eyeprolog 1.5.82 → 1.5.84

package/README.md

@@ -95,11 +95,10 @@ npm test
 ```
 The npm command list is deliberately small:
 
-- `npm test` (or `npm run test`): run the release gate, including live upstream conformity checks.
-- `npm run conformance:update:wg17`: update the vendored WG17 syntax cases and their inventory documentation.
+- `npm test` (or `npm run test`): run the release gate, including live upstream conformity checks (WG17 syntax among them).
 - `npm run generate`: rebuild generated library and book files.
 - `npm run benchmark`: run the wall-clock benchmarks.
 
-Use `npm test -- --offline` for a network-free local pass. Focused checks remain available directly, for example `node test/run-regression.mjs docs` or `node test/run-wg17.mjs`; see [test runners](test/README.md). The automatic version hooks still run the release gate, refresh and stage conformance reports, and push the release. Detailed upstream report maintenance is documented in the [conformance guide](test/conformance/README.md).
+Use `npm test -- --offline` for a network-free local pass (this also skips the live-discovered WG17 syntax check, since it has no offline snapshot). Focused checks remain available directly, for example `node test/run-regression.mjs docs`; see [test runners](test/README.md). The automatic version hooks still run the release gate, refresh and stage conformance reports, and push the release. Detailed upstream report maintenance is documented in the [conformance guide](test/conformance/README.md).
 
 EyeProlog is released under the [MIT License](LICENSE.md).

package/conformance-report.md

@@ -1,25 +1,17 @@
 # EyeProlog conformance report
 
-This report combines an executable external conformance gate with the file-based
-conformance corpus under `test/conformance/`. The executable result is measured
-when this report is generated; it is not inferred from fixture counts.
+This report combines a live external conformance gate with the file-based
+conformance corpus under `test/conformance/`. The file-based corpus is
+measured when this report is generated; it is not inferred from fixture counts.
 
 ## Latest Neumerkel evidence
 
-See the tracked [latest Neumerkel conformity report](test/conformance/NEUMERKEL-LATEST.md).
-`npm test` fetches all eight TU Wien sources once and executes the discovered inventory.
-The release workflow then synchronizes this tracked report from those exact successful
-cached source bytes, avoiding a second live fetch and its race window.
-
-## Executable conformance status
-
-| Gate | Passed | Total | Status |
-|---|---:|---:|---|
-| WG17 syntax | 379 | 379 | pass |
-
-The WG17 syntax row executes the vendored 379-case conformity-testing matrix
-against EyeProlog's strict ISO reader/writer. A behavior fix such as operator-token
-spelling therefore changes this report even when no corpus file is added or removed.
+See the tracked [latest Neumerkel conformity report](test/conformance/NEUMERKEL-LATEST.md)
+for the executable external gate, including WG17 syntax conformance: `npm test`
+fetches all eight TU Wien sources (syntax discovered and executed live, not a
+vendored fixture) and executes the discovered inventory. The release workflow
+then synchronizes this tracked report from those exact successful cached source
+bytes, avoiding a second live fetch and its race window.
 
 ## File-based corpus inventory
 
@@ -32,7 +24,7 @@ spelling therefore changes this report even when no corpus file is added or remo
 | context | 11 | 0 | 0 | 0 | 11 |
 | control | 15 | 0 | 0 | 0 | 15 |
 | explicit-tabling | 6 | 0 | 0 | 0 | 6 |
-| iso | 174 | 219 | 0 | 0 | 393 |
+| iso | 176 | 219 | 0 | 0 | 395 |
 | lists | 52 | 3 | 0 | 0 | 55 |
 | modules | 2 | 0 | 0 | 0 | 2 |
 | negation | 8 | 0 | 19 | 0 | 27 |
@@ -45,7 +37,7 @@ spelling therefore changes this report even when no corpus file is added or remo
 | terms | 26 | 3 | 0 | 0 | 29 |
 | unification | 18 | 0 | 0 | 0 | 18 |
 | variables | 16 | 7 | 0 | 0 | 23 |
-| **Total** | **496** | **274** | **19** | **21** | **810** |
+| **Total** | **498** | **274** | **19** | **21** | **812** |
 
 ## DCG conformance clarification
 

package/package.json

@@ -3,7 +3,7 @@
   "publishConfig": {
     "access": "public"
   },
-  "version": "1.5.82",
+  "version": "1.5.84",
   "description": "EyeProlog turns facts and rules into answers and proofs.",
   "type": "module",
   "main": "./index.js",
@@ -46,7 +46,6 @@
   },
   "scripts": {
     "test": "node test/run-all.mjs",
-    "conformance:update:wg17": "node tools/upgrade-wg17.mjs",
     "benchmark": "node test/benchmark.mjs",
     "generate": "node tools/generate-library-autoload-index.mjs && node tools/generate-predicate-reference.mjs && node tools/extract-book-examples.mjs",
     "preversion": "npm test && node test/run-neumerkel.mjs --cached --update-report && node test/run-conformance-report.mjs conformance-report.md && git add test/conformance/NEUMERKEL-LATEST.md conformance-report.md",

package/test/README.md

@@ -1,7 +1,8 @@
 # Test runners
 
-Normally, run `npm test`. Use `npm test -- --offline` to skip live upstream fetching.
-`npm run conformance:update:wg17` refreshes the vendored WG17 snapshot.
+Normally, run `npm test`. Use `npm test -- --offline` to skip live upstream fetching
+(this also skips the WG17 syntax check, which is discovered live and has no
+offline snapshot).
 
 For development, run a focused check directly from the repository root:
 
@@ -9,7 +10,6 @@ For development, run a focused check directly from the repository root:
 node test/run-regression.mjs        # all regression sections
 node test/run-regression.mjs docs   # documentation checks only
 node test/run-conformance-all.mjs   # conformance layers; accepts --offline
-node test/run-wg17.mjs              # vendored WG17 syntax
 node test/run-iso-strict.mjs        # strict ISO core
 node test/run-iso-part2-amendment.mjs
 node test/run-neumerkel.mjs         # live upstream; --cached reproduces last fetch

package/test/conformance/ISO-COMPLIANCE.md

@@ -23,7 +23,7 @@ relevant implementation-dependent overlap choice to an explicit review outcome.
 
 | Requirement | Status | EyeProlog evidence / remaining work |
 | --- | --- | --- |
-| 5.1(a) prepare conforming Prolog text | covered | Clause 6 token/term production and rejection families are mapped by the strict production gate and the complete 379-case vendored WG17 syntax matrix; 7.4 preparation/directive behavior is separately closed. |
+| 5.1(a) prepare conforming Prolog text | covered | Clause 6 token/term production and rejection families are mapped by the strict production gate and the complete WG17 syntax matrix, discovered dynamically and executed live as part of the Neumerkel conformity gate; 7.4 preparation/directive behavior is separately closed. |
 | 5.1(b) execute conforming Prolog goals | covered | Clause 7 term/control/execution semantics, the higher-level 7.10 stream model, 7.11 flags, the 7.12 error envelope, all 8.2-8.17 built-in rows, and Clause 9 arithmetic have explicit executable dispositions. |
 | 5.1(c) reject nonconforming text/read-terms | covered | WG17 negative syntax cases, focused malformed-production/escape/comment/operator cases, read-term syntax errors, and strict rejection of implementation-specific syntax provide explicit rejection evidence across the Clause 6 families. |
 | 5.1(d) document permitted variations | covered | The clause-by-clause [ISO 5.4 decision index](ISO-IMPLEMENTATION-DEFINED.md) records every explicit implementation-defined decision found in the Part 1 + Corrigenda baseline and separately inventories implementation-specific extension families. The index retains an `gap` category for future discoveries; no current release-facing row depends on one. |
@@ -35,7 +35,7 @@ relevant implementation-dependent overlap choice to an explicit review outcome.
 
 | Standard area | Status | Current evidence |
 | --- | --- | --- |
-| Clause 6 — tokens, terms, lists, operators, quoted text | covered | The strict Clause 6 gate maps atomic/variable/compound/operator/list/curly/double-quoted term forms, layout/comments, quoted escapes, integer bases/character codes, floating tokens, and solo/meta token boundaries, with malformed counterparts rejected. The complete 379-case WG17 syntax matrix provides externally sourced production-level expectations, and a cross-profile gate verifies that normal syntax extensions preserve every strict-accepted standard observation. The implementation-defined 6.5/6.6 PCS/collation choices are documented separately. |
+| Clause 6 — tokens, terms, lists, operators, quoted text | covered | The strict Clause 6 gate maps atomic/variable/compound/operator/list/curly/double-quoted term forms, layout/comments, quoted escapes, integer bases/character codes, floating tokens, and solo/meta token boundaries, with malformed counterparts rejected. The complete WG17 syntax matrix, discovered dynamically at run time, provides externally sourced production-level expectations, and a cross-profile gate verifies that normal syntax extensions preserve every strict-accepted standard observation. The implementation-defined 6.5/6.6 PCS/collation choices are documented separately. |
 | 7.1-7.3 — term types, term order, unification | covered | [ISO-TERM-SEMANTICS-MATRIX.md](ISO-TERM-SEMANTICS-MATRIX.md) records the five mutually exclusive strict term types, derived variable/compound/list notions, the complete standard-order classes and implementation-dependent variable-order choice, NSTO MGU/failure behavior, and EyeProlog's permitted consistent occurs-check failure for Part 1's undefined STO ordinary-unification cases. The normal JavaScript API string term is documented as a 5.5.4 extension and rejected at strict program/goal entry. |
 | 7.4 — Prolog text and directives | covered | [ISO-PROLOG-TEXT-EXECUTION-MATRIX.md](ISO-PROLOG-TEXT-EXECUTION-MATRIX.md) closes the preparation/directive rows: declaration semantics, source clauses, cross-text operators/character conversion/flags, initialization order/lifetime, `include/1`, and one-time `ensure_loaded/1`, with implementation-defined cross-text choices indexed under 5.4. |
 | 7.5-7.6 — database and term/clause conversion | covered | [ISO-PROLOG-TEXT-EXECUTION-MATRIX.md](ISO-PROLOG-TEXT-EXECUTION-MATRIX.md) closes static/dynamic and private/public procedure semantics, clause order, logical-update visibility, empty/unknown lifetime, and term/body/clause conversion. Source and runtime assertion conversion recurse through `,/2`, `;/2`, and `->/2`, preserve variable identity, and protect standardized static/control procedures. |
@@ -278,10 +278,11 @@ A release intended to advance ISO conformance uses one canonical command:
 npm test
 ```
 
-That command first fetches and runs the **latest seven Neumerkel conformity
-sources**, then executes the deterministic local conformance, strict ISO,
-vendored WG17 regression, regression/API, examples, documentation and
-architecture gates. The upstream case counts are discovered dynamically.
+That command first fetches and runs the **latest eight Neumerkel conformity
+sources** (WG17 syntax among them, discovered live, not vendored), then
+executes the deterministic local conformance, strict ISO,
+regression/API, examples, documentation and architecture gates. The upstream
+case counts are discovered dynamically.
 
 Useful focused commands are:
 
@@ -289,7 +290,6 @@ Useful focused commands are:
 node test/run-conformance-all.mjs
 node test/run-neumerkel.mjs
 node test/run-iso-strict.mjs
-node test/run-wg17.mjs
 npm test -- --offline
 ```
 
@@ -307,7 +307,7 @@ post-N289 STC drafts remain review input until standardized.
 | Criterion | Status | Evidence |
 | --- | --- | --- |
 | Clause 5 processor obligations have explicit dispositions | covered | `ISO-PROCESSOR-REQUIREMENTS.md` gives explicit outcomes for 5.1-5.5, including the strict/normal extension boundary. |
-| Clause 6 lexical/syntactic requirements have explicit dispositions | covered | the strict Clause 6 production/rejection gate and the 379-case vendored WG17 matrix cover the standard families; cross-profile preservation verifies that normal extensions do not reinterpret strict-accepted text. |
+| Clause 6 lexical/syntactic requirements have explicit dispositions | covered | the strict Clause 6 production/rejection gate and the live-discovered WG17 matrix cover the standard families; cross-profile preservation verifies that normal extensions do not reinterpret strict-accepted text. |
 | Clause 7 semantic requirements have explicit dispositions | covered | 7.1-7.12 now have explicit semantic, implementation-defined, or error-envelope dispositions backed by strict tests and the specialized matrices. |
 | Clause 8 built-in modes/errors have explicit dispositions | covered | `ISO-BUILTIN-MODE-ERROR-MATRIX.md` accounts for 8.2-8.17 condition by condition, using grouped rows only where every grouped condition and its evidence are named |
 | Clause 9 evaluable-functor requirements have explicit dispositions | covered | `ISO-EVALUABLE-FUNCTOR-MATRIX.md` plus the strict arithmetic regression closes the published Part 1 + Corrigenda arithmetic rows |
@@ -315,7 +315,7 @@ post-N289 STC drafts remain review input until standardized.
 | Implementation-specific strict/normal boundary is documented and tested | covered | all 5.5 hooks have explicit dispositions; the WG17 cross-profile gate verifies syntax-preservation for standard text accepted by the strict reader. |
 | Published Corrigenda 1-3 are incorporated | covered | `ISO-CORRIGENDA-MATRIX.md` inventories every published amendment cluster with executable, editorial, or superseded disposition |
 | Current post-N289 draft is tracked without silently changing the published baseline | covered | `STC-DRAFT-STATUS.md` tracks reviewed draft items separately from normative requirements |
-| Latest Neumerkel conformity is a live release gate | covered | `npm test` fetches and executes the eight current TU Wien conformity sources with dynamic inventories; the vendored WG17 matrix remains an offline reviewed-outcome regression layer |
+| Latest Neumerkel conformity is a live release gate | covered | `npm test` fetches and executes the eight current TU Wien conformity sources with dynamic inventories, including WG17 syntax; a small offline corpus of reviewed cases (`test/conformance/wg17-syntax-cases.json`) supplies an additional exact-outcome regression lock, cross-referenced by id against the live-discovered cases |
 | Third-party standard-core regression provenance is retained | covered | adapted Logtalk, Scryer, Trealla, and SWI-Prolog cases retain source identifiers and licenses in `THIRD_PARTY.md` |
 | No unexplained deviation remains in the release-facing ledger | covered | the release-facing ledger contains no remaining `review` rows; documented variation points are implementation-defined/specific or draft-only rather than unexplained deviations. |
 

package/test/conformance/NEUMERKEL-LATEST.md

@@ -1,6 +1,6 @@
 # EyeProlog — latest Neumerkel conformity
 
-Status: **PASS** — **719/719** discovered upstream cases passed.
+Status: **PASS — 1 divergence to be addressed** — **718/719** discovered upstream cases passed.
 
 This tracked report records the latest upstream inventory successfully checked by EyeProlog.
 `npm test` fetches the eight TU Wien sources again and executes the discovered cases.
@@ -15,9 +15,13 @@ Counts are output from upstream, not hard-coded test constants.
 | dif/2 | 26 | 26 |
 | length/2 | 37 | 37 |
 | phrase/2,3 | 58 | 58 |
-| Prologue draft | 33 | 33 |
+| Prologue draft | 32 | 33 |
 | setup_call_cleanup/3 | 25 | 25 |
-| **Total** | **719** | **719** |
+| **Total** | **718** | **719** |
+
+## Known divergences
+
+- **Prologue draft** (line 118): bounded=false: EyeProlog reports no max_integer value at all (ISO 7.11.1.1), so this quad's evaluation_error(int_overflow) | Max = unbounded pair never applies
 
 ## Upstream sources
 

package/test/conformance/NEUMERKEL-LIVE.md

@@ -19,9 +19,12 @@ The Prologue corpus carries one permanent, documented divergence (its
 `max_integer` quad — see `KNOWN_QUAD_DIVERGENCES` in `test/neumerkel.mjs`):
 EyeProlog's `bounded=false` reports no `max_integer` value at all, which
 neither of the quad's two anticipated answers describes. That quad is still
-executed and verified every run; it is reported as a passing, documented
-divergence for as long as it keeps reproducing, and would fail loudly if it
-ever stopped.
+executed every run and counted honestly (it does not report as passing), but
+it does not abort the run the way an unexplained failure would; the tracked
+report and console summary both say "N divergence(s) to be addressed" rather
+than a bare pass/fail count. See `NEUMERKEL-LATEST.md`'s own "Known
+divergences" section for the explanation. If the divergence ever stopped
+reproducing, the counts would simply climb back to a plain, unqualified PASS.
 
 The runner discovers the inventory at run time. A new upstream row is therefore
 executed automatically and a removed row disappears automatically. The syntax

package/test/conformance/README.md

@@ -22,8 +22,10 @@ implements of Part 2 (modules) and Part 3 (definite clause grammar rules),
 including the 2013 module-amendment evidence and the known gaps and deviations.
 Neither part carries a conformance claim. Built-in rows may group closely related conditions only when the
 row names every grouped condition and its executable evidence.
-The exit checklist is embedded in [ISO-COMPLIANCE.md](ISO-COMPLIANCE.md). [WG17-SYNTAX-STATUS.md](WG17-SYNTAX-STATUS.md) records the
-complete one-to-one trace for the vendored active upstream WG17 syntax cases.
+The exit checklist is embedded in [ISO-COMPLIANCE.md](ISO-COMPLIANCE.md). WG17
+syntax cases are discovered live and executed as part of the Neumerkel
+conformity gate (see [NEUMERKEL-LIVE.md](NEUMERKEL-LIVE.md)); there is no
+separate offline coverage ledger to keep in sync.
 [STC-DRAFT-STATUS.md](STC-DRAFT-STATUS.md) separately tracks executable
 implementation questions from the post-N289 working draft (reviewed through the 2026-08-23 items #73-#76); those cases are
 review evidence, not normative ISO claims.
@@ -96,24 +98,19 @@ node test/run-neumerkel.mjs --cached     # exact last fetched bytes; reproductio
 node test/run-neumerkel.mjs --cached --verify-report # verify tracked report against last successful live snapshot
 node test/run-iso-strict.mjs                  # Part 1 + Corrigenda strict-core processor gate
 node test/run-iso-part2-amendment.mjs  # 2013 Part 2 amendment module requirements
-node test/run-wg17.mjs                 # vendored reviewed WG17 syntax regression
 ```
 
 `node test/run-neumerkel.mjs` always fetches the current TU Wien sources. It does not skip a
 fetch because a cache exists. The runner discovers the number of active tests from those sources and fails on any newly introduced case EyeProlog does not pass. If the stable, tracked [NEUMERKEL-LATEST.md](NEUMERKEL-LATEST.md) is stale, normal tests warn rather than turning a passing engine run into a failure. `node test/run-neumerkel.mjs --update-report` performs a fresh live run and refreshes the report; after `npm test`, `node test/run-neumerkel.mjs --cached --update-report` refreshes it from the exact successful cached snapshot; and `node test/run-neumerkel.mjs --cached --verify-report` verifies that snapshot without a second network fetch. Exact bytes, SHA-256 hashes, timestamps, and HTTP validators stay under Git-ignored `.cache/neumerkel/` for inspection/reproduction only. See [NEUMERKEL-LIVE.md](NEUMERKEL-LIVE.md).
 
-The vendored WG17 syntax snapshot is intentionally secondary. Update it with `npm run conformance:update:wg17`. For focused upstream report maintenance and verification:
-
-```sh
-npm run conformance:update:wg17
-node test/run-neumerkel.mjs --update-report
-node test/run-neumerkel.mjs --cached --update-report
-node test/run-wg17.mjs
-```
-
-Every vendored syntax case is still checked against its upstream Codex
-expectation; reviewed exact outcomes are an additional regression lock rather
-than an alternative acceptance rule.
+WG17 syntax cases are discovered live, the same way the other seven TU Wien
+sources are: there is no separate vendored snapshot or update step. A small
+offline corpus, `test/conformance/wg17-syntax-cases.json`, additionally pins
+the exact reviewed EyeProlog outcome for most cases, cross-referenced by id
+against the live-discovered cases in `test/neumerkel.mjs`. Every case, live or
+reviewed, is still checked against its upstream Codex expectation
+(`matchesUpstreamExpectation` in `test/run-wg17.mjs`); a reviewed exact outcome
+is an additional regression lock, never an alternative acceptance rule.
 
 Regenerate the top-level local-corpus report with `node test/run-conformance-report.mjs conformance-report.md`. It links to [NEUMERKEL-LATEST.md](NEUMERKEL-LATEST.md) rather than duplicating live evidence. Counts in the tracked Neumerkel report are generated evidence, not permanent constants in project policy.
 
@@ -141,7 +138,7 @@ Selected cases are adapted from the ISO and standard-core suites of Logtalk,
 Scryer Prolog, Trealla Prolog, and SWI-Prolog. Their upstream identifiers and licenses
 are recorded in [THIRD_PARTY.md](THIRD_PARTY.md).
 
-The corpus has 393 cases in `iso/` and 810 file-based conformance cases in total. Of those, 11 cases in `stc/` are explicitly labelled working-draft review evidence rather than normative ISO claims. The separate vendored strict-reader WG17 matrix is a deterministic regression snapshot; the live Neumerkel gate discovers the current upstream inventory at run time; release/report checks separately verify the tracked `NEUMERKEL-LATEST.md`. The generated `conformance-report.md` records local corpus totals and links to that live evidence. Together with regression, documentation-sync, API, example, and book-example checks, `npm test` is the release gate.
+The corpus has 395 cases in `iso/` and 812 file-based conformance cases in total. Of those, 11 cases in `stc/` are explicitly labelled working-draft review evidence rather than normative ISO claims. The strict-reader WG17 syntax matrix is discovered dynamically and executed live by the Neumerkel gate, the same as its other seven TU Wien sources; release/report checks separately verify the tracked `NEUMERKEL-LATEST.md`. The generated `conformance-report.md` records local corpus totals and links to that live evidence. Together with regression, documentation-sync, API, example, and book-example checks, `npm test` is the release gate.
 
 ## Updating expected output
 

package/test/conformance/cases/iso/trealla_bagof_multi_free_variables.pl

@@ -0,0 +1,23 @@
+% Adapted from Trealla Prolog tests/tests/test0027.pl, test0030.pl, and test0031.pl.
+% See test/conformance/THIRD_PARTY.md.
+foo(a, b, cc).
+foo(a, b, dd).
+foo(b, c, ee).
+foo(b, c, ff).
+foo(c, c, gg).
+foo(d, e, gg).
+
+%% goal: bagof_grouped_by_two_free_variables(X0)
+
+bagof_grouped_by_two_free_variables(Cs) :-
+    bagof(C, foo(_, _, C), Cs).
+
+%% goal: bagof_two_free_variables_quantified(X0)
+
+bagof_two_free_variables_quantified(Cs) :-
+    bagof(C, A^B^foo(A, B, C), Cs).
+
+%% goal: setof_two_free_variables_quantified(X0)
+
+setof_two_free_variables_quantified(Cs) :-
+    setof(C, A^B^foo(A, B, C), Cs).

package/test/conformance/cases/iso/trealla_sub_atom_modes.pl

@@ -0,0 +1,32 @@
+% Adapted from Trealla Prolog tests/tests/test0058.pl.
+% See test/conformance/THIRD_PARTY.md.
+
+%% goal: sub_atom_all(X0, X1, X2, X3)
+
+sub_atom_all(B, L, A, S) :-
+    sub_atom(abc, B, L, A, S).
+
+%% goal: sub_atom_length_2(X0, X1, X2)
+
+sub_atom_length_2(B, A, S) :-
+    sub_atom(abc, B, 2, A, S).
+
+%% goal: sub_atom_before_0_length_2(X0, X1)
+
+sub_atom_before_0_length_2(A, S) :-
+    sub_atom(abc, 0, 2, A, S).
+
+%% goal: sub_atom_before_0(X0, X1, X2)
+
+sub_atom_before_0(L, A, S) :-
+    sub_atom(abc, 0, L, A, S).
+
+%% goal: sub_atom_after_0(X0, X1, X2)
+
+sub_atom_after_0(B, L, S) :-
+    sub_atom(abc, B, L, 0, S).
+
+%% goal: sub_atom_length_2_after_0(X0, X1)
+
+sub_atom_length_2_after_0(B, S) :-
+    sub_atom(abc, B, 2, 0, S).

package/test/conformance/expected/iso/trealla_bagof_multi_free_variables.pl

@@ -0,0 +1,5 @@
+bagof_grouped_by_two_free_variables([cc, dd]).
+bagof_grouped_by_two_free_variables([ee, ff]).
+bagof_grouped_by_two_free_variables([gg]).
+bagof_two_free_variables_quantified([cc, dd, ee, ff, gg, gg]).
+setof_two_free_variables_quantified([cc, dd, ee, ff, gg]).

package/test/conformance/expected/iso/trealla_sub_atom_modes.pl

@@ -0,0 +1,22 @@
+sub_atom_all(0, 0, 3, '').
+sub_atom_all(0, 1, 2, a).
+sub_atom_all(0, 2, 1, ab).
+sub_atom_all(0, 3, 0, abc).
+sub_atom_all(1, 0, 2, '').
+sub_atom_all(1, 1, 1, b).
+sub_atom_all(1, 2, 0, bc).
+sub_atom_all(2, 0, 1, '').
+sub_atom_all(2, 1, 0, c).
+sub_atom_all(3, 0, 0, '').
+sub_atom_length_2(0, 1, ab).
+sub_atom_length_2(1, 0, bc).
+sub_atom_before_0_length_2(1, ab).
+sub_atom_before_0(0, 3, '').
+sub_atom_before_0(1, 2, a).
+sub_atom_before_0(2, 1, ab).
+sub_atom_before_0(3, 0, abc).
+sub_atom_after_0(0, 3, abc).
+sub_atom_after_0(1, 2, bc).
+sub_atom_after_0(2, 1, c).
+sub_atom_after_0(3, 0, '').
+sub_atom_length_2_after_0(1, bc).

package/test/neumerkel.mjs

@@ -11,10 +11,11 @@ import {
   htmlCellText,
   parseWg17SyntaxTable,
   setupInput,
-} from '../tools/upgrade-wg17.mjs';
+} from '../tools/wg17-syntax.mjs';
 import {
   executeWg17Item,
   matchesUpstreamExpectation,
+  readWg17SyntaxOutcomes,
   wg17TestDescription,
 } from './run-wg17.mjs';
 
@@ -308,10 +309,11 @@ function expectedMatches(expected, actual) {
 
 // A permanent, deliberate divergence from an upstream quad: EyeProlog's own
 // implementation choice, not a bug to chase. Keyed by corpus key, then by the
-// answer description's source line. Reported as its own test either way, so a
-// future upstream change that makes the divergence stop reproducing (as
-// happened with the length corpus's occurs-check quad) is caught immediately
-// instead of silently masked.
+// answer description's source line. It is still reported and counted as an
+// ordinary failing test (so the totals stay honest and it self-corrects the
+// moment it stops reproducing, exactly as happened with the length corpus's
+// occurs-check quad); it just does not abort the run the way an unexplained
+// failure would.
 const KNOWN_QUAD_DIVERGENCES = {
   prologue: new Map([
     [118, 'bounded=false: EyeProlog reports no max_integer value at all (ISO 7.11.1.1), ' +
@@ -358,17 +360,9 @@ function runQuadCorpus(reporter, label, item, knownDivergences = new Map()) {
     const divergence = knownDivergences.get(description.line);
     const name = divergence == null
       ? `${label} ${locator} (line ${description.line ?? '?'}) ?- ${formatQuadTerm(program, description.query)}`
-      : `${label} ${locator} (line ${description.line ?? '?'}, documented divergence) ?- ${formatQuadTerm(program, description.query)}`;
+      : `${label} ${locator} (line ${description.line ?? '?'}, documented divergence: ${divergence}) ?- ${formatQuadTerm(program, description.query)}`;
     try {
       reporter.test(name, () => {
-        if (divergence != null) {
-          // Verify the divergence itself, not just tolerate it: if it stops
-          // reproducing, that is exactly the kind of change worth noticing.
-          if (description.ok) {
-            throw new Error(`documented divergence (${divergence}) no longer reproduces; remove it from KNOWN_QUAD_DIVERGENCES`);
-          }
-          return;
-        }
         if (!description.ok) {
           const reason = description.reason ? `: ${description.reason}` : '';
           throw new Error(`${description.kind ?? 'failed'}${reason}`);
@@ -376,10 +370,12 @@ function runQuadCorpus(reporter, label, item, knownDivergences = new Map()) {
       });
       passed++;
     } catch (error) {
-      // Upstream changes often arrive in small clusters. Keep running the
-      // corpus so one live test run exposes every new mismatch instead of
-      // forcing a fix/rerun cycle for each row.
-      failures.push(error);
+      // A documented divergence is expected to keep failing; note it without
+      // aborting the run over it. Upstream changes otherwise often arrive in
+      // small clusters -- keep running the corpus so one live test run
+      // exposes every new mismatch instead of forcing a fix/rerun cycle for
+      // each row.
+      if (divergence == null) failures.push(error);
     }
   });
   return { total: result.results.length, passed, failures };
@@ -392,6 +388,7 @@ export async function executeNeumerkel({ reporter, mode = 'live', cacheDir = def
   reporter.section(`Neumerkel conformity (${mode === 'live' && sourceDir == null ? 'live upstream' : sourceDir != null ? 'source fixtures' : 'cached'})`);
 
   const syntaxCases = materializeSyntaxCases(sources.get('syntax').text);
+  const syntaxOutcomes = readWg17SyntaxOutcomes();
   const syntaxFailures = [];
   let syntaxPassed = 0;
   for (const item of syntaxCases) {
@@ -401,6 +398,16 @@ export async function executeNeumerkel({ reporter, mode = 'live', cacheDir = def
         if (!matchesUpstreamExpectation(item.expected, actual, item)) {
           throw new Error(`syntax #${item.id} expected ${item.expected}; actual ${JSON.stringify(actual)}`);
         }
+        // A case can also carry a reviewed exact-behavior lock (see
+        // wg17-syntax-outcomes.json): a stronger regression check than the
+        // upstream assertion alone, catching a change that still happens to
+        // satisfy the loose Codex expectation.
+        const reviewed = syntaxOutcomes.get(item.id);
+        if (reviewed != null && JSON.stringify(actual) !== JSON.stringify(reviewed)) {
+          throw new Error(
+            `syntax #${item.id} reviewed outcome changed; expected ${JSON.stringify(reviewed)}; actual ${JSON.stringify(actual)}`,
+          );
+        }
       });
       syntaxPassed++;
     } catch (error) {
@@ -514,10 +521,21 @@ export function formatNeumerkelMarkdown({ summary }) {
   const rows = labels.map(([key, label]) => ({ key, label, ...summary[key] }));
   const passed = rows.reduce((sum, row) => sum + row.passed, 0);
   const total = rows.reduce((sum, row) => sum + row.total, 0);
+  const divergenceCount = Object.values(KNOWN_QUAD_DIVERGENCES)
+    .reduce((sum, byLine) => sum + byLine.size, 0);
+  const shortfall = total - passed;
+  // A shortfall fully accounted for by known, documented divergences (see
+  // KNOWN_QUAD_DIVERGENCES and the list below) is not a build failure; any
+  // other shortfall is.
+  const status = shortfall === 0
+    ? 'PASS'
+    : shortfall === divergenceCount
+      ? `PASS — ${shortfall} divergence${shortfall === 1 ? '' : 's'} to be addressed`
+      : 'FAIL';
   const lines = [
     '# EyeProlog — latest Neumerkel conformity',
     '',
-    `Status: **${passed === total ? 'PASS' : 'FAIL'}** — **${passed}/${total}** discovered upstream cases passed.`,
+    `Status: **${status}** — **${passed}/${total}** discovered upstream cases passed.`,
     '',
     'This tracked report records the latest upstream inventory successfully checked by EyeProlog.',
     '`npm test` fetches the eight TU Wien sources again and executes the discovered cases.',
@@ -530,6 +548,14 @@ export function formatNeumerkelMarkdown({ summary }) {
   for (const row of rows) lines.push(`| ${row.label} | ${row.passed} | ${row.total} |`);
   lines.push(`| **Total** | **${passed}** | **${total}** |`);
 
+  if (divergenceCount > 0) {
+    lines.push('', '## Known divergences', '');
+    for (const [key, byLine] of Object.entries(KNOWN_QUAD_DIVERGENCES)) {
+      const label = labels.find(([labelKey]) => labelKey === key)?.[1] ?? key;
+      for (const [line, reason] of byLine) lines.push(`- **${label}** (line ${line}): ${reason}`);
+    }
+  }
+
   lines.push(
     '',
     '## Upstream sources',

package/test/regression/cases-documentation-sync.mjs

@@ -7,8 +7,7 @@ import * as publicApi from '../../src/index.js';
 import { createDefaultRegistry, eyePrologLibraryIndicators, eyePrologNativeLibraryIndicators, eyePrologPortableLibraryIndicators, standardLibrarySources } from '../../src/index.js';
 import { assertEqual, assertIncludes, assertNotIncludes } from '../test-style.mjs';
 import { buildConformanceReport, formatConformanceReport } from '../run-conformance-report.mjs';
-import { renderWg17SyntaxStatus } from '../../tools/report-wg17-syntax-coverage.mjs';
-import { parseWg17SyntaxTable, updateWg17InventoryReferences } from '../../tools/upgrade-wg17.mjs';
+import { parseWg17SyntaxTable } from '../../tools/wg17-syntax.mjs';
 import { executeWg17Item, matchesUpstreamExpectation, readWg17SyntaxFixture } from '../run-wg17.mjs';
 import {
   assertArrayEqual,
@@ -163,29 +162,10 @@ ${profile}`;
       },
     },
     {
-      name: 'WG17 syntax status matches its executable-coverage manifest',
-      run: () => {
-        const filename = path.join(testRoot, 'conformance', 'WG17-SYNTAX-STATUS.md');
-        assertEqual(fs.readFileSync(filename, 'utf8'), renderWg17SyntaxStatus(), 'WG17 syntax status');
-      },
-    },
-    {
-      name: 'WG17 inventory documentation accepts omitted counts and repairs partial updates',
-      run: () => {
-        const countFree = 'The vendored WG17 syntax snapshot is intentionally secondary.';
-        assertEqual(updateWg17InventoryReferences(countFree, 379), countFree, 'no count required');
-        const stale = 'The 366-case vendored WG17 matrix and 365-case WG17 syntax matrix; WG17 matrix has 366 executable cases. An unrelated 366-case suite stays unchanged.';
-        const expected = 'The 379-case vendored WG17 matrix and 379-case WG17 syntax matrix; WG17 matrix has 379 executable cases. An unrelated 366-case suite stays unchanged.';
-        const updated = updateWg17InventoryReferences(stale, 379);
-        assertEqual(updated, expected, 'all stale WG17 counts repaired independently of fixture state');
-        assertEqual(updateWg17InventoryReferences(updated, 379), updated, 'retry is idempotent');
-      },
-    },
-    {
-      name: 'WG17 upgrader accepts omitted HTML table end tags',
+      name: 'WG17 syntax parser accepts omitted HTML table end tags',
       run: () => {
         // HTML permits </td> and </tr> to be omitted. TU Wien uses this
-        // compact form, so the upgrader must not depend on explicit closes.
+        // compact form, so the parser must not depend on explicit closes.
         const rows = Array.from({ length: 120 }, (_, index) =>
           `<tr><td>${index + 1}<td><code>write(${index + 1}).</code><td>ok`).join('\n');
         const html = `<table><tr><th>#<th>Query<th>Codex${rows}</table>`;
@@ -197,7 +177,7 @@ ${profile}`;
       },
     },
     {
-      name: 'WG17 upgrader normalizes presentation non-breaking spaces',
+      name: 'WG17 syntax parser normalizes presentation non-breaking spaces',
       run: () => {
         const rows = Array.from({ length: 120 }, (_, index) =>
           `<tr><td>${index + 1}<td>set_prolog_flag(&nbsp;double_quotes,chars).<td>succeeds`).join('\n');
@@ -207,7 +187,7 @@ ${profile}`;
       },
     },
     {
-      name: 'WG17 upgrader removes presentation footnote markers from Codex text',
+      name: 'WG17 syntax parser removes presentation footnote markers from Codex text',
       run: () => {
         const rows = Array.from({ length: 120 }, (_, index) =>
           `<tr><td>${index + 1}<td>writeq(-(1^2)).<td>- (1^2)&sup3;`).join('\n');
@@ -649,7 +629,7 @@ ${profile}`;
         const publishIndex = publishWorkflow.indexOf('run: npm publish');
         assertEqual(testIndex >= 0 && testIndex < publishIndex, true, 'publish workflow test gate');
         assertEqual(packIndex >= 0 && packIndex < publishIndex, true, 'publish workflow package gate');
-        assertArrayEqual(Object.keys(pkg.scripts).sort(), ['benchmark', 'conformance:update:wg17', 'generate', 'postversion', 'preversion', 'test'], 'small npm command surface');
+        assertArrayEqual(Object.keys(pkg.scripts).sort(), ['benchmark', 'generate', 'postversion', 'preversion', 'test'], 'small npm command surface');
         assertEqual(pkg.scripts.test, 'node test/run-all.mjs', 'full release gate');
         const runner = fs.readFileSync(path.join(packageRoot, 'test', 'run-all.mjs'), 'utf8');
         assertIncludes(runner, 'runOpenRuleBenchChecks(reporter)', 'OpenRuleBench remains in release gate');
@@ -667,28 +647,11 @@ ${profile}`;
         assertEqual(report.total.total >= 475, true, 'conformance case count');
         assertEqual(report.total.positive + report.total.errors + report.total.warnings + report.total.proofs, report.total.total, 'conformance total');
         assertEqual(report.rows.some((row) => row.category === 'legacy-numbered'), false, 'legacy-numbered category');
-        const wg17 = report.executable.find((gate) => gate.name === 'WG17 syntax');
-        const wg17FixtureTotal = readWg17SyntaxFixture().cases.length;
-        assertEqual(wg17?.total, wg17FixtureTotal, 'WG17 executable total');
-        assertEqual(wg17?.passed, wg17FixtureTotal, 'WG17 executable passed');
-        assertArrayEqual(wg17?.failures ?? [], [], 'WG17 executable failures');
         const text = formatConformanceReport(report);
-        assertIncludes(text, `| WG17 syntax | ${wg17FixtureTotal} | ${wg17FixtureTotal} | pass |`, 'report');
+        assertIncludes(text, 'latest Neumerkel conformity report', 'report links live evidence');
         assertIncludes(text, '| variables |', 'report');
         assertIncludes(text, '| Proofs |', 'report');
         assertIncludes(text, '| **Total** |', 'report');
-
-        const failing = buildConformanceReport({
-          wg17Suite: (reporter) => {
-            reporter.section('WG17 syntax');
-            reporter.test('synthetic syntax failure', () => { throw new Error('synthetic mismatch'); });
-            reporter.sectionTotal('WG17 syntax');
-          },
-        });
-        assertEqual(failing.executable[0].passed, 0, 'failing WG17 executable passed');
-        assertEqual(failing.executable[0].total, 1, 'failing WG17 executable total');
-        assertEqual(failing.executionIssues.length, 1, 'failing WG17 report issue count');
-        assertIncludes(formatConformanceReport(failing), '| WG17 syntax | 0 | 1 | fail |', 'failing report');
       },
     },
 

package/test/regression/support.mjs

@@ -512,7 +512,6 @@ export function documentedConformanceMetricIssues() {
   const report = buildConformanceReport();
   const iso = report.rows.find((row) => row.category === 'iso')?.total;
   const total = report.total.total;
-  const wg17 = report.executable.find((gate) => gate.name === 'WG17 syntax')?.total;
   const checks = [
     {
       file: path.join(packageRoot, 'the-art-of-eyeprolog.md'),
@@ -542,6 +541,9 @@ export function documentedConformanceMetricIssues() {
       }
     }
   }
+  // WG17 syntax cases are discovered live (test/neumerkel.mjs), never
+  // vendored, so no document should hard-code a specific count for them: it
+  // would just go stale the next time upstream adds or removes a case.
   for (const file of [
     path.join(packageRoot, 'test', 'conformance', 'README.md'),
     path.join(packageRoot, 'test', 'conformance', 'ISO-COMPLIANCE.md'),
@@ -552,13 +554,9 @@ export function documentedConformanceMetricIssues() {
       ...[...text.matchAll(/\b(\d+)-case[^\n|]*WG17/g)].map((match) => Number(match[1])),
       ...[...text.matchAll(/WG17[^\n|]*?\b(\d+) executable/g)].map((match) => Number(match[1])),
     ];
-    if (claims.length === 0) {
-      const dynamicPolicy = /live Neumerkel|current upstream inventory|discovered dynamically/i.test(text);
-      if (!dynamicPolicy) issues.push(`${relative}: WG17 total or dynamic-upstream policy not found`);
-    }
-    for (const claim of claims) {
-      if (claim !== wg17) issues.push(`${relative}: WG17 count ${claim} != ${wg17}`);
-    }
+    const dynamicPolicy = /live Neumerkel|current upstream inventory|discovered dynamically/i.test(text);
+    if (!dynamicPolicy) issues.push(`${relative}: WG17 dynamic-upstream policy not found`);
+    for (const claim of claims) issues.push(`${relative}: WG17 count ${claim} is hard-coded, not dynamic`);
   }
   return issues.sort();
 }