tip is pre-merge SHA, `git log origin/<branch>..HEAD` shows other people's work). * Two safe fixes by intent: git push origin --delete <branch> # done git push --force-with-lease origin <branch> # align * Note: --force-with-lease is safe here because there's no unmerged remote work to clobber. Pure documentation. Same branch / same PR shape as the previous ops-notes commit. Refs PR #344 (the trigger case)
hyperpolymath added a commit
that referenced
this pull request
May 24, 2026
#348) …erience Captures five operational findings from the parallel-claude coordinator-feedback turn (2026-05-24), as a single "Agent operations notes" section: * CI signal reliability — "PR merged" does NOT mean "build green" on this repo today; auto-merge fires through red builds. Recently-merged PRs (#334/#335/#336/#344) all landed with `build` red; the red persisted until PR #346. * Reading CI logs — WebFetch on the Actions UI returns React skeleton; use mcp__github__pull_request_read get_check_runs / get_status, hand back to user for actual log lines via `gh run view --log-failed`. * Known-failing baseline checks — vscode-smoke (npm 404), migration-assistant (fixed by #342 but stale-base branches red), governance/Language anti-pattern policy (flags approved TS exemptions), Hypatia 143-finding comment (mostly the same exemption hits). Don't waste turns investigating per-PR; only investigate *changes* in this set. * Branching discipline — `git fetch origin main && git rebase origin/main` immediately before push, not just at branch- creation. Claude 1's #337 accidentally reverted #334+#335 from a stale base; cheap to prevent. * Test-fixture hygiene — when adding a new declaration shape (extern fn, etc.), test it against EVERY downstream consumer (parse/resolve/typecheck/interp/codegen). PR #346's FnExtern interp bug survived since the interpreter was written because no test fed an inline `extern fn` to Interp.eval_program. Pure documentation. Zero behavioural risk. Refs PR #346, #335, #337, #344 --------- Co-authored-by: Claude <noreply@anthropic.com>
hyperpolymath added a commit
that referenced
this pull request
May 24, 2026
#349) …gate) Two `print(*r)` example fragments inside doc-comment blocks (lines 1115 and 1338) contained the literal substring `(*r`, which OCaml's lexer treats as a nested comment opener — bringing the outer comment to depth 2 with only one matching `*)` to close it. Result: "Comment not terminated" at EOF on every `dune build`. Marker count was 126 `(*` vs 124 `*)` (off by 2, one per offending example); after this patch it is 124 / 124. This has blocked main's `dune build` since #335 (CORE-01 pt3 Slice A) merged, which is why every subsequent PR (#341, #344, the open #346) has shown `build` red regardless of its own changes. Tiny fix; high leverage. Fix: insert a single space inside the parens — `print( *r)` — so the lexer sees `(` `*` `r` `)` rather than the comment-open token `(*` followed by `r`. Reader-friendly; the example still reads the same. No behavioural change. No code touched. Comment text only. Diagnosed by parallel claude (INT-03 / S5 thread). Refs #335 Co-authored-by: Claude <noreply@anthropic.com>
3 tasks
hyperpolymath added a commit
that referenced
this pull request
May 24, 2026
…quash) (#350) ## Summary Recovers three doc artefacts that were dropped by PR #344's squash-merge despite being on `claude/happy-mendel-b4IeU` when it merged: - `.machine_readable/6a2/META.a2ml` — ADR-020 + ADR-021 records (both `status="accepted"`, `ratified="2026-05-24"`). - `docs/specs/TYPED-WASM-COORDINATION-LEDGER.adoc` + `.a2ml` — the cross-repo coordination queue mandated by ADR-021. Live entry **Q-001** (ADR-020 v1→v2 schema versioning) in state `verifier-parse-shipping`. - `docs/TECH-DEBT.adoc` — DOC-10..16 ledger entries restored. Cherry-picked from the original local commits (`977896c` + `243f11b`) and squashed; the restored files are byte-identical to their pre-loss state. ## Why this matters - ADR-020 and ADR-021 are **owner-ratified** (2026-05-24). Their cross-repo coordination protocol is what the `claude/tw-verify-v2-parse` PR (incoming) lands its code under. - Without these records on main, sister repos (`hyperpolymath/typed-wasm`, `hyperpolymath/ephapax`) have no canonical text to mirror. - Cross-references from `TYPED-WASM-INTERFACE.adoc` and `TYPED-WASM-ROADMAP.adoc` to these ADRs are currently dangling on main. ## Why zero risk Pure documentation. No source code touched. CI surface unchanged. ## Test plan - [ ] `dune build` clean (no source files changed) - [ ] `dune runtest --force` baseline holds - [ ] No new lints Refs PR #344 (the squash-merge that lost these), ADR-020, ADR-021. --- _Generated by [Claude Code](https://claude.ai/code/session_01WHUYQEPKgQU6jBgUj4snYU)_ Co-authored-by: Claude <noreply@anthropic.com>
Open
3 tasks
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 freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.